CN119203006A - A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft - Google Patents
A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft Download PDFInfo
- Publication number
- CN119203006A CN119203006A CN202411699558.3A CN202411699558A CN119203006A CN 119203006 A CN119203006 A CN 119203006A CN 202411699558 A CN202411699558 A CN 202411699558A CN 119203006 A CN119203006 A CN 119203006A
- Authority
- CN
- China
- Prior art keywords
- flight
- value
- motor
- time
- vertical take
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
A method for diagnosing the abnormal flight state of an electric vertical take-off and landing aircraft belongs to the technical field of the flight management of the electric vertical take-off and landing aircraft. The method comprises the steps of constructing flight monitoring indexes of an electric vertical take-off and landing aircraft, constructing influence factors of the flight monitoring indexes of the electric vertical take-off and landing aircraft, including environment temperature, wind speed, wind direction, load, flight time, battery voltage and battery current, constructing a flight abnormal state diagnosis data set of the electric vertical take-off and landing aircraft, including a training set, a verification set and a test set, constructing a flight monitoring index prediction model of the electric vertical take-off and landing aircraft based on an improved LSTM neural network, training the flight monitoring index prediction model of the electric vertical take-off and landing aircraft by using the training set to obtain a trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft, and diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft based on the prediction data.
Description
Technical Field
The invention belongs to the technical field of flight management of electric vertical take-off and landing aircrafts, and particularly relates to a flight abnormality diagnosis method of an electric vertical take-off and landing aircrafts.
Background
EVTOL aircraft (electric vertical takeoff and landing aircraft) have a number of significant features and advantages. First, they are able to take off and land vertically in very small take off and land spaces, enabling operation in cities and crowded areas, greatly saving the space required for conventional aircraft that require long runways. And secondly, as the electric propulsion system is adopted, the eVTOL aircraft is more environment-friendly than the traditional internal combustion engine aircraft, has lower emission and greatly reduces noise, and is beneficial to reducing the influence on the environment and improving the life quality of residents. In addition, eVTOL aircraft's maintenance cost is lower relatively, and flexible operation can respond to urban short-distance transportation demand fast. These characteristics make eVTOL an ideal choice for the fields of future urban air traffic, rapid rescue and the like.
During flight of eVTOL aircraft, monitoring battery charge, motor temperature, and rotational speed are critical. Monitoring of the battery charge ensures that the aircraft can safely return or land before the power is exhausted, and also helps to maintain the battery and extend its useful life. Real-time monitoring of motor temperature can prevent equipment damage or performance degradation caused by motor overheating while providing important data regarding motor health. The accurate monitoring of the rotating speed directly influences the lift force and the thrust force of the aircraft, and is important for flight control and energy consumption optimization. Through the continuous monitoring of the data, the safety and efficiency of the flight can be greatly improved.
EVTOL (electric vertical takeoff and landing aircraft) during flight, a series of serious problems are faced by lack of detection of abnormal states of battery power, motor temperature and rotation speed. First, the battery power is a basic energy source for eVTOL flight, and abnormal power may cause the aircraft to fail to complete a predetermined flight task, even power is suddenly cut off in the air, resulting in a flight safety accident. Secondly, the motor is a core component for driving eVTOL to fly, and abnormal temperature rise of the motor can indicate overload or fault occurrence, and if the motor is not detected and processed in time, the motor can be damaged, and fire can be caused when the motor is serious. Meanwhile, abnormal changes of the motor rotation speed are directly related to control stability and operation responsiveness of the aircraft, and too high or too low rotation speed can cause difficulty in flight control and influence flight safety. The lack of an effective anomaly detection method not only increases the safety risk in the flight process, but also may result in increased maintenance costs and reduced flight efficiency of the aircraft. In addition, this can also affect the confidence level of the passengers and operators, negatively impacting the business operations.
In the patent of the invention with the application number 201810790066.3 and the invention name of 'a method and a system for predicting the flight state of an unmanned aerial vehicle based on LSTM', the invention provides the method and the system for predicting the flight state based on LSTM aiming at the problem that the existing unmanned aerial vehicle control model can not predict the flight action according to the flight state data. By constructing an action tag dictionary, collecting and preprocessing data and training and verifying by using an improved LSTM model, the method can accurately predict the flight action of the unmanned aerial vehicle and improve the accuracy and safety of the unmanned aerial vehicle operation. However, the above-described invention is dedicated to motion prediction in normal operation situations, and has not been concerned with abnormal state detection or processing.
Disclosure of Invention
The invention aims to solve the problem of detecting and processing the abnormal state of an electric vertical take-off and landing aircraft and provides a method for diagnosing the abnormal flight state of the electric vertical take-off and landing aircraft.
In order to achieve the above purpose, the present invention is realized by the following technical scheme:
a method for diagnosing the abnormal flight state of an electric vertical take-off and landing aircraft comprises the following steps:
s1, constructing flight monitoring indexes of an electric vertical take-off and landing aircraft, wherein the flight monitoring indexes are battery electric quantity, motor temperature and rotating speed;
s2, constructing influence factors of flight monitoring indexes of the electric vertical take-off and landing aircraft, wherein the influence factors comprise environmental temperature, wind speed, wind direction, load, flight time, battery voltage and battery current;
S3, acquiring flight monitoring index data of the electric vertical take-off and landing aircraft and influence factor data of the flight monitoring index of the electric vertical take-off and landing aircraft at the same time, and constructing an abnormal flight state diagnosis data set of the electric vertical take-off and landing aircraft, wherein the diagnosis data set comprises a training set, a verification set and a test set;
s4, constructing an electric vertical take-off and landing aircraft flight monitoring index prediction model based on the improved LSTM neural network;
s5, training the electric vertical take-off and landing aircraft flight monitoring index prediction model by using the training set obtained in the step S3 to obtain a trained electric vertical take-off and landing aircraft flight monitoring index prediction model;
S6, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft, and diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft based on prediction data.
Further, in step S1, a coulometer equipped with the eVTOL aircraft is used to measure the battery power, an infrared sensor is used to measure the motor temperature, and an optical encoder is used to detect the rotation speed.
Further, the specific implementation method of the step S3 includes the following steps:
S3.1, acquiring flight monitoring index data of the electric vertical take-off and landing aircrafts at the same time and influence factor data of the flight monitoring indexes of the electric vertical take-off and landing aircrafts;
Collecting unmanned aerial vehicle in flight process The battery electric quantity corresponding to the moment is sequentially as followsThe corresponding motor temperatures are in turnThe corresponding motor rotation speeds are in turn;
Collecting unmanned aerial vehicle flight process by utilizing thermal infrared imagerThe ambient temperature corresponding to the time is in turn;
Adopt hot wire anemometer to gather unmanned aerial vehicle flight in-processThe wind speeds corresponding to the moments are in turn;
Adopt ultrasonic wave anemoscope to gather unmanned aerial vehicle flight in-processThe wind directions corresponding to the moments are in turn;
Adopt piezoelectric weighing sensor to gather unmanned aerial vehicle flight in-processThe load corresponding to the moment is in turn;
Adopt GPS module to gather unmanned aerial vehicle flight in-processThe flight time corresponding to the moment is in turn;
Adopt voltage sensor to gather unmanned aerial vehicle flight in-processThe voltages of the batteries corresponding to the moments are sequentially as follows;
Adopt current sensor to gather unmanned aerial vehicle flight in-processThe current of the batteries corresponding to the moments is sequentially as follows;
S3.2, dividing the data at the time t 1-tn obtained in the step S3.1 into a plurality of continuous time sequences, wherein the length of each time sequence is j, the number of the time sequences obtained by dividing is K, namely, the time corresponding to the first time sequence is t 1-tj, the time corresponding to the second time sequence is t 2-t (j+1), the time corresponding to the third time sequence is t 3-t (j+2), and the time corresponding to the Kth time sequence is tK-t (j+K-1);
and dividing all the K time series data into a training set, a verification set and a test set according to the proportion relation of 8:1:1.
Further, the specific implementation method of the step S4 includes the following steps:
s4.1, setting an improved amnestic door activation value calculation expression in an improved LSTM neural network as follows:
;
Wherein, In order to forget the activation value of the door,For the sigmoid activation function,For the layer normalization,For the Dropout regularization,For the weight of the forgetting gate,As the final output of the last time step,For the input of the current time step,In order to perform the connection operation,In forgetting to leave the doorIs used for the weight of the (c),In order to be the cell state of the previous time step,Bias for forgetting the door;
S4.2, in the improved LSTM neural network, the method for calculating the activation value of the improved input gate is as follows:
;
Wherein, For the input of the activation value of the gate,For the weight of the input gate,Is input into the doorIs used for the weight of the (c),Bias for the input gate;
s4.3, in the improved LSTM neural network, the improved cell candidate value calculation method is as follows:
;
Wherein, As a candidate value for the cell,For the hyperbolic tangent activation function,As a weight for the candidate value of the cell,Among cell candidatesIs used for the weight of the (c),Bias for cell candidates;
s4.4, in the improved LSTM neural network, the improved calculation method of the cell state update is as follows:
;
Wherein, In order for the attention to be weighted,The function is activated for softmax and,For the coefficient of the attention weighting,In the attention weightingIs used for the weight of the (c),Bias for attention weight;
;
Wherein, Is in a cellular state; is a dot product operation;
S4.5, in the improved LSTM neural network, the improved output gate and the final output calculation method are as follows:
;
Wherein, In order to output the activation value of the gate,In order to output the weight of the gate,In order to output the weight of the gate,Offset for the output gate;
;
Wherein, Is the final output value;
And S4.6, updating an improved LSTM model calculation formula into an algorithm corresponding to the LSTM model, and deploying the improved LSTM model calculation formula into a computer to complete construction of the electric vertical take-off and landing aircraft flight monitoring index prediction model.
Further, the specific implementation method of the step S5 includes the following steps:
s5.1, inputting a training set into an electric vertical take-off and landing aircraft flight monitoring index prediction model for training, wherein in the training process, a loss function adopts a mean square error MSE, and the calculation method is as follows:
;
Wherein, For the number of samples to be taken,To be a true value of the value,Is a predicted value;
s5.2, setting the corresponding mean square error of the battery electric quantity, the motor temperature and the motor rotating speed to be sequentially recorded as 、、Then the total loss functionThe method comprises the following steps:
;
Wherein, The weight coefficient corresponding to the mean square error of the battery power,The weight coefficient corresponding to the mean square error of the motor temperature,The weight coefficient corresponds to the mean square error of the motor rotation speed;
S5.3, error correction is realized through a reverse propagation and random gradient descent SGD method, when the loss function value tends to be stable or reaches a preset training round, training is finished, and training of an electric vertical take-off and landing aircraft flight monitoring index prediction model is completed and is used for predicting battery power, motor temperature and motor rotating speed in the flight process of the eVTOL aircraft.
Further, the specific implementation method of the step S6 includes the following steps:
S6.1, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft obtained in the step S5, and predicting to obtain the first step Battery predictive power at timeFirst, thePredicted motor temperature at timeFirst, theMotor predictive rotational speed at time;
S6.2. For the first obtained in step S6.1Battery predictive power at timeActual monitoring with coulometer to obtain the firstThe actual electric quantity of the battery at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable battery electric quantity error determined by an expert shows that the battery electric quantity value predicted by the model is consistent with the actual monitoring value;
When (when) In this case, the battery power is predicted at time tp+1And the actual battery power at tp+1Further comparisons are made whenWhen the power consumption of the aircraft is reduced, the deviation between the predicted value and the actual monitoring value of the battery is considered to be caused by the deviation of the sensor, and eVTOL aircraft equipment is not required to be processed;
When (when) At the tp+2 time, the battery power is predictedAnd the actual battery power at tp+2Further comparisons are made whenIf the deviation of the predicted value and the actual monitored value of the battery power is the sensor deviation, the eVTOL aircraft equipment is not required to be processed, and ifWhen the battery power value predicted by the model is larger than the actual monitoring value, the deviation between the battery power value predicted by the model and the actual monitoring value is considered to be a problem of the battery itself, and the battery is required to be checked and maintained;
s6.3. For the first obtained in step S6.1 Predicted motor temperature at timeThe first step obtained by actual monitoring by adopting an infrared sensorThe actual temperature of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor temperature error determined by an expert shows that the motor temperature predicted by the model is consistent with the actual monitoring value;
When (when) At the time, the motor at the tp+1 time is predicted to have a temperatureAnd the actual temperature of the motor at time tp+1Further comparisons are made whenWhen the deviation of the predicted value and the monitored value of the motor temperature is considered to be caused by the deviation of the sensor, the eVTOL aircraft equipment is not required to be processed, and whenWhen the motor temperature value predicted by the model is larger than the actual monitoring value, the deviation between the motor temperature predicted value and the actual monitoring value is considered to be the problem of the motor itself, and the motor needs to be checked and maintained;
s6.4. For the first obtained in step S6.1 Motor predictive rotational speed at timeAnd the first is obtained by actual monitoring by an optical encoderThe actual rotation speed of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor rotation speed error determined by an expert shows that the motor rotation speed value predicted by the model is consistent with the actual monitoring value;
When (when) When the model is used, the predicted motor rotating speed value and the actual monitoring value are larger, and at the moment, the unmanned aerial vehicle motor needs to be immediately checked and maintained;
s6.5, when the battery electric quantity, the motor temperature and the motor rotating speed are abnormal at the same time, the motor rotating speed problem needs to be processed preferentially, the motor temperature problem needs to be processed, and finally the battery electric quantity problem is processed.
The invention has the beneficial effects that:
According to the method for diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft, provided by the invention, the problems of large data fluctuation, data time complexity, important characteristic excavation in data, overfitting and the like are considered, an LSTM prediction model is improved, and the accuracy of prediction on battery electric quantity, motor temperature and motor rotating speed is improved. The method has important values for ensuring the flight safety of eVTOL aircraft, optimizing the motor performance management, prolonging the service life of the battery, improving the energy efficiency and reducing the maintenance cost, so that the energy utilization is more efficient and economical.
According to the method for diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft, through accurately identifying the data abnormality, potential eVTOL aircraft equipment faults or performance degradation can be found in time, so that preventive measures can be taken or necessary maintenance can be carried out. The method is not only beneficial to ensuring the safe operation of eVTOL aircraft equipment and reducing the accident risk, but also can effectively prolong the service life of the equipment, reduce the operation cost and ensure the production efficiency and quality.
Drawings
Fig. 1 is a flowchart of a method for diagnosing an abnormal flight state of an electric vertical takeoff and landing aircraft according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings and detailed description. It should be understood that the embodiments described herein are for purposes of illustration only and are not intended to limit the invention, i.e., the embodiments described are merely some, but not all, of the embodiments of the invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein can be arranged and designed in a wide variety of different configurations, and the present invention can have other embodiments as well.
Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
For further understanding of the invention, the following detailed description is presented in conjunction with the accompanying drawings 1 to provide a further understanding of the invention in its aspects, features and efficacy:
Example 1:
a method for diagnosing the abnormal flight state of an electric vertical take-off and landing aircraft comprises the following steps:
s1, constructing flight monitoring indexes of an electric vertical take-off and landing aircraft, wherein the flight monitoring indexes are battery electric quantity, motor temperature and rotating speed;
Further, in step S1, a coulometer equipped with eVTOL aircraft is used to measure battery power, an infrared sensor is used to measure motor temperature, and an optical encoder is used to detect rotation speed;
s2, constructing influence factors of flight monitoring indexes of the electric vertical take-off and landing aircraft, wherein the influence factors comprise environmental temperature, wind speed, wind direction, load, flight time, battery voltage and battery current;
furthermore, when the ambient temperature is low, the chemical reaction rate of the battery can be reduced, the internal resistance is increased, the available electric quantity is reduced, the electric quantity consumption is accelerated, the low temperature ensures that the motor needs to reach the optimal working temperature for a longer time, meanwhile, the friction of mechanical parts between the motors is increased, and the rotating speed of the motor needs to be increased to overcome additional resistance.
When the wind speed is high and the wind goes downwind, the power consumption of the battery can be reduced, meanwhile, the cooling effect of the motor is increased, and the motor needs to reduce the rotating speed to keep the current flying speed.
The greater the load of eVTOL aircraft, the greater the battery power consumption, the higher the motor temperature, and the faster the motor speed. And measuring the load weight of the aircraft by adopting a piezoelectric weighing sensor.
The longer the flight time of eVTOL aircraft, the greater the power consumption, the higher the motor temperature, and the motor efficiency may be slightly reduced, requiring an increase in rotational speed to maintain the same performance.
The voltage and the electric quantity have nonlinear relation, and when the voltage is reduced, the temperature of the motor is increased while the same power is maintained, and when the voltage is reduced, the highest reachable rotating speed of the motor is reduced. And measuring the voltage of the battery terminal by adopting a voltage sensor.
The current directly determines the power output and the energy consumption. The increase in current results in an increase in the rate of electricity consumption, an increase in motor temperature, and an increase in motor speed. The current in the circuit is measured using a current sensor.
S3, acquiring flight monitoring index data of the electric vertical take-off and landing aircraft and influence factor data of the flight monitoring index of the electric vertical take-off and landing aircraft at the same time, and constructing an abnormal flight state diagnosis data set of the electric vertical take-off and landing aircraft, wherein the diagnosis data set comprises a training set, a verification set and a test set;
further, the specific implementation method of the step S3 includes the following steps:
S3.1, acquiring flight monitoring index data of the electric vertical take-off and landing aircrafts at the same time and influence factor data of the flight monitoring indexes of the electric vertical take-off and landing aircrafts;
Collecting unmanned aerial vehicle in flight process The battery electric quantity corresponding to the moment is sequentially as followsThe corresponding motor temperatures are in turnThe corresponding motor rotation speeds are in turn;
Collecting unmanned aerial vehicle flight process by utilizing thermal infrared imagerThe ambient temperature corresponding to the time is in turn;
Adopt hot wire anemometer to gather unmanned aerial vehicle flight in-processThe wind speeds corresponding to the moments are in turn;
Adopt ultrasonic wave anemoscope to gather unmanned aerial vehicle flight in-processThe wind directions corresponding to the moments are in turn;
Adopt piezoelectric weighing sensor to gather unmanned aerial vehicle flight in-processThe load corresponding to the moment is in turn;
Adopt GPS module to gather unmanned aerial vehicle flight in-processThe flight time corresponding to the moment is in turn;
Adopt voltage sensor to gather unmanned aerial vehicle flight in-processThe voltages of the batteries corresponding to the moments are sequentially as follows;
Adopt current sensor to gather unmanned aerial vehicle flight in-processThe current of the batteries corresponding to the moments is sequentially as follows;
S3.2, dividing the data at the time t 1-tn obtained in the step S3.1 into a plurality of continuous time sequences, wherein the length of each time sequence is j, the number of the time sequences obtained by dividing is K, namely, the time corresponding to the first time sequence is t 1-tj, the time corresponding to the second time sequence is t 2-t (j+1), the time corresponding to the third time sequence is t 3-t (j+2), and the time corresponding to the Kth time sequence is tK-t (j+K-1);
and dividing all the K time series data into a training set, a verification set and a test set according to the proportion relation of 8:1:1.
In the process of constructing a prediction model, two problems exist (1) in the traditional model, long-term dependency relationship between data is difficult to capture, namely, the model is difficult to memorize and use battery power, motor temperature and motor rotating speed at an earlier moment to predict the battery power, motor temperature and motor rotating speed at a future moment. (2) When predicting battery power, motor temperature, and motor speed, not all the information at the earlier time is equally important, and it is necessary to distinguish the importance of the information at the existing time.
Aiming at the problems, a scholars puts forward an LSTM (long-short term memory network) model, and the model introduces a gating mechanism comprising an input gate, a forget gate and an output gate, and effectively solves the long-term problem of data through long-term storage and forget of control information. At the same time, the gating mechanism of LSTM may allow the model to learn the information importance of different time nodes.
However, in LSTM, there are still the following problems, (1) in training LSTM, since the data in time series fluctuates greatly, the performance of the model may be affected, especially in the case of uneven data distribution. (2) Conventional LSTM models may not be accurate enough to handle data that contains complex time dependencies, such as motor speed or battery power changes, because the method typically relies on short-term inputs and hidden states only. (3) Conventional LSTM may ignore information that is particularly important for prediction when processing long sequence data, which is typically used to process the entire sequence. (4) In training a complex LSTM network, the model may overfit the training data, resulting in performance degradation in practical applications.
The above problems will lead to eVTOL that it is difficult to accurately predict the battery power, the motor temperature, and the motor speed based on the monitored data during the flight of the aircraft.
The LSTM model was improved. The method comprises the following steps:
(1) Layer normalization techniques are introduced. The layer normalization can stabilize the learning process of LSTM, and improve the adaptability of the model under different data conditions, which is very important for unmanned aerial vehicle to fly under different environments and load conditions. This technique can help the model adapt better to environmental changes, such as temperature changes affecting motor temperature, or load changes affecting battery consumption and motor performance.
(2) Peak connection techniques are introduced. The peak connection allows the model to directly use long-term memory to adjust the opening and closing of the gate, which is useful for predicting long-term consumption patterns such as battery or temperature and rotational speed of the motor under continuous operation.
(3) Attention mechanisms are introduced. The attention mechanism may help the model identify the most critical flight phase to battery power consumption, as well as specific operating conditions affecting motor temperature and speed. For example, the model may learn the mode in which the battery is most consumed at take-off and landing, and the changes in motor temperature and speed at high speed of flight.
(4) Dropout regularization technique was introduced. In the prediction model of the battery power, the motor temperature and the rotating speed of the unmanned aerial vehicle, the Dropout can be used for increasing the generalization capability of the model and reducing the prediction error under new or unseen flight conditions. This helps to maintain the stability and accuracy of the predictions when the drone is subjected to different environmental conditions or flight modes.
By adopting the 4 techniques, the forget gate, the input gate, the cell candidate value, the cell state update value, the output gate and the final output in the traditional LSTM are improved.
The cell state is the core of network memory and communicates information through a series of time steps. The cell status can be seen as a flowing "information road" carrying all the information that may be important for future decisions. In LSTM this cell state can be transferred in a virtually constant fashion at different time steps, with only fine adjustments being made by means of gating.
S4, constructing an electric vertical take-off and landing aircraft flight monitoring index prediction model based on the improved LSTM neural network;
further, the specific implementation method of the step S4 includes the following steps:
s4.1, setting an improved amnestic door activation value calculation expression in an improved LSTM neural network as follows:
;
Wherein, In order to forget the activation value of the door,For the sigmoid activation function,For the layer normalization,For the Dropout regularization,For the weight of the forgetting gate,As the final output of the last time step,For the input of the current time step,In order to perform the connection operation,In forgetting to leave the doorIs used for the weight of the (c),In order to be the cell state of the previous time step,Bias for forgetting the door;
further, the function of the forgetting gate is to determine the information to discard from the cell state. It calculates a value between 0 and 1 by a sigmoid activation function, which multiplies the cell state element by element, thus achieving "forgetting" of the state.
For eVTOL aircraft flight, for battery power prediction, a forgetting door can help the model ignore insignificant changes if the discharge rate of the battery is not changed under certain conditions, for motor temperature prediction, a forgetting door can help the model discard historical temperature information which is irrelevant to current temperature prediction, such as past abnormal temperature fluctuation, and for motor speed prediction, when the speed is changed rapidly, the forgetting door can make the model ignore outdated speed information and concentrate on more recent data.
S4.2, in the improved LSTM neural network, the method for calculating the activation value of the improved input gate is as follows:
;
Wherein, For the input of the activation value of the gate,For the weight of the input gate,Is input into the doorIs used for the weight of the (c),Bias for the input gate;
Further, the input gate controls the importance of the new information, determining the new information to be added to the cell state. The cell candidate is then a transformation of the new information that may be added to the cell state.
For eVTOL in the flight process of the aircraft, the input door controls the importance of new electric quantity input information (such as charging state, using time and the like) for battery electric quantity prediction, the cell candidate value updates the electric quantity state according to the information, for motor temperature prediction, new environment or machine operation information updates a temperature prediction model through the input door and the cell candidate value, and for motor rotating speed prediction, the input door and the cell candidate value help the model integrate the latest operation command and environment factors, and the rotating speed prediction is adjusted.
S4.3, in the improved LSTM neural network, the improved cell candidate value calculation method is as follows:
;
Wherein, As a candidate value for the cell,For the hyperbolic tangent activation function,As a weight for the candidate value of the cell,Among cell candidatesIs used for the weight of the (c),Bias for cell candidates;
s4.4, in the improved LSTM neural network, the improved calculation method of the cell state update is as follows:
;
Wherein, In order for the attention to be weighted,The function is activated for softmax and,For the coefficient of the attention weighting,In the attention weightingIs used for the weight of the (c),Bias for attention weight;
;
Wherein, Is in a cellular state; is a dot product operation;
Further, the cell state is where the LSTM stores past information, which is updated at each time step to maintain long-term dependent information.
For eVTOL in the flight process of the aircraft, when the battery electric quantity, the motor temperature and the motor rotating speed are predicted, the updating of the cell state allows the model to retain and forget information, so that only useful information is ensured to be retained, and the prediction accuracy is improved.
S4.5, in the improved LSTM neural network, the improved output gate and the final output calculation method are as follows:
;
Wherein, In order to output the activation value of the gate,In order to output the weight of the gate,In order to output the weight of the gate,Offset for the output gate;
;
Wherein, Is the final output value;
The output gate controls the amount of information output from the LSTM cell. The final output is based on the current cell state and the output of the output gate, which is the output that is passed on to the next time step and the portion of the final output.
For eVTOL in the flight process of the aircraft, the output gate and the final output help to determine the electric quantity output of each step for battery management, data support is provided for battery management, for motor temperature prediction, the model can accurately predict the instant temperature of the motor by adjusting the output gate and the final output, help to prevent overheat, and for motor rotating speed prediction, the updating of the output gate and the final output enables the model to reflect the current and future possible rotating speed changes, and optimize the machine performance.
And S4.6, updating an improved LSTM model calculation formula into an algorithm corresponding to the LSTM model, and deploying the improved LSTM model calculation formula into a computer to complete construction of the electric vertical take-off and landing aircraft flight monitoring index prediction model.
S5, training the electric vertical take-off and landing aircraft flight monitoring index prediction model by using the training set obtained in the step S3 to obtain a trained electric vertical take-off and landing aircraft flight monitoring index prediction model;
Further, the specific implementation method of the step S5 includes the following steps:
s5.1, inputting a training set into an electric vertical take-off and landing aircraft flight monitoring index prediction model for training, wherein in the training process, a loss function adopts a mean square error MSE, and the calculation method is as follows:
;
Wherein, For the number of samples to be taken,To be a true value of the value,Is a predicted value;
s5.2, setting the corresponding mean square error of the battery electric quantity, the motor temperature and the motor rotating speed to be sequentially recorded as 、、Then the total loss functionThe method comprises the following steps:
;
Wherein, The weight coefficient corresponding to the mean square error of the battery power,The weight coefficient corresponding to the mean square error of the motor temperature,The weight coefficient corresponds to the mean square error of the motor rotation speed;
S5.3, error correction is realized by a reverse propagation and random gradient descent SGD method, and when the loss function value tends to be stable or reaches a preset training round, training is finished;
S6, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft, and diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft based on prediction data.
Further, the specific implementation method of the step S6 includes the following steps:
S6.1, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft obtained in the step S5, and predicting to obtain the first step Battery predictive power at timeFirst, thePredicted motor temperature at timeFirst, theMotor predictive rotational speed at time;
S6.2. For the first obtained in step S6.1Battery predictive power at timeActual monitoring with coulometer to obtain the firstThe actual electric quantity of the battery at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable battery electric quantity error determined by an expert shows that the battery electric quantity value predicted by the model is consistent with the actual monitoring value;
When (when) In this case, the battery power is predicted at time tp+1And the actual battery power at tp+1Further comparisons are made whenWhen the power consumption of the aircraft is reduced, the deviation between the predicted value and the actual monitoring value of the battery is considered to be caused by the deviation of the sensor, and eVTOL aircraft equipment is not required to be processed;
When (when) At the tp+2 time, the battery power is predictedAnd the actual battery power at tp+2Further comparisons are made whenIf the deviation of the predicted value and the actual monitored value of the battery power is the sensor deviation, the eVTOL aircraft equipment is not required to be processed, and ifWhen the battery power value predicted by the model is larger than the actual monitoring value, the deviation between the battery power value predicted by the model and the actual monitoring value is considered to be a problem of the battery itself, and the battery is required to be checked and maintained;
s6.3. For the first obtained in step S6.1 Predicted motor temperature at timeThe first step obtained by actual monitoring by adopting an infrared sensorThe actual temperature of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor temperature error determined by an expert shows that the motor temperature predicted by the model is consistent with the actual monitoring value;
When (when) At the time, the motor at the tp+1 time is predicted to have a temperatureAnd the actual temperature of the motor at time tp+1Further comparisons are made whenWhen the deviation of the predicted value and the monitored value of the motor temperature is considered to be caused by the deviation of the sensor, the eVTOL aircraft equipment is not required to be processed, and whenWhen the motor temperature value predicted by the model is larger than the actual monitoring value, the deviation between the motor temperature predicted value and the actual monitoring value is considered to be the problem of the motor itself, and the motor needs to be checked and maintained;
s6.4. For the first obtained in step S6.1 Motor predictive rotational speed at timeAnd the first is obtained by actual monitoring by an optical encoderThe actual rotation speed of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor rotation speed error determined by an expert shows that the motor rotation speed value predicted by the model is consistent with the actual monitoring value;
When (when) When the model is used, the predicted motor rotating speed value and the actual monitoring value are larger, and at the moment, the unmanned aerial vehicle motor needs to be immediately checked and maintained;
s6.5, when the battery electric quantity, the motor temperature and the motor rotating speed are abnormal at the same time, the motor rotating speed problem needs to be processed preferentially, the motor temperature problem needs to be processed, and finally the battery electric quantity problem is processed.
The key points and the points to be protected of the invention are as follows:
(1) A battery power, motor temperature and motor rotation speed prediction model.
(2) A method for discriminating abnormal data of battery power, motor temperature and motor rotation speed is provided.
It is noted that relational terms such as "first" and "second", and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
Although the application has been described above with reference to specific embodiments, various modifications may be made and equivalents may be substituted for elements thereof without departing from the scope of the application. In particular, the features of the disclosed embodiments may be combined with each other in any manner so long as there is no structural conflict, and the exhaustive description of these combinations is not given in this specification solely for the sake of brevity and resource saving. Therefore, it is intended that the application not be limited to the particular embodiments disclosed herein, but that the application will include all embodiments falling within the scope of the appended claims.
Claims (6)
1. The method for diagnosing the abnormal flight state of the electric vertical take-off and landing aircraft is characterized by comprising the following steps of:
s1, constructing flight monitoring indexes of an electric vertical take-off and landing aircraft, wherein the flight monitoring indexes are battery electric quantity, motor temperature and rotating speed;
s2, constructing influence factors of flight monitoring indexes of the electric vertical take-off and landing aircraft, wherein the influence factors comprise environmental temperature, wind speed, wind direction, load, flight time, battery voltage and battery current;
S3, acquiring flight monitoring index data of the electric vertical take-off and landing aircraft and influence factor data of the flight monitoring index of the electric vertical take-off and landing aircraft at the same time, and constructing an abnormal flight state diagnosis data set of the electric vertical take-off and landing aircraft, wherein the diagnosis data set comprises a training set, a verification set and a test set;
s4, constructing an electric vertical take-off and landing aircraft flight monitoring index prediction model based on the improved LSTM neural network;
s5, training the electric vertical take-off and landing aircraft flight monitoring index prediction model by using the training set obtained in the step S3 to obtain a trained electric vertical take-off and landing aircraft flight monitoring index prediction model;
S6, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft, and diagnosing the flight abnormal state of the electric vertical take-off and landing aircraft based on prediction data.
2. The method for diagnosing a flight anomaly of an electric vertical takeoff and landing aircraft according to claim 1, wherein in step S1, a coulometer equipped with eVTOL aircraft is used to measure battery power, an infrared sensor is used to measure motor temperature, and an optical encoder is used to detect rotation speed.
3. The method for diagnosing a flight anomaly of an electric vertical takeoff and landing aircraft according to claim 2, wherein the specific implementation method of step S3 includes the steps of:
S3.1, acquiring flight monitoring index data of the electric vertical take-off and landing aircrafts at the same time and influence factor data of the flight monitoring indexes of the electric vertical take-off and landing aircrafts;
Collecting unmanned aerial vehicle in flight process The battery electric quantity corresponding to the moment is sequentially as followsThe corresponding motor temperatures are in turnThe corresponding motor rotation speeds are in turn;
Collecting unmanned aerial vehicle flight process by utilizing thermal infrared imagerThe ambient temperature corresponding to the time is in turn;
Adopt hot wire anemometer to gather unmanned aerial vehicle flight in-processThe wind speeds corresponding to the moments are in turn;
Adopt ultrasonic wave anemoscope to gather unmanned aerial vehicle flight in-processThe wind directions corresponding to the moments are in turn;
Adopt piezoelectric weighing sensor to gather unmanned aerial vehicle flight in-processThe load corresponding to the moment is in turn;
Adopt GPS module to gather unmanned aerial vehicle flight in-processThe flight time corresponding to the moment is in turn;
Adopt voltage sensor to gather unmanned aerial vehicle flight in-processThe voltages of the batteries corresponding to the moments are sequentially as follows;
Adopt current sensor to gather unmanned aerial vehicle flight in-processThe current of the batteries corresponding to the moments is sequentially as follows;
S3.2, dividing the data at the time t 1-tn obtained in the step S3.1 into a plurality of continuous time sequences, wherein the length of each time sequence is j, the number of the time sequences obtained by dividing is K, namely, the time corresponding to the first time sequence is t 1-tj, the time corresponding to the second time sequence is t 2-t (j+1), the time corresponding to the third time sequence is t 3-t (j+2), and the time corresponding to the Kth time sequence is tK-t (j+K-1);
and dividing all the K time series data into a training set, a verification set and a test set according to the proportion relation of 8:1:1.
4. The method for diagnosing a flight anomaly of an electric vertical takeoff and landing aircraft according to claim 3, wherein the specific implementation method of step S4 includes the steps of:
s4.1, setting an improved amnestic door activation value calculation expression in an improved LSTM neural network as follows:
;
Wherein, In order to forget the activation value of the door,For the sigmoid activation function,For the layer normalization,For the Dropout regularization,For the weight of the forgetting gate,As the final output of the last time step,For the input of the current time step,In order to perform the connection operation,In forgetting to leave the doorIs used for the weight of the (c),In order to be the cell state of the previous time step,Bias for forgetting the door;
S4.2, in the improved LSTM neural network, the method for calculating the activation value of the improved input gate is as follows:
;
Wherein, For the input of the activation value of the gate,For the weight of the input gate,Is input into the doorIs used for the weight of the (c),Bias for the input gate;
s4.3, in the improved LSTM neural network, the improved cell candidate value calculation method is as follows:
;
Wherein, As a candidate value for the cell,For the hyperbolic tangent activation function,As a weight for the candidate value of the cell,Among cell candidatesIs used for the weight of the (c),Bias for cell candidates;
s4.4, in the improved LSTM neural network, the improved calculation method of the cell state update is as follows:
;
Wherein, In order for the attention to be weighted,The function is activated for softmax and,For the coefficient of the attention weighting,In the attention weightingIs used for the weight of the (c),Bias for attention weight;
;
Wherein, Is in a cellular state; is a dot product operation;
S4.5, in the improved LSTM neural network, the improved output gate and the final output calculation method are as follows:
;
Wherein, In order to output the activation value of the gate,In order to output the weight of the gate,In order to output the weight of the gate,Offset for the output gate;
;
Wherein, Is the final output value;
And S4.6, updating an improved LSTM model calculation formula into an algorithm corresponding to the LSTM model, and deploying the improved LSTM model calculation formula into a computer to complete construction of the electric vertical take-off and landing aircraft flight monitoring index prediction model.
5. The method for diagnosing a flight anomaly of an electric vertical takeoff and landing aircraft according to claim 4, wherein the specific implementation method of step S5 includes the steps of:
s5.1, inputting a training set into an electric vertical take-off and landing aircraft flight monitoring index prediction model for training, wherein in the training process, a loss function adopts a mean square error MSE, and the calculation method is as follows:
;
Wherein, For the number of samples to be taken,To be a true value of the value,Is a predicted value;
s5.2, setting the corresponding mean square error of the battery electric quantity, the motor temperature and the motor rotating speed to be sequentially recorded as 、、Then the total loss functionThe method comprises the following steps:
;
Wherein, The weight coefficient corresponding to the mean square error of the battery power,The weight coefficient corresponding to the mean square error of the motor temperature,The weight coefficient corresponds to the mean square error of the motor rotation speed;
S5.3, error correction is realized through a reverse propagation and random gradient descent SGD method, when the loss function value tends to be stable or reaches a preset training round, training is finished, and training of an electric vertical take-off and landing aircraft flight monitoring index prediction model is completed and is used for predicting battery power, motor temperature and motor rotating speed in the flight process of the eVTOL aircraft.
6. The method for diagnosing a flight anomaly of an electric vertical takeoff and landing aircraft according to claim 5, wherein the specific implementation method of step S6 includes the steps of:
S6.1, predicting the flight monitoring index of the electric vertical take-off and landing aircraft by using the trained flight monitoring index prediction model of the electric vertical take-off and landing aircraft obtained in the step S5, and predicting to obtain the first step Battery predictive power at timeFirst, thePredicted motor temperature at timeFirst, theMotor predictive rotational speed at time;
S6.2. For the first obtained in step S6.1Battery predictive power at timeActual monitoring with coulometer to obtain the firstThe actual electric quantity of the battery at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable battery electric quantity error determined by an expert shows that the battery electric quantity value predicted by the model is consistent with the actual monitoring value;
When (when) In this case, the battery power is predicted at time tp+1And the actual battery power at tp+1Further comparisons are made whenWhen the power consumption of the aircraft is reduced, the deviation between the predicted value and the actual monitoring value of the battery is considered to be caused by the deviation of the sensor, and eVTOL aircraft equipment is not required to be processed;
When (when) At the tp+2 time, the battery power is predictedAnd the actual battery power at tp+2Further comparisons are made whenIf the deviation of the predicted value and the actual monitored value of the battery power is the sensor deviation, the eVTOL aircraft equipment is not required to be processed, and ifWhen the battery power value predicted by the model is larger than the actual monitoring value, the deviation between the battery power value predicted by the model and the actual monitoring value is considered to be a problem of the battery itself, and the battery is required to be checked and maintained;
s6.3. For the first obtained in step S6.1 Predicted motor temperature at timeThe first step obtained by actual monitoring by adopting an infrared sensorThe actual temperature of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor temperature error determined by an expert shows that the motor temperature predicted by the model is consistent with the actual monitoring value;
When (when) At the time, the motor at the tp+1 time is predicted to have a temperatureAnd the actual temperature of the motor at time tp+1Further comparisons are made whenWhen the deviation of the predicted value and the monitored value of the motor temperature is considered to be caused by the deviation of the sensor, the eVTOL aircraft equipment is not required to be processed, and whenWhen the motor temperature value predicted by the model is larger than the actual monitoring value, the deviation between the motor temperature predicted value and the actual monitoring value is considered to be the problem of the motor itself, and the motor needs to be checked and maintained;
s6.4. For the first obtained in step S6.1 Motor predictive rotational speed at timeAnd the first is obtained by actual monitoring by an optical encoderThe actual rotation speed of the motor at the moment is;
Comparison ofAnd (3) withWhen (when)In whichThe allowable motor rotation speed error determined by an expert shows that the motor rotation speed value predicted by the model is consistent with the actual monitoring value;
When (when) When the model is used, the predicted motor rotating speed value and the actual monitoring value are larger, and at the moment, the unmanned aerial vehicle motor needs to be immediately checked and maintained;
s6.5, when the battery electric quantity, the motor temperature and the motor rotating speed are abnormal at the same time, the motor rotating speed problem needs to be processed preferentially, the motor temperature problem needs to be processed, and finally the battery electric quantity problem is processed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411699558.3A CN119203006B (en) | 2024-11-26 | 2024-11-26 | A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411699558.3A CN119203006B (en) | 2024-11-26 | 2024-11-26 | A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119203006A true CN119203006A (en) | 2024-12-27 |
| CN119203006B CN119203006B (en) | 2025-04-22 |
Family
ID=94042415
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411699558.3A Active CN119203006B (en) | 2024-11-26 | 2024-11-26 | A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119203006B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119203007A (en) * | 2024-11-26 | 2024-12-27 | 深圳市城市交通规划设计研究中心股份有限公司 | An emergency landing method based on abnormal battery voltage of eVTOL aircraft |
| CN120199134A (en) * | 2025-03-19 | 2025-06-24 | 安胜（天津）飞行模拟系统有限公司 | Low-altitude pilot AI model training application system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190354644A1 (en) * | 2018-05-18 | 2019-11-21 | Honeywell International Inc. | Apparatuses and methods for detecting anomalous aircraft behavior using machine learning applications |
| CN112633361A (en) * | 2020-12-20 | 2021-04-09 | 中国人民解放军空军预警学院 | Flight emergency prediction method and device based on LSTM neural network |
| CN116704641A (en) * | 2023-06-16 | 2023-09-05 | 四川大学 | Method, device, equipment and storage medium for abnormal detection of UAV flight data |
| CN118570683A (en) * | 2024-07-30 | 2024-08-30 | 天之翼(苏州)科技有限公司 | Unmanned aerial vehicle state monitoring method and system based on deep learning |
-
2024
- 2024-11-26 CN CN202411699558.3A patent/CN119203006B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190354644A1 (en) * | 2018-05-18 | 2019-11-21 | Honeywell International Inc. | Apparatuses and methods for detecting anomalous aircraft behavior using machine learning applications |
| CN112633361A (en) * | 2020-12-20 | 2021-04-09 | 中国人民解放军空军预警学院 | Flight emergency prediction method and device based on LSTM neural network |
| CN116704641A (en) * | 2023-06-16 | 2023-09-05 | 四川大学 | Method, device, equipment and storage medium for abnormal detection of UAV flight data |
| CN118570683A (en) * | 2024-07-30 | 2024-08-30 | 天之翼(苏州)科技有限公司 | Unmanned aerial vehicle state monitoring method and system based on deep learning |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119203007A (en) * | 2024-11-26 | 2024-12-27 | 深圳市城市交通规划设计研究中心股份有限公司 | An emergency landing method based on abnormal battery voltage of eVTOL aircraft |
| CN120199134A (en) * | 2025-03-19 | 2025-06-24 | 安胜（天津）飞行模拟系统有限公司 | Low-altitude pilot AI model training application system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119203006B (en) | 2025-04-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN119203006B (en) | A method for diagnosing abnormal flight status of an electric vertical take-off and landing aircraft | |
| CN117952592B (en) | Intelligent management method for charging pile | |
| CN117613421A (en) | An equalization method and medium for energy storage battery pack management system based on neural network | |
| CN111157898A (en) | New energy vehicle online battery fault detection and analysis method and device | |
| CN114757086B (en) | Real-time remaining service life prediction method and system for multi-rotor unmanned aerial vehicles | |
| CN119447611B (en) | A thermal management control system and method for power battery packs | |
| Yang et al. | The remaining useful life estimation of lithium-ion battery based on improved extreme learning machine algorithm | |
| CN120410254A (en) | A predictive maintenance method for integrated photovoltaic, storage and charging power stations based on deep learning | |
| CN116184214A (en) | On-line rapid prediction method, device and storage medium for vehicle power battery life | |
| Kang et al. | Satellite power system state prediction based on online learning with parameter association rules | |
| Gao et al. | A dynamic warning method for electric vehicle charging safety based on CNN-BiGRU hybrid model | |
| CN118198542A (en) | Lithium battery formation component cabinet | |
| CN120821201B (en) | Photovoltaic snow-proof shed tunnel self-adaptive snow melting method and system | |
| CN119203007A (en) | An emergency landing method based on abnormal battery voltage of eVTOL aircraft | |
| CN120336789A (en) | Intelligent fault diagnosis method for renewable energy equipment based on deep learning | |
| CN119091401B (en) | Future traffic volume and energy consumption prediction method for unmanned aerial vehicle in traffic industry | |
| CN120511654A (en) | Industrial park load prediction method based on artificial intelligence | |
| CN119202716A (en) | A temperature sensor fault diagnosis and fault data reconstruction method, system, device and medium | |
| CN118211094A (en) | A method for detecting faults in a gearbox of a wind turbine generator set | |
| Mathi et al. | SoC Estimation and Comparative Analysis of Lithium Polymer and Lithium-Ion Batteries in Unmanned Aerial Vehicles | |
| Xiaoyang | Deep Learning Based Prediction of Lithium Battery State of Health | |
| CN120449719B (en) | Unmanned aerial vehicle climbing process energy-saving effect prediction analysis method and system | |
| CN120728092B (en) | A heat pipe-assisted immersion battery pack thermal management system | |
| US20250371962A1 (en) | Alerting method in the event of failure of an energy production device and associated electronic device | |
| US20250189593A1 (en) | Battery system control based on state space aging modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |