CN118534229A - Transformer fault determination method and device based on multi-sensor - Google Patents

Abstract

The application provides a fault determining method and device for a transformer based on multiple sensors. The method comprises the following steps: acquiring operation information of a transformer; acquiring a weight coefficient corresponding to the operation information, wherein the weight coefficient corresponds to the operation information one by one; calculating a weighted average value of the weight coefficient and the running information to obtain a comprehensive calculation result; and obtaining a first analysis result according to the magnitude relation between the comprehensive calculation result and the preset operation threshold, wherein the first analysis result is used for representing whether the transformer has faults or not. The method solves the problem of lower accuracy of transformer fault detection in the prior art.

Description

Transformer fault determination method and device based on multi-sensor

Technical Field

The present application relates to the technical field of transformer fault diagnosis, and in particular to a transformer fault determination method, device, computer program product and transformer fault detection system based on multiple sensors.

Background Art

Transformer fault diagnosis refers to monitoring and analyzing various parameters and signals during the operation of the transformer to determine whether the transformer has a fault and find out the specific cause and location of the fault. However, in the current transformer fault detection scheme, the accuracy of fault detection is low.

Summary of the invention

The main purpose of the present application is to provide a transformer fault determination method, device, computer program product and transformer fault detection system based on multiple sensors, so as to at least solve the problem of low accuracy of transformer fault detection in the prior art.

In order to achieve the above-mentioned purpose, according to one aspect of the present application, a transformer fault determination method based on multiple sensors is provided, including: obtaining operating information of the transformer, wherein there are multiple operating information, the operating information is collected by sensors, there are multiple sensors, and the operating information corresponds to the sensors one-to-one; obtaining a weight coefficient corresponding to the operating information, wherein the weight coefficient corresponds to the operating information one-to-one; calculating a weighted average of the weight coefficient and the operating information to obtain a comprehensive calculation result; obtaining a first analysis result based on the size relationship between the comprehensive calculation result and a preset operating threshold, wherein the first analysis result is used to characterize whether the transformer is faulty, and when the comprehensive calculation result is greater than or equal to the preset operating threshold, the first analysis result characterizes that the transformer is normal, and when the comprehensive calculation result is less than the preset operating threshold, the first analysis result characterizes that the transformer is faulty.

Optionally, before obtaining the operating information of the transformer, the method further includes: obtaining initial operating information; performing a data consistency check on the initial operating information to obtain a check result, wherein the check method includes at least one or more of numerical analysis, variance analysis, correlation analysis, and time series analysis; when the check result indicates that the data is inconsistent, performing data repair on the initial operating information to obtain the operating information, wherein the data repair method includes at least one or more of re-collecting data, filling in missing data, and eliminating abnormal data.

Optionally, after obtaining the operating information of the transformer, the method further includes: constructing a first detection model, wherein the first detection model is trained by a Bayesian algorithm using multiple sets of training data, and each set of training data in the multiple sets of training data includes historical operating information obtained within a historical time period, and a historical second analysis result corresponding to the historical operating information, wherein the historical second analysis result is used to characterize whether the historical transformer is faulty; inputting the operating information into the first detection model to obtain a second analysis result corresponding to the operating information.

Optionally, after obtaining the operating information of the transformer, the method further includes: constructing a second detection model, wherein the second detection model is trained using multiple sets of training data through a Bayesian algorithm and an auxiliary algorithm, and each set of training data in the multiple sets of training data includes historical operating information obtained within a historical time period, and a historical third analysis result corresponding to the historical operating information, wherein the historical third analysis result is used to characterize whether the historical transformer is faulty, and the auxiliary algorithm includes at least one or more of an LSTM algorithm, a LightGBM algorithm, a TPE algorithm, a BOA-LSTM algorithm, and a BART algorithm; the operating information is input into the second detection model to obtain a third analysis result corresponding to the operating information.

Optionally, before obtaining the operating information of the transformer, the method further includes: obtaining interface definition information, wherein the interface definition information is information on a data transmission format, a transmission protocol and a communication method of a predefined interface of the sensor; and configuring the interface of the sensor according to the interface definition information.

Optionally, before obtaining the operating information of the transformer, the method further includes: obtaining configuration definition information, wherein the configuration definition information is pre-defined information of a sampling frequency, resolution, abnormal alarm threshold and power mode of the sensor; and configuring the sensor according to the configuration definition information.

Optionally, after obtaining the first analysis result based on the size relationship between the comprehensive calculation result and the preset operating threshold, the method also includes: generating early warning information when the first analysis result characterizes the transformer fault; based on the early warning information, controlling the early warning device to turn on, wherein the early warning device includes at least a buzzer and/or an LED light.

According to another aspect of the present application, a multi-sensor based transformer fault determination device is provided, comprising: a first acquisition unit, used to acquire operating information of the transformer, wherein there are multiple operating information, the operating information is acquired by using sensors, there are multiple sensors, and the operating information and the sensors correspond one-to-one; a second acquisition unit, used to acquire a weight coefficient corresponding to the operating information, wherein the weight coefficient and the operating information correspond one-to-one; a calculation unit, used to calculate a weighted average of the weight coefficient and the operating information to obtain a comprehensive calculation result; a determination unit, used to obtain a first analysis result based on the size relationship between the comprehensive calculation result and a preset operating threshold, wherein the first analysis result is used to characterize whether the transformer is faulty, and when the comprehensive calculation result is greater than or equal to the preset operating threshold, the first analysis result characterizes that the transformer is normal, and when the comprehensive calculation result is less than the preset operating threshold, the first analysis result characterizes that the transformer is faulty.

According to another aspect of the present application, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the steps of any one of the multi-sensor based transformer fault determination methods are implemented.

According to another aspect of the present application, a transformer fault detection system is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing any one of the multi-sensor based transformer fault determination methods.

By applying the technical solution of the present application, information detected by different sensors can be collected at the same time to obtain multiple operating information, which can be comprehensively analyzed to evaluate whether the transformer is faulty. Compared with the single sensor evaluation method, it has higher accuracy and can improve the accuracy of transformer fault detection.

BRIEF DESCRIPTION OF THE DRAWINGS

The drawings constituting part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

FIG1 shows a hardware structure block diagram of a mobile terminal for executing a transformer fault determination method based on multiple sensors provided in an embodiment of the present application;

FIG2 is a schematic flow chart of a transformer fault determination method based on multiple sensors according to an embodiment of the present application;

FIG3 shows a schematic diagram of a process flow of transformer fault detection based on multiple sensors;

FIG4 shows a schematic diagram of a process flow of data consistency detection;

FIG5 shows a schematic diagram of the process of uncertain diagnosis;

FIG6 shows a schematic diagram of the process of interface standardization;

FIG7 shows a schematic diagram of a dynamic configuration process;

FIG8 shows a schematic diagram of the flow of the plug-in mechanism;

FIG9 shows a structural block diagram of a transformer fault determination device based on multiple sensors according to an embodiment of the present application.

The above drawings include the following reference numerals:

102, processor; 104, memory; 106, transmission device; 108, input and output devices.

DETAILED DESCRIPTION

It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

As introduced in the background technology, the accuracy of transformer fault detection in the prior art is low. To solve the above problem, the embodiments of the present application provide a transformer fault determination method, device, computer program product and transformer fault detection system based on multiple sensors.

The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, FIG1 is a hardware structure block diagram of a mobile terminal of a fault determination method for a transformer based on a multi-sensor according to an embodiment of the present invention. As shown in FIG1 , the mobile terminal may include one or more (only one is shown in FIG1 ) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input/output device 108 for communication functions. It can be understood by those skilled in the art that the structure shown in FIG1 is only for illustration and does not limit the structure of the mobile terminal. For example, the mobile terminal may also include more or fewer components than those shown in FIG1 , or have a configuration different from that shown in FIG1 .

The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the display method of device information in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific network example may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

In this embodiment, a transformer fault determination method based on multiple sensors is provided, which runs on a mobile terminal, a computer terminal or a similar computing device. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

FIG2 is a flow chart of a method for determining a transformer fault based on multiple sensors according to an embodiment of the present application. As shown in FIG2 , the method includes the following steps:

Step S201, obtaining operation information of the transformer, wherein the operation information is multiple, the operation information is collected by using a sensor, there are multiple sensors, and the operation information corresponds to the sensors one by one;

Specifically, multiple sensors can be used to detect the transformer, with one sensor detecting one type of operating information. Of course, multiple sensors can also detect one type of operating information, for example, a temperature sensor detects the temperature of the transformer, and a humidity sensor detects the humidity of the transformer.

Step S202, obtaining a weight coefficient corresponding to the operation information, wherein the weight coefficient corresponds to the operation information one by one;

Specifically, weight coefficients of different operation information are preset, and the weight information represents the weight of the operation information in the data fusion.

Step S203, calculating the weighted average of the weight coefficient and the operation information to obtain a comprehensive calculation result;

Specifically, by integrating the data detected by different sensors through a fusion algorithm, the calculation result of the comprehensive transformer, that is, the comprehensive calculation result, can be obtained.

For example, multiple sensors can be divided into temperature sensors, humidity sensors and gas sensors. The data collected by the temperature sensor is T(t), the data collected by the humidity sensor is H(t), and the data collected by the gas sensor is G(t). Finally, S(t)=f(T(t), H(t), G(t)) is obtained; the weight coefficients α=0.4, β=0.3, and γ=0.3 are set respectively, which specifically means that the temperature accounts for 40% of the weight in the fusion, and the humidity and gas concentration each account for 30% of the weight. The data at time t are as follows:

T(t) = 80 degrees Celsius;

H(t)＝60％;

G(t) = 0.5 ppm;

According to the information fusion algorithm, the comprehensive information S(t)=0.4×80+0.3×60+0.3×0.5=2+18+0.15=50.15 can be calculated. Finally, by adjusting the weight coefficient, the influence of different sensor data in the fusion can be flexibly controlled, thereby achieving more accurate and effective information fusion.

Step S204, obtaining a first analysis result based on the size relationship between the above-mentioned comprehensive calculation result and the preset operating threshold, wherein the above-mentioned first analysis result is used to characterize whether the above-mentioned transformer is faulty. When the above-mentioned comprehensive calculation result is greater than or equal to the above-mentioned preset operating threshold, the above-mentioned first analysis result characterizes that the above-mentioned transformer is normal; when the above-mentioned comprehensive calculation result is less than the above-mentioned preset operating threshold, the above-mentioned first analysis result characterizes that the above-mentioned transformer is faulty.

Specifically, after obtaining the comprehensive calculation result of the comprehensive evaluation of the transformer, the first analysis result is determined directly by threshold comparison. For example, the preset operating threshold may be 50, 60, 70, 80, etc. In the above embodiment, if the preset operating threshold is 50 and the comprehensive calculation result is 50.15, then the transformer is not faulty.

Through this embodiment, information detected by different sensors can be collected at the same time to obtain multiple operating information, which can be comprehensively analyzed to evaluate whether the transformer is faulty. Compared with the single sensor evaluation method, it has higher accuracy and can improve the accuracy of transformer fault detection.

In addition, the above scheme can be applied to the transformer fault detection system. The different functions in this scheme can adopt a modular architecture in the system. Each module is independently responsible for a specific function, which is convenient for adding or replacing modules, thereby ensuring the flexible scalability of the system and effectively managing the uncertainty of the diagnostic results, thereby significantly enhancing the system's fault diagnosis capabilities in complex environments.

The transformer fault diagnosis scheme based on multi-sensor information fusion proposed in this scheme refers to the simultaneous collection and fusion of multiple parameter information from different sensors, and the comprehensive analysis of this information, so as to improve the accuracy and reliability of transformer faults. In this scheme, data fusion technology can be used to integrate data from different sensors, and these data can be processed and analyzed using pattern recognition, statistical analysis and other methods to achieve accurate diagnosis of transformer faults. By comprehensively considering multiple parameter information, the accuracy of fault diagnosis can be improved, and the operating status of the transformer can be more comprehensively understood, and timely measures can be taken for maintenance and repair to ensure the safe and reliable operation of the transformer.

Specifically, as shown in Figure 3, the main functions of the solution include data acquisition, information fusion, fault diagnosis and system feedback. Data acquisition is carried out through multiple sensors and also includes system expansion. The system expansion mainly includes interface standardization, dynamic configuration and plug-in mechanism configuration, as well as data consistency detection, data reception, data preprocessing and uncertain diagnosis.

Specifically, data collection refers to collecting data through sensors, fusing the collected data, calculating the comprehensive calculation results according to the preset weight coefficients, and judging whether the transformer has a fault based on the comprehensive calculation results. The results of these diagnoses can also be fed back to the operator. Through the collaborative operation between the multi-functional modules, this solution can realize the full process operation of data collection, information fusion, fault diagnosis and automatic feedback, which improves the overall accuracy and stability of the solution and helps to timely discover and deal with potential fault problems.

Specifically, the preprocessed data is fused, and the data from different sensors are fused together through a fusion algorithm to obtain comprehensive transformer status information, wherein the multiple sensors can be divided into temperature sensors, humidity sensors and gas sensors. The data collected by the temperature sensor is T(t), the data collected by the humidity sensor is H(t), and the data collected by the gas sensor is G(t). Finally, S(t)=f(T(t), H(t), G(t)) is obtained.

The three indicators of temperature, humidity and gas concentration are currently monitored, and the weight coefficients are set to α = 0.4, β = 0.3, and γ = 0.3, which specifically means that temperature accounts for 40% of the weight in the fusion, and humidity and gas concentration each account for 30% of the weight. The data at time t are as follows:

T(t) = 80 degrees Celsius;

H(t)＝60％;

G(t) = 0.5 ppm;

According to the information fusion algorithm, the comprehensive information S(t)=0.4×80+0.3×60+0.3×0.5=2+18+0.15=50.15 can be calculated. Finally, by adjusting the weight coefficient, the influence of different sensor data in the fusion can be flexibly controlled, thereby achieving more accurate and effective information fusion.

In order to further improve the validity and accuracy of the data, before obtaining the operating information of the transformer, the above method also includes the following steps: obtaining initial operating information; performing data consistency check on the above initial operating information to obtain a test result, wherein the test method includes at least one or more of numerical analysis, variance analysis, correlation analysis, and time series analysis; when the above test result indicates that the data is inconsistent, performing data repair on the above initial operating information to obtain the above operating information, wherein the data repair method includes at least one or more of re-collecting data, filling in missing data, and eliminating abnormal data.

In this solution, the data can be checked for consistency to determine whether there is any abnormal data. If the data is inconsistent, that is, if there is abnormal data, the data can be repaired, thereby further improving the validity and accuracy of the data.

The initial operation information is the operation information obtained at the initial moment, that is, the operation information obtained when the transformer just starts to operate.

Specifically, the transformer data can be numerically analyzed to calculate the mean, median, standard deviation and other statistics of each variable to understand the distribution of the data and the situation of outliers. Variance analysis can be used to test whether there are significant differences in the means of transformer data. The ANOVA method can be used to compare the means between multiple variables to determine whether they are consistent. The correlation between transformer data can be tested by correlation analysis to determine whether there is a consistent relationship between the data. The Pearson correlation coefficient or the Spearman rank correlation coefficient can be calculated to evaluate the correlation between variables. Time series analysis can be performed on transformer data to test whether the data shows a consistent time trend and periodicity. Time series graphs and autocorrelation graphs can be drawn to observe the characteristics of the data.

For example, there is a set of transformer data, including variables such as input current, output current, and temperature. First, statistical analysis is performed on these data to calculate statistics such as mean, median, and standard deviation. Then, variance analysis is used to test whether there are significant differences in the means of these variables to determine whether there is consistency between them. Then, correlation analysis is used to calculate the correlation coefficients between the variables to determine whether the relationship between them is consistent. Finally, time series analysis is used to test whether the data has consistent time trends and periodicity.

The results obtained by assuming that the correlation coefficient between the input current and the output current is 0.95, indicating that there is a high correlation between them; and the mean of the temperature variable is not significantly different in the variance analysis, indicating that the data consistency of the temperature variable is good. The time series analysis shows that the output current shows a gradual increase trend, indicating that the transformer output current may have a consistent time variation law. Through the above analysis, it can be concluded that there is a high consistency between the input current and the output current of the transformer, and the data of the temperature variable is also relatively consistent.

Specifically, the data of the transformer can be recollected to ensure the accuracy and completeness of the data. Data can be obtained by real-time monitoring through sensors or monitoring equipment, or by regular inspection and testing of the transformer. For missing data, it can be filled by interpolation or mean method. The interpolation method estimates the value of the missing data point based on the relationship between known data points, such as linear interpolation, polynomial interpolation, etc.; the mean method replaces the missing data with the average value of the known data. By performing statistical analysis and outlier detection on the data, abnormal data points are found and eliminated. Common outlier detection methods include box plots, Z-score methods, etc. Eliminating abnormal data can improve the accuracy and reliability of the data.

Specifically, as shown in FIG. 4 , the main process of the above embodiment includes data collection, consistency detection, data repair, repair execution, abnormality alarm and log recording.

Specifically, the sensor data in the database can be regularly checked to compare whether the data between different sensors are consistent, including whether the data values are within a reasonable range and whether there are abnormal data. When data inconsistency is detected, it will be processed according to a pre-set data repair strategy. The data repair strategy can be generated through the following steps: data analysis, anomaly detection, strategy generation, strategy evaluation, strategy optimization, strategy application, result monitoring, and feedback mechanism.

Data analysis: collect historical data and perform data analysis, including statistical analysis, anomaly detection, etc.;

Anomaly detection: Detecting outliers or inconsistencies in sensor data using data mining or machine learning algorithms;

Strategy generation: Automatically generate data repair strategies based on anomaly detection results and data analysis; strategy generation can be based on rule engines, machine learning models, or expert systems;

Strategy evaluation: Evaluate the generated data repair strategy, including accuracy, efficiency, and feasibility, to ensure that the strategy can effectively repair data inconsistency issues;

Strategy optimization: If necessary, optimize and adjust the generated strategy to improve the repair effect and performance;

Strategy application: Apply the generated data repair strategy to the actual data to repair inconsistent data;

Result monitoring: monitor the repaired data to ensure data consistency and accuracy;

Feedback mechanism: Feedback to the strategy generation module based on the repair results to optimize and improve the strategy generation process.

Through the above process, data repair strategies can be automatically generated to improve the efficiency and accuracy of data repair. Repair strategies suitable for the current situation can be dynamically generated based on historical data and real-time data, reducing manual intervention and improving the adaptive ability of the solution. At the same time, through the evaluation and optimization of strategies, the effect of data repair can be continuously improved to ensure the consistency and quality of data.

In the specific implementation process, after obtaining the operating information of the transformer, the method also includes the following steps: constructing a first detection model, wherein the first detection model is obtained by training with a Bayesian algorithm using multiple sets of training data, and each set of training data in the multiple sets of training data includes historical operating information obtained within a historical time period, and historical second analysis results corresponding to the historical operating information, wherein the historical second analysis results are used to characterize whether the historical transformer is faulty; inputting the operating information into the first detection model to obtain a second analysis result corresponding to the operating information.

In this scheme, the Bayesian network model, i.e., the first detection model, can be trained through the training data set. The first detection model is used to analyze whether the transformer is faulty. The Bayesian algorithm can calculate the probability of whether the transformer is faulty based on the existing data and prior knowledge, thereby quickly and accurately judging the working status of the transformer. The Bayesian algorithm can effectively handle uncertainty and noise, further improving the accuracy and reliability of fault diagnosis.

Specifically, as shown in FIG5 , the main process of the above embodiment includes uncertainty modeling, fault diagnosis, uncertainty processing, result interpretation, feedback optimization, and also includes additional node addition, node definition, and data integration.

Specifically, first, real-time data is collected from sensors and other devices and stored in a database. Then, the collected data is cleaned, denoised and feature extracted to prepare for fault diagnosis analysis. Then, multiple fault diagnosis algorithms including rule engines, machine learning models and expert systems are integrated to form an integrated fault detection model (first detection model). At this time, probabilistic inference technology is introduced, and the diagnosis results generated by the algorithm are modeled using Bayesian networks for uncertainty. According to data features and algorithm fusion results, fault diagnosis analysis is performed and possible fault diagnosis results are generated. The uncertainty of the diagnosis results is processed, including calculating confidence or probability distribution to improve the reliability of the diagnosis results. It should be particularly noted that when using Bayesian networks for uncertainty modeling, temperature, humidity and wind speed are additionally added as inputs to the model. Then, based on domain knowledge and data analysis, the relationship between geographical environment and weather factors and transformer status is defined. Specifically, in a high temperature environment, the temperature inside the transformer may rise, resulting in a decrease in electrical insulation performance, thereby increasing the risk of transformer failure. Therefore, a relationship node representing the impact of temperature on the transformer status is established. A high humidity environment may cause the transformer insulation material to become damp, increasing the possibility of leakage and short circuit. Therefore, humidity can be associated with the transformer status node to reflect the impact of humidity on transformer performance. In the case of high wind speed, the transformer may be affected by the external environment, the stress caused by wind blowing may increase, or the insulation performance may decrease due to wind and rain erosion. Therefore, wind speed can be used as a node and associated with the transformer status node to describe the impact of wind speed on the transformer. At this time, when additional nodes are introduced to represent geographical environment and weather factors, the relationship between nodes can be described by the conditional probability distribution matrix. Specifically, the following nodes are set:

Transformer status node: T (normal, fault);

Current node: I (low (less than the first current threshold), normal (greater than or equal to the first current threshold and less than the second current threshold), high (greater than the second current threshold), the first current threshold is less than the second current threshold);

Temperature node: Temp (low (less than the first temperature threshold), normal (greater than or equal to the first temperature threshold and less than the second temperature threshold), high (greater than the second temperature threshold), the first temperature threshold is less than the second temperature threshold);

Humidity node: Humidity (low (less than the first humidity threshold), normal (greater than or equal to the first humidity threshold and less than the second humidity threshold), high (greater than the second humidity threshold), the first humidity threshold is less than the second humidity threshold);

Wind speed node: Wind (low (less than the first wind speed threshold), normal (greater than or equal to the first wind speed threshold and less than the second wind speed threshold), high (greater than the second wind speed threshold), the first wind speed threshold is less than the second wind speed threshold);

Finally, the matrix P(T|I, Temp, Humidity, Wind) (conditional probability distribution of transformer status nodes with respect to other nodes) contains all possible combinations, such as the probability that the transformer status is normal when the current is low, the temperature is high, the humidity is normal, and the wind speed is low. Such a matrix can help the model to infer and predict, and it takes multiple factors into account more comprehensively.

Then, the geographic environment and weather data are integrated into the training data of the Bayesian network model, and then the model is trained. By learning the patterns and correlations in the data, the model can more accurately infer and predict the status of the transformer, and then feedback the diagnosis results to the user or operator according to the region, and explain the credibility of the diagnosis results and the recommended treatment measures. Finally, according to the actual situation and user feedback, the algorithm fusion and uncertainty modeling process are optimized to improve the overall fault diagnosis accuracy and reliability of the scheme. Finally, the temperature, humidity and wind speed factors of the geographical location are used as additional nodes and associated with the transformer status node. By analyzing the relationship between these factors and transformer faults, the Bayesian network model can more accurately predict the operating status and potential faults of the transformer.

Specifically, integrating geographic environment and weather data into the training data of the Bayesian network model and then training the model requires the following steps: data preparation, building a Bayesian network model, integrating data, and model training.

Data preparation: Prepare geographic environment and weather data, as well as sensor data. Ensure that the data format is consistent and perform necessary preprocessing, such as missing value processing and data normalization;

Building a Bayesian network model: Use appropriate libraries or tools to build a Bayesian network model, specifically using pgmpy;

Integrate data: Integrate geographic environment and weather data into sensor data to form a complete training data set;

Model training: Use the integrated training data set to train the Bayesian network model. This includes inputting data into the model and calling the corresponding training function to fit the model.

The specific model training code is as follows:

from pgmpy.models import BayesianModel;

from pgmpy.estimators import ParameterEstimator;

from pgmpy.estimators import MaximumLikelihoodEstimator;

import pandas as pd;

#Assume that sensor data, geographic environment and weather data have been prepared and stored in DataFrame;

#Assume that sensor data is stored in sensor_data, geographic environment data is stored in geo_data, and weather data is stored in weather_data;

#Merge data;

merged_data=pd.concat([sensor_data,geo_data,weather_data],axis=1);

#Build a Bayesian network model;

model=BayesianModel([('sensor1','sensor2'),('geo_feature','sensor1'),('weather_feature','sensor1')]);

#Use maximum likelihood estimation to estimate parameters;

estimator=ParameterEstimator(model,merged_data);

estimator.get_parameters();

#Use maximum likelihood estimation to estimate parameters;

model.fit(merged_data,estimator=MaximumLikelihoodEstimator);

Through the above process, the uncertainty of the fault diagnosis algorithm can be effectively handled, and the reliability and interpretability of the fault diagnosis results can be improved. The combination of algorithm fusion and probabilistic inference can make full use of the advantages of different algorithms and effectively manage the uncertainty of the diagnosis results, thereby improving the fault diagnosis capability of this solution in complex environments.

As mentioned above, since the data between different sensors may be inconsistent, including different units and sampling frequencies, this will affect the accuracy of information fusion, and with the increase in the scale of the transformer detection system and the introduction of new sensors, the scalability of the transformer detection system is also a challenge, and the performance and resource requirements of the transformer detection system need to be considered. At the same time, the fault diagnosis algorithm within the transformer detection system may have misjudgments and missed judgments, especially in complex fault conditions. The accuracy of the algorithm may be affected. In the process of information fusion, the Bayesian network model is often used for transformer fault detection. However, after deep learning through a fixed Bayesian network model, when it is applied to substations in different regions, different erroneous judgments will be generated due to the influence of the geographical environment and external weather, which may cause limitations in predicting faults. Therefore, deep learning training can be performed through the fusion of multiple algorithms. The following embodiments are introduced in detail.

In the specific implementation process, after obtaining the operating information of the transformer, the method also includes the following steps: constructing a second detection model, wherein the second detection model is trained by a Bayesian algorithm and an auxiliary algorithm using multiple sets of training data, and each set of training data in the multiple sets of training data includes historical operating information obtained within a historical time period, and a historical third analysis result corresponding to the historical operating information, wherein the historical third analysis result is used to characterize whether the historical transformer is faulty, and the auxiliary algorithm includes at least one or more of the LSTM algorithm, LightGBM algorithm, TPE algorithm, BOA-LSTM algorithm, and BART algorithm; inputting the operating information into the second detection model to obtain the third analysis result corresponding to the operating information.

In this solution, the fusion of multiple algorithms can make up for the shortcomings of a single algorithm, thereby reducing the misjudgment rate and further accurately determining whether the transformer is faulty.

Specifically, the second detection model can be constructed by using the Bayesian algorithm and the LSTM algorithm to determine whether the transformer is faulty. The extracted features are analyzed using the Bayesian algorithm to calculate the probability of transformer failure. The Bayesian algorithm can effectively handle uncertainty and missing data. The time series data is input into the LSTM neural network for training to predict the future state of the transformer. The LSTM algorithm can capture long-term dependencies in time series data. The results of the Bayesian algorithm and the LSTM algorithm are comprehensively analyzed to determine whether the transformer is faulty. The Bayesian algorithm can effectively handle uncertainty and missing data, improving the robustness of the solution. The LSTM algorithm can capture long-term dependencies in time series data, improving the accuracy and stability of the prediction. Combining the two algorithms can comprehensively consider static and dynamic features and comprehensively analyze the health status of the transformer.

Specifically, the second detection model can be constructed by using the Bayesian algorithm and the LightGBM algorithm to determine whether the transformer is faulty. The data set is divided into a training set and a test set. The training set is trained using the Bayesian algorithm to obtain a model. The training set is trained using the LightGBM algorithm to obtain a model. The model is evaluated using the test set to compare the accuracy, recall and other indicators of the two algorithms. Determine whether the transformer is faulty based on the model prediction results. The LightGBM algorithm is an efficient gradient boosting algorithm that can handle large-scale data sets and has a high accuracy rate. The combination of the two algorithms can improve the accuracy and robustness of the model, making fault judgment more reliable.

Specifically, the second detection model can be constructed by the Bayesian algorithm and the TPE algorithm to determine whether the transformer is faulty. An initial probability model is established using the Bayesian algorithm, which can represent the probability distribution of whether the transformer is faulty. The TPE algorithm is used to optimize the probability model to find the parameter combination that is most likely to cause the transformer to fail. Based on the optimized parameter combination, the Bayesian algorithm is used again to update the probability model to more accurately represent the probability distribution of whether the transformer is faulty. Based on the updated probability model, a more accurate prediction result can be obtained to determine whether the transformer is faulty. The TPE algorithm can quickly and effectively find the parameter combination that is most likely to cause a transformer failure, thereby reducing computing time and resource consumption. The combination of the Bayesian algorithm and the TPE algorithm can better adapt to the potential laws and characteristics of transformer failures in the process of continuously updating the probability model, thereby improving the accuracy and reliability of the prediction.

Specifically, the second detection model can be constructed by using the Bayesian algorithm and the BOA-LSTM algorithm to determine whether the transformer is faulty. Use the Bayesian algorithm for fault prediction: Use the Bayesian algorithm to analyze and model the transformer operation data to obtain the probability of whether the transformer is faulty. Input the preprocessed data into the BOA-LSTM algorithm for training to obtain the prediction result of whether the transformer is faulty. Combine the fault prediction results obtained by the Bayesian algorithm and the BOA-LSTM algorithm to comprehensively analyze whether the transformer is faulty. The Bayesian algorithm can provide the results of probability analysis, and the BOA-LSTM algorithm can provide the prediction results of time series data. The combination of the two can more comprehensively evaluate the fault condition of the transformer. Combining the two algorithms can effectively reduce the false alarm rate and the missed alarm rate, and improve the effect and efficiency of fault prediction.

Specifically, the second detection model can be constructed by using the Bayesian algorithm and the BART algorithm to determine whether the transformer is faulty. The Bayesian algorithm is used to analyze the data and calculate the posterior probability distribution of each parameter of the transformer. The obtained posterior probability distribution is used as the input of the BART algorithm, and a prediction model is established using the BART algorithm to predict whether the transformer is faulty. According to the prediction results of the BART algorithm, the transformer is maintained or repaired. The BART algorithm can establish a prediction model through a non-parametric method, which improves the accuracy and stability of the prediction. Combining the two algorithms can more comprehensively analyze the status of the transformer and improve the efficiency and accuracy of fault diagnosis.

Specifically, this solution can effectively identify abnormal values or inconsistencies in sensor data by collecting and analyzing historical data and combining data mining and machine learning algorithms. Furthermore, by combining rule engines, machine learning models and expert systems, corresponding repair strategies can be automatically generated for various data repair needs and situations. This automated process not only greatly improves the efficiency and accuracy of data repair, but also generates the best repair strategy for the current situation based on historical and real-time data through dynamic adjustment of strategy generation, thereby significantly reducing manual intervention and enhancing adaptive capabilities. At the same time, by continuously evaluating and optimizing strategies, the data repair effect can be continuously optimized to ensure data consistency and quality. In terms of fault diagnosis algorithms, by considering geographical environment and weather factors, the Bayesian network model with geographical location and weather condition nodes can better adapt to transformer systems in different regions. This customized model can more accurately reflect the environmental characteristics of a specific region, improve the applicability and generalization of the model, and combine the prediction model and repair strategy generation to more intelligently adjust the repair strategy to adapt to future data changes. In this way, it is possible to respond to different situations more flexibly and further improve the adaptability and generalization of the solution.

In some embodiments, before obtaining the operating information of the transformer, the method further includes the following steps: obtaining interface definition information, wherein the interface definition information is pre-defined information on the data transmission format, transmission protocol and communication method of the interface of the sensor; and configuring the interface of the sensor according to the interface definition information.

In this solution, standardized interface specifications can be pre-defined to ensure that communication and data interaction can be achieved through standard interfaces, which greatly promotes the integration of third-party modules or third-party devices and further enhances the scalability of this solution.

Specifically, by obtaining the information of the data transmission format, transmission protocol and communication mode of the predefined sensor interface, and then configuring the sensor interface according to this information, the advantage of doing so is that it can ensure that the data transmission between the sensor and other devices (such as transformers) or systems is effective and accurate, and it can also improve the stability and reliability of the system.

For example, suppose there is a temperature sensor. According to the interface definition information provided by the manufacturer, the sensor uses the Modbus protocol for communication, the data transmission format is a 16-bit integer, and the communication method is RS485. Based on this information, the interface parameters of the sensor can be configured to ensure normal communication with other devices or systems and avoid data transmission errors or losses. By properly configuring the sensor interface, the overall efficiency and performance of the system can be improved, system failures caused by data transmission problems can be reduced, and the accuracy and reliability of data can be ensured.

Specifically, in the above specific embodiment, as shown in FIG. 6 , the specific process of interface standardization includes the following steps: interface definition, interface management, interface registration, interface verification, interface communication, and interface update.

Interface definition: Determine the interfaces required for data exchange and communication between various sensors and systems. This may include defining the data format, transmission protocol, communication method, and other specifications of the interface;

Interface management: Interface standardization management is responsible for the definition and specification of all interfaces. You can maintain an interface list, including interface name, description, data format and other information;

Interface registration: When a new module or a new sensor is added to the overall system, the interface it provides needs to be registered. The registration includes information such as interface name, function description, data format, etc.

Interface verification: Verify the registered interface to ensure compliance with the specification. Verification includes whether the data format is correct and whether the data transmission is reliable;

Interface communication: When you need to communicate with other sensors or devices, query the specifications of the corresponding interface. Perform data exchange and communication according to the specifications to ensure the correct transmission and analysis of data;

Interface update: As the overall solution evolves and requirements change, the interface specifications may need to be updated. Update the interface specifications in a timely manner and notify the corresponding devices, modules, or sensors to make corresponding adjustments.

Through the above process, interface standardization can ensure the consistency and reliability of interface specifications between various modules and sensors in the system, and improve maintainability and scalability. At the same time, through unified interface specifications, the integration and interaction between different devices and modules become simpler and more efficient.

In some embodiments, before obtaining the operating information of the transformer, the method further includes the following steps: obtaining configuration definition information, wherein the configuration definition information is pre-defined information of a sampling frequency, resolution, abnormal alarm threshold and power mode of the sensor; and configuring the sensor according to the configuration definition information.

In this solution, standardized sensor specifications can be predefined to ensure that the sensors can be automatically configured through standardized settings, thereby further enhancing the scalability of the solution and further improving the performance and stability of the sensors.

Specifically, obtain the configuration information of the sensor, including pre-defined parameters such as sampling frequency, resolution, abnormal alarm threshold and power mode; then configure the sensor accordingly according to these configuration information. In this way, the working parameters of the sensor can be customized according to actual needs and environmental conditions to ensure that the sensor can work normally and provide accurate data. For example, if you need to monitor temperature changes, you can set the sampling frequency of the temperature sensor to collect data once a minute, the resolution to 0.1 degrees Celsius, the abnormal alarm threshold to alarm when it exceeds 40 degrees Celsius, and the power mode to select low power mode. By configuring the sensor, you can improve the performance and stability of the sensor, and you can also effectively manage power consumption and extend the service life of the sensor.

Specifically, in the above specific embodiment, the configuration information can be managed through dynamic configuration, and the user or administrator is allowed to dynamically modify the configuration as needed. As shown in Figure 7, the following is the specific process of dynamic configuration: configuration management, configuration query, configuration modification, configuration verification, configuration validation, and configuration rollback.

Configuration management: Maintain a configuration list, including configuration items and their default values for each module and sensor. Each configuration item has a unique identifier and description to identify and explain the function of the configuration item;

Configuration query: Users or administrators can query configuration items and their current values. Provide an interface or screen for users to view and manage configuration information;

Configuration modification: When users modify configuration items by changing their values, the configuration of related modules will be automatically updated according to the new values;

Configuration verification: When users modify configuration items, the new configuration values will be verified. Verification includes whether the data format is correct and whether the value range is legal.

Configuration takes effect: After the user modifies the configuration, the new configuration information will be applied to the relevant modules and sensors. Ensure that the new configuration takes effect immediately without restarting the main control device;

Configuration rollback: If a user finds a problem after modifying the configuration, the configuration item can be rolled back to the value before the modification. This ensures that the configuration modification process can be safe and reliable.

Through the above process, flexible management and dynamic adjustment of configuration can be achieved, which improves the adaptability and maintainability of the solution. Users can modify the configuration at any time as needed without shutting down or restarting the system, thus improving the flexibility and customizability of the solution.

Specifically, this solution also introduces a plug-in mechanism, which allows users or developers to write custom plug-ins according to actual needs and integrate them seamlessly into the solution, providing users with great functional expansion space.

Specifically, in the above specific embodiment, the plug-in mechanism can be responsible for loading, managing and calling the plug-in. As shown in Figure 8, the following are the steps of the process of introducing the plug-in mechanism: plug-in registration, plug-in loading, plug-in calling, plug-in unloading, plug-in updating, and plug-in communication.

Plugin registration: Developers write plugins and register them with the system including sensors. The registration includes plugin name, description, version number, author, etc.

Plugin loading: When the system starts, all registered plugins are loaded. The loading process includes checking plugin dependencies, initializing the plugin environment, and other operations.

Plugin call: You can call loaded plugins. Plugins provide some interfaces or functions for calling and extending functions.

Plugin uninstallation: When a plugin is no longer needed, you can uninstall it. The uninstallation process includes releasing the resources occupied by the plugin and cleaning up the plugin-related data.

Plugin update: It supports plugin update operations. When the plugin author releases a new version, the plugin can be updated.

Plugin communication: Plugins can communicate and collaborate with each other. Some communication mechanisms are provided for data exchange and collaborative work between plugins.

Through the above process, the plug-in mechanism can realize dynamic expansion and customization of functions, improving the flexibility and scalability of the solution. Developers can write plug-ins as needed without modifying the core code, thereby reducing the difficulty of solution maintenance and upgrades. At the same time, the plug-in mechanism also promotes decoupling and collaboration between modules, making the solution more modular and combinable.

In summary, the design of this solution aims to achieve flexible expansion and high customization to meet the actual needs of different users, thereby ensuring the high scalability and strong adaptability of the solution.

In some embodiments, after obtaining the first analysis result based on the size relationship between the above-mentioned comprehensive calculation result and the preset operating threshold, the above-mentioned method also includes the following steps: when the above-mentioned first analysis result characterizes the above-mentioned transformer fault, generating early warning information; based on the above-mentioned early warning information, controlling the early warning device to turn on, wherein the above-mentioned early warning device at least includes a buzzer and/or an LED light.

In this solution, if the transformer fails, the early warning device can be controlled to turn on, so that the operating personnel can be promptly reminded of the transformer failure, so that the operating personnel can understand the transformer failure in time.

In summary, the scalability of this solution is good, and the fault diagnosis algorithm has fewer misjudgments and missed judgments.

In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the multi-sensor based transformer fault determination method of the present application will be described in detail below in combination with specific embodiments.

This embodiment relates to a specific transformer fault determination method based on multiple sensors, comprising the following steps:

Step S1: data collection and processing;

First, data is collected from different sensors in real time and passed to the information fusion module. While receiving data, the data between different sensors are compared for consistency. At the same time, the data values are checked to see if they are within a reasonable range and whether there are abnormal data. When inconsistent sensor data is detected, the data consistency check module will process it according to the pre-set data repair strategy, including taking the average, taking the median, and removing abnormal values. When the system needs to be expanded, the interfaces for data exchange and communication between the modules in the system are first determined, and then the data format, transmission protocol and communication method of the interface are defined.

Step S2: information fusion;

Then, the data from different sensors are received and the comprehensive information S is calculated by weighted average. The specific calculation formula is S = α*T + β*H + γ*G, where α, β and γ are weight coefficients used to adjust the importance of different sensors in fusion;

Step S3: fault diagnosis;

Fault diagnosis is performed based on the fused information S to determine whether the system has a fault. When S exceeds the preset threshold, it is determined that the system has a fault, an alarm is triggered and maintenance suggestions are provided. When S does not exceed the threshold, the system operates normally and no operation is performed. In the fault diagnosis process, multiple fault diagnosis algorithms including rule engines, machine learning models and expert systems are integrated to form an integrated fault diagnosis system. Then, the Bayesian network is used to model the uncertainty of the diagnosis results generated by the algorithm, and then the uncertainty of the diagnosis results is processed.

The embodiment of the present application also provides a fault determination device for a transformer based on multiple sensors. It should be noted that the fault determination device for a transformer based on multiple sensors in the embodiment of the present application can be used to execute the fault determination method for a transformer based on multiple sensors provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and those that have been described will not be repeated. As used below, the term "module" can implement a combination of software and/or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

The following introduces a transformer fault determination device based on multiple sensors provided in an embodiment of the present application.

FIG9 is a structural block diagram of a transformer fault determination device based on multiple sensors according to an embodiment of the present application. As shown in FIG9 , the device includes:

A first acquisition unit 10 is used to acquire operation information of the transformer, wherein the operation information is multiple and is acquired by using a sensor. There are multiple sensors, and the operation information corresponds to the sensors one by one.

A second acquisition unit 20 is used to acquire a weight coefficient corresponding to the operation information, wherein the weight coefficient corresponds to the operation information in a one-to-one manner;

The calculation unit 30 is used to calculate the weighted average of the weight coefficient and the operation information to obtain a comprehensive calculation result;

The determination unit 40 is used to obtain a first analysis result based on the size relationship between the above-mentioned comprehensive calculation result and the preset operation threshold, wherein the above-mentioned first analysis result is used to characterize whether the above-mentioned transformer is faulty. When the above-mentioned comprehensive calculation result is greater than or equal to the above-mentioned preset operation threshold, the above-mentioned first analysis result characterizes that the above-mentioned transformer is normal; when the above-mentioned comprehensive calculation result is less than the above-mentioned preset operation threshold, the above-mentioned first analysis result characterizes that the above-mentioned transformer is faulty.

Through this embodiment, information detected by different sensors can be collected at the same time to obtain multiple operating information, which can be comprehensively analyzed to evaluate whether the transformer is faulty. Compared with the single sensor evaluation method, it has higher accuracy and can improve the accuracy of transformer fault detection.

In order to further improve the validity and accuracy of the data, the above-mentioned device also includes a third acquisition unit, a verification unit and a repair unit. The third acquisition unit is used to obtain initial operating information before obtaining the operating information of the transformer; the verification unit is used to perform data consistency verification on the above-mentioned initial operating information to obtain a verification result, wherein the verification method at least includes one or more of numerical analysis, variance analysis, correlation analysis, and time series analysis; the repair unit is used to perform data repair on the above-mentioned initial operating information to obtain the above-mentioned operating information when the above-mentioned verification result indicates that the data is inconsistent, wherein the data repair method at least includes one or more of re-collecting data, filling missing data, and eliminating abnormal data.

In this solution, the data can be checked for consistency to determine whether there is any abnormal data. If the data is inconsistent, that is, if there is abnormal data, the data can be repaired, thereby further improving the validity and accuracy of the data.

During the specific implementation process, the above-mentioned device also includes a first construction unit and a first processing unit. The first construction unit is used to construct a first detection model after obtaining the operating information of the transformer, wherein the above-mentioned first detection model is obtained by training through a Bayesian algorithm using multiple sets of training data, and each set of training data in the above-mentioned multiple sets of training data includes historical operating information obtained within a historical time period and historical second analysis results corresponding to the above-mentioned historical operating information, wherein the above-mentioned historical second analysis results are used to characterize whether the historical transformer is faulty; the first processing unit is used to input the above-mentioned operating information into the above-mentioned first detection model to obtain the second analysis result corresponding to the above-mentioned operating information.

In this scheme, the Bayesian network model, i.e., the first detection model, can be trained through the training data set. The first detection model is used to analyze whether the transformer is faulty. The Bayesian algorithm can calculate the probability of whether the transformer is faulty based on the existing data and prior knowledge, thereby quickly and accurately judging the working status of the transformer. The Bayesian algorithm can effectively handle uncertainty and noise, further improving the accuracy and reliability of fault diagnosis.

During the specific implementation process, the above-mentioned device also includes a second construction unit and a second processing unit. The second construction unit is used to construct a second detection model after obtaining the operating information of the transformer, wherein the above-mentioned second detection model is obtained by training through a Bayesian algorithm and an auxiliary algorithm using multiple sets of training data, and each set of training data in the above-mentioned multiple sets of training data includes historical operating information obtained within a historical time period, and a historical third analysis result corresponding to the above-mentioned historical operating information, wherein the above-mentioned historical third analysis result is used to characterize whether the historical transformer is faulty, and the above-mentioned auxiliary algorithm includes at least one or more of the LSTM algorithm, LightGBM algorithm, TPE algorithm, BOA-LSTM algorithm, and BART algorithm; the second processing unit is used to input the above-mentioned operating information into the above-mentioned second detection model to obtain the third analysis result corresponding to the above-mentioned operating information.

In this solution, the fusion of multiple algorithms can make up for the shortcomings of a single algorithm, thereby reducing the misjudgment rate and further accurately determining whether the transformer is faulty.

In some embodiments, the above-mentioned device also includes a fourth acquisition unit and a first configuration unit, the fourth acquisition unit is used to obtain interface definition information before obtaining the operation information of the transformer, wherein the above-mentioned interface definition information is pre-defined information on the data transmission format, transmission protocol and communication method of the interface of the above-mentioned sensor; the first configuration unit is used to configure the interface of the above-mentioned sensor according to the above-mentioned interface definition information.

In this solution, standardized interface specifications can be pre-defined to ensure that communication and data interaction can be achieved through standard interfaces, which greatly promotes the integration of third-party modules or third-party devices and further enhances the scalability of this solution.

In some embodiments, the above-mentioned device also includes a fifth acquisition unit and a second configuration unit, the fifth acquisition unit is used to obtain configuration definition information before obtaining the operation information of the transformer, wherein the above-mentioned configuration definition information is pre-defined information of the sampling frequency, resolution, abnormal alarm threshold and power mode of the above-mentioned sensor; the second configuration unit is used to configure the above-mentioned sensor according to the above-mentioned configuration definition information.

In this solution, standardized sensor specifications can be predefined to ensure that the sensors can be automatically configured through standardized settings, thereby further enhancing the scalability of the solution and further improving the performance and stability of the sensors.

In some embodiments, the above-mentioned device also includes a generation unit and a control unit. The generation unit is used to generate early warning information after obtaining a first analysis result based on the size relationship between the above-mentioned comprehensive calculation result and a preset operating threshold, when the above-mentioned first analysis result characterizes the above-mentioned transformer fault; the control unit is used to control the early warning device to start based on the above-mentioned early warning information, wherein the above-mentioned early warning device at least includes a buzzer and/or an LED light.

In this solution, if the transformer fails, the early warning device can be controlled to turn on, so that the operating personnel can be promptly reminded of the transformer failure, so that the operating personnel can understand the transformer failure in time.

The above-mentioned transformer fault determination device based on multiple sensors includes a processor and a memory, and the above-mentioned first acquisition unit, second acquisition unit, calculation unit and determination unit are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are respectively located in different processors in the form of any combination.

The processor includes a kernel, and the kernel retrieves the corresponding program unit from the memory. One or more kernels can be provided, and the problem of low accuracy of transformer fault detection in the prior art can be solved by adjusting kernel parameters.

The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the multi-sensor based transformer fault determination method.

An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein the multi-sensor-based transformer fault determination method is executed when the program is run.

An embodiment of the present invention provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, at least steps of a transformer fault determination method based on multiple sensors are implemented.

The devices in this article can be servers, PCs, PADs, mobile phones, etc.

A computer program product includes a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the multi-sensor-based transformer fault determination method in each embodiment of the present application are implemented.

The present application also provides a transformer fault detection system, comprising one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing any one of the above-mentioned multi-sensor based transformer fault determination methods.

Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, and the combination of the process and/or box in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

In a typical configuration, a computing device includes one or more processors (CPU), input/output interfaces, network interfaces, and memory.

The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and/or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

1) The multi-sensor transformer fault determination method of the present application can simultaneously collect information detected by different sensors, obtain multiple operating information, and comprehensively analyze the multiple operating information to evaluate whether the transformer is faulty. Compared with the single sensor evaluation method, it has higher accuracy and can improve the accuracy of transformer fault detection.

2) The multi-sensor based transformer fault determination device of the present application can simultaneously collect information detected by different sensors, obtain multiple operating information, and comprehensively analyze the multiple operating information to evaluate whether the transformer is faulty. Compared with the single sensor evaluation method, it has higher accuracy and can improve the accuracy of transformer fault detection.

The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims (10)

1. A transformer fault determination method based on multiple sensors, characterized by comprising: Acquire operation information of the transformer, wherein there are multiple pieces of operation information, the operation information is collected by using a sensor, there are multiple sensors, and the operation information corresponds to the sensors one by one; Obtaining a weight coefficient corresponding to the operation information, wherein the weight coefficient corresponds to the operation information one-to-one; Calculating a weighted average of the weight coefficient and the operation information to obtain a comprehensive calculation result; A first analysis result is obtained based on the size relationship between the comprehensive calculation result and the preset operating threshold, wherein the first analysis result is used to characterize whether the transformer is faulty. When the comprehensive calculation result is greater than or equal to the preset operating threshold, the first analysis result characterizes that the transformer is normal; when the comprehensive calculation result is less than the preset operating threshold, the first analysis result characterizes that the transformer is faulty. 2. The method according to claim 1, characterized in that, before obtaining the operation information of the transformer, the method further comprises: Get initial operation information; Performing a data consistency check on the initial operation information to obtain a check result, wherein the check method includes at least one or more of numerical analysis, variance analysis, correlation analysis, and time series analysis; In the case where the test result characterizes data inconsistency, data repair is performed on the initial operation information to obtain the operation information, wherein the data repair method includes at least one or more of re-collecting data, filling missing data, and eliminating abnormal data. 3. The method according to claim 1, characterized in that, after obtaining the operation information of the transformer, the method further comprises: Constructing a first detection model, wherein the first detection model is obtained by training with a Bayesian algorithm using multiple sets of training data, each set of training data in the multiple sets of training data includes historical operation information obtained within a historical time period and a historical second analysis result corresponding to the historical operation information, wherein the historical second analysis result is used to characterize whether a historical transformer is faulty; The operation information is input into the first detection model to obtain a second analysis result corresponding to the operation information. 4. The method according to claim 1, characterized in that, after obtaining the operation information of the transformer, the method further comprises: Construct a second detection model, wherein the second detection model is obtained by training with a Bayesian algorithm and an auxiliary algorithm using multiple sets of training data, each set of training data in the multiple sets of training data includes historical operation information obtained in a historical time period, and a historical third analysis result corresponding to the historical operation information, wherein the historical third analysis result is used to characterize whether the historical transformer is faulty, and the auxiliary algorithm includes at least one or more of an LSTM algorithm, a LightGBM algorithm, a TPE algorithm, a BOA-LSTM algorithm, and a BART algorithm; The operation information is input into the second detection model to obtain a third analysis result corresponding to the operation information. 5. The method according to any one of claims 1 to 4, characterized in that, before obtaining the operation information of the transformer, the method further comprises: Acquire interface definition information, wherein the interface definition information is predefined information of a data transmission format, a transmission protocol, and a communication method of the interface of the sensor; The interface of the sensor is configured according to the interface definition information. 6. The method according to any one of claims 1 to 4, characterized in that, before obtaining the operation information of the transformer, the method further comprises: Acquiring configuration definition information, wherein the configuration definition information is predefined information of a sampling frequency, a resolution, an abnormal alarm threshold, and a power mode of the sensor; The sensor is configured according to the configuration definition information. 7. The method according to any one of claims 1 to 4, characterized in that after obtaining a first analysis result according to the magnitude relationship between the comprehensive calculation result and a preset operation threshold, the method further comprises: In the case where the first analysis result indicates that the transformer is faulty, generating warning information; Based on the warning information, the warning device is controlled to turn on, wherein the warning device at least includes a buzzer and/or an LED light. 8. A transformer fault determination device based on multiple sensors, characterized by comprising: A first acquisition unit is used to acquire operation information of the transformer, wherein the operation information has multiple pieces, the operation information is collected by using a sensor, the sensor has multiple pieces, and the operation information corresponds to the sensor one by one; A second acquisition unit, configured to acquire a weight coefficient corresponding to the operation information, wherein the weight coefficient corresponds to the operation information in a one-to-one manner; A calculation unit, used to calculate the weighted average of the weight coefficient and the operation information to obtain a comprehensive calculation result; A determination unit is used to obtain a first analysis result based on the size relationship between the comprehensive calculation result and a preset operating threshold, wherein the first analysis result is used to characterize whether the transformer is faulty. When the comprehensive calculation result is greater than or equal to the preset operating threshold, the first analysis result characterizes that the transformer is normal; when the comprehensive calculation result is less than the preset operating threshold, the first analysis result characterizes that the transformer is faulty. 9. A computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, the steps of the transformer fault determination method based on multiple sensors as claimed in any one of claims 1 to 7 are implemented. 10. A transformer fault detection system, characterized in that it comprises: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing the multi-sensor based transformer fault determination method according to any one of claims 1 to 7.