KR102718986B1 - Data Logger System for Intelligent Data Analysis - Google Patents

Abstract

The embodiment relates to an AI-based data logger system.
Specifically, these AI-based data logger systems are provided with a separate AI processing unit so that AI-based data logger functions can be added in addition to existing input/output information. In addition, in the event of a power outage or communication failure in the central control unit, the AI processing unit can later check the data at the time of the problem and restore it to the central control unit's computer.
In addition, in these cases, we provide a data analysis solution that detects and responds to process abnormalities in advance through real-time monitoring by connecting various data gathering and communications in general factories or HACCP factories, and supports alarm alerts through data analysis.
That is, it provides a remote monitoring support hub role by collecting various data from general factories or HACCP sites and comprehensively collecting and analyzing heterogeneous protocols using a data collection and analysis tool.
In addition, it provides support services for implementing device hardware based on SBC (Single Board Computing) architecture and data collection/processing/transmission from various general-purpose facilities, and provides various I/Fs that can collect data from manufacturing facilities and devices (sensors).
Accordingly, through this, a separate AI processing unit is provided so that an AI-based data logger function can be added in addition to the existing input/output information. In addition, an AI-based data logger system is provided that has the function of checking the data at the time of the problem through the AI processing unit and restoring it to the computer of the central control unit in the event of a power outage or communication failure of the central control unit.
In addition, these data logger systems collect and process signals acquired from various sensors through data collection and processing devices, and provide data processing and analysis to enable analysis and diagnosis.
In addition, these data collection and analysis devices solve the problem of complicated measurements and many connecting devices, which are inconvenient to use. Therefore, a signal collection device that acts as a comprehensive protocol gateway that can collect and analyze various signals such as temperature/pressure/noise/vibration by connecting to a portable PC or laptop is provided.

Description

{Data Logger System for Intelligent Data Analysis}

The subject matter disclosed in this specification relates to a data logger system, and more specifically, to a technology for transmitting field monitoring data monitored by a field control panel equipped with an external sensor to a server, storing and processing the data based on the data, and then receiving and using the necessary data from each control panel.

Unless otherwise indicated herein, the materials described in this section are not prior art to the claims of this application and their inclusion in this section is not an admission that they are prior art.

In general, a field measurement system consists of a field control panel and a central management information processing unit located at a remote location. The field control panel generates field monitoring data using various external sensors connected to it and transmits this to the remote management information processing unit using a communication network.

And, these central management information processing devices manage the on-site situation based on these transmitted surveillance data, and based on these, they can remotely operate various devices connected to the on-site control panel.

For example, the existing system uses PLC (Programmable Logic Control) and RTU (Remote Terminal Unit) as the field control unit. Therefore, when the PLC collects the status information of the monitoring target from the measuring device, it transmits the status information to the management information processing device through RTU, and receives control according to the current status to appropriately operate the load equipment such as motors and valves. Therefore, the monitoring target is operated according to the state desired by the manager.

On the other hand, among the conventional field measurement systems, the field control panel simply transmits or receives input/output information without the function of a data logger, so it is not easy to check the field monitoring data during the period when a power outage or communication line abnormality occurs. In addition, there is a problem in that the data cannot be checked even after the power outage or communication line abnormality ends because there is no backup function for this data.

That is, when these systems exhibit failures in the PLC, RTU, or communication lines, they must be quickly controlled in a different way, which is not easy. In addition, if a separate emergency system is installed at this time, there may be various difficulties in operation, installation, and maintenance.

(Patent Document 0001) KR10-2108932 B1

(Patent Document 0002) KR10-2019-0046244 A

(Patent Document 0003) KR10-2229387 B1

For reference, the technology of this patent document 1 allows for easy setting completion by connecting sensors necessary for facility measurement to collect sensor detection values even if the type or number of sensors to be installed differs depending on the classification, type, shape, or size of the facility.

In addition, Patent Document 2 can be installed and used in multiple categories of social infrastructure facilities that utilize different types of sensors for remote monitoring, and selectively connects to different types of IoT communication networks.

In addition, patent document 3 discloses a technology for measuring data in a remote location using a modem data logger.

The disclosed content is to provide an AI-based data logger system that has a separate AI processing unit so that an AI-based data logger function can be added in addition to existing input/output information, and has the function of checking the data at the time of the problem through the AI processing unit and restoring it to the central server computer even in the event of a power outage or communication failure of the central control unit.

In addition, in these cases, the AI processing unit provides a data analysis solution that connects various data gathering and communications in general factories or HACCP factories, detects and responds to process abnormalities in advance through real-time monitoring, and supports alarm alerts through data analysis.

In addition, the AI processing unit provides services for data collection/processing/transmission from various general-purpose facilities and device hardware implementation based on the SBC (Single Board Computing) architecture, and provides various I/Fs that can collect data from manufacturing facilities and devices (sensors).

The AI-based data logger system according to the embodiment is:

First, in the AI processing unit, that is, the AI-based intelligent data collection and analysis device provides a communication protocol (Unified Gateway) with the PLC or SCADA/HMI process monitoring and control system and heterogeneous utility linkage system to receive data from field equipment. In addition, in such cases, it supports a reliable protocol that supports various devices and data transmission methods.

For example, compared to managing a single data item or sensing method, it provides various data interfaces such as various sensors, devices, IO boards, and PLCs.

In particular, it is characterized by being an AI-based intelligent data collection and analysis device that links with on-site sensors and devices and acts as a communication gateway, providing analysis, management, and monitoring alarms for data, and in addition, providing an optimized data collection and analysis control system.

According to embodiments, a separate AI processing unit is provided so that an AI-based data logger function can be added in addition to existing input/output information, and an AI-based data logger system having a function of checking data at a problematic point in time through the AI processing unit and restoring it to the computer of the central control unit in the event of a power outage or communication failure of the central control unit is provided.

And also, this AI processing unit, that is, the AI-based intelligent data collection and analysis device, collects and processes signals acquired from various sensors, and provides data processing and analysis to enable analysis and diagnosis.

In addition, these data collection and analysis devices solve the problem of complicated measurements and many connecting devices, which are inconvenient to use. Therefore, a signal collection device that acts as a comprehensive protocol gateway that can collect and analyze various signals such as temperature/pressure/noise/vibration by connecting to a portable PC or laptop is provided.

Figures 1 to 3 are drawings conceptually explaining an AI-based data logger system according to an embodiment.
Figure 4 is a diagram illustrating an overall AI-based data logger system according to an embodiment of the present invention.
Figure 5 is a block diagram illustrating the configuration of an AI-based data logger system according to an embodiment.
Figure 6 is a flow chart showing the operation of an AI-based data logger system according to an embodiment in sequence.
Figure 7 is a drawing showing an HMI monitoring screen in a factory applied to an AI-based data logger system according to an embodiment of the present invention.

Figures 1 to 3 are drawings schematically illustrating an AI-based data logger system according to one embodiment.

As illustrated in FIGS. 1 to 3, the data logger system of one embodiment is provided with a separate AI processing unit so that an AI-based data logger function can be added in addition to the existing input/output information as illustrated in FIG. 1. In addition, even in the event of a power outage or communication failure of the central server, the system has a function of checking the data at the time of the problem later by the AI processing unit and restoring it to the central server computer.

In addition, in these cases, we provide a data analysis solution that detects and responds to process abnormalities in advance through real-time monitoring by connecting various data gathering and communications in general factories or HACCP factories, and supports alarm alerts through data analysis.

That is, it provides a remote monitoring support hub role by collecting various data from general factories or HACCP sites and comprehensively collecting and analyzing heterogeneous protocols using a data collection and analysis tool.

In addition, it provides support services for implementing device hardware based on SBC (Single Board Computing) architecture and data collection/processing/transmission from various general-purpose facilities, and provides various I/Fs that can collect data from manufacturing facilities and devices (sensors).

Specifically, for this purpose, the data logger system is equipped with an AI processing unit, i.e., an AI-based data collection and analysis device, as shown in Fig. 2.

That is, these devices act as heterogeneous data protocol gateways, perform data control suitable for field situations, and predefine the data collection and analysis controller in the protocol platform and application that controls the transmission of data from the data collection system.

In addition, these devices also provide field data collection devices, and these collection devices search for various sensors and provide an OPC-UA information collection device interface that is optimized for each factory and capable of collecting field data. In addition, they also provide a user interface and HMI that are convenient for users to handle the information of the sensors collected in this way.

On the one hand, these data logger systems provide separate configurations for gathering and connecting heterogeneous field data in such cases, thereby enabling smooth collection and analysis of heterogeneous field data.

For example, it communicates wired or wirelessly depending on the factory environment and situation, and applies wired and wireless hybridization. Specifically, it monitors the surrounding conditions such as temperature and humidity with sensors, converts the data values into digital, and transmits controller speed and On, Off control commands to the operating unit to control the temperature and organic matter concentration inside the factory.

Accordingly, an AI-based data logger system according to one embodiment of the present invention is provided with a separate AI processing unit so that an AI-based data logger function can be added in addition to existing input/output information. In addition, an AI-based data logger system having a function of checking the data at the time of the problem through the AI processing unit and restoring it to the computer of the central control unit is provided even in the event of a power outage or communication failure of the central control unit.

In addition, these data logger systems collect and process signals acquired from various sensors through data collection and processing devices, and provide data processing and analysis to enable analysis and diagnosis.

In addition, these data collection and analysis devices solve the problem of complicated measurements and many connecting devices, which are inconvenient to use. Therefore, a signal collection device that acts as a comprehensive protocol gateway that can collect and analyze various signals such as temperature/pressure/noise/vibration by connecting to a portable PC or laptop is provided.

Figure 4 is a diagram showing an overall AI-based data logger system according to an embodiment.

As illustrated in FIG. 4, an AI-based data logger system according to an embodiment is largely composed of a number of different factory facilities and field control panels and AI processing units installed for each situation, and a central management information processing device located in a remote location (see a).

Additionally, these systems may further include a fault repair center information processing device (not shown) that is connected to a management information processing device to provide services such as fault repair.

The above control panel is connected to various external sensors that monitor the on-site situation, and the data monitored by the external sensors is transmitted to the AI processing unit via the Internet. Of course, the control panel can receive drive signals for various devices from the management information processing device and control various devices connected to it based on these signals.

The above AI processing unit collects various data transmitted from the remote field control panel, and the collected data is classified and processed by a deep learning module equipped with an artificial neural network, and is configured to provide various statistical data and alarms for efficient data management (see b). In addition, the AI processing unit stores various transmitted data in the form of a database, and can be configured to provide various statistical data and graph data in response to user requests via Internet access. In addition, the system's self-diagnosis function is provided, so that accidents can be prevented in advance by providing the occurrence of abnormal situations to pre-registered users through various routes (e.g., text message transmission, etc.). In this way, the AI processing unit functions as a kind of data logger. This AI processing unit can be more useful when managing multiple workplaces. When multiple workplaces are managed by the AI processing unit, duplicate investment due to the construction of an independent system can be prevented, and efficient management is possible with a minimum number of personnel. In other words, several to dozens of field control units are provided, and data from these dozens of field control units is transmitted to an AI-based processing unit, so that data management of the field control units can be performed centrally and in batches.

The above management information processing device can access the AI processing unit using the Internet, receive various data stored in the AI processing unit, output the same, and in some cases, download and store various data. Therefore, even if the management information processing device experiences a power outage or communication failure, the data for the corresponding period can be downloaded and restored after the end of the failure. Accordingly, the AI-based data logger system according to the embodiment of the present disclosure provides an AI-based data logger system in which a separate AI processing unit is provided so that an AI-based data logger function can be added in addition to existing input/output information, and even in the event of a power outage or communication failure of the central control unit, the AI-based data logger system has a function of checking the data at the time of the problem in the future by the AI processing unit and restoring it to the computer of the central control unit.

Meanwhile, another AI-based data logger system according to the present disclosure can be applied as follows.

It can be applied to the development of smart products optimized for smart factories of small businesses, and specifically, it can be applied to the development of interfaces with sensor product groups such as power sensors, heat detection sensors, coefficient sensors, scanners, temperature sensors, and vibration sensors that are mostly required in small smart factories.

In addition, it can be applied to companies participating in data collection during the business period to present a representative application model, and can also be applied to the technology development of products that contribute to the spread of smart factories for small businesses by effectively collecting data from various small factories in the future.

In addition, it can be applied to the development of a field data collection device for smart failures, and specifically, it can be optimally applied to small-scale smart factories by exploring, developing, and verifying various sensors, and it can also be applied to the development of an OPC-UA information collection device capable of device investigation and field data collection.

In addition, it can be applied to serial communication data collection functions and user interface program development (there are various communication standards, and RS232 is the most common, but it only maintains a transmission distance of up to 50 feet, RS485 is superior to RS232 and supports a transmission distance of up to 5,000 feet), and it can also be applied to selectable GUIs by industry and purpose, development of smart data collection devices and smart sensors for expanding smart factories of small businesses, and development of user interface programs for system operation.

Figure 5 is a block diagram illustrating the configuration of an AI processing unit applied to an AI-based data logger system according to an embodiment.

As illustrated in FIG. 5, the AI processing unit according to one embodiment largely includes a key signal input unit, an I/F unit, a signal processing unit, a main processing unit, a database, and a display unit.

The above key signal input section receives various administrator setting information, etc. used to service the data logger function according to an embodiment of the present invention, through user key operations.

The above I/F section collects various data on factory facilities and situations from a remote field control panel and transmits the data logger information accordingly to the remote registration management information processing device.

The signal processing unit includes a power supply device such as AC or ICP for compatibility with various sensors, a low pass filter for preventing aliasing, and an amplifier for amplifying the input signal. In addition, the signal processing unit further includes a data acquisition device combining an A/D, a D/A converter, and a DSP.

The above main processing unit controls each of the above units, and collects heterogeneous field data in bulk from the field control panel and transmits it to the registration management information processing device. In addition, the main processing unit classifies and processes the heterogeneous field data by different factory facilities and situations to provide a data logger function. In addition, the main processing unit provides optimal field analysis information for each data according to one embodiment, provides a monitoring function for repetitive and standardized errors and their automatic control, and provides a method for pattern analysis and collation analysis monitoring. In addition, it improves data errors that have occurred in advance and provides monitoring and control and fault occurrence prevention alarms. In addition, the main processing unit provides a data analysis solution that detects and responds to process abnormalities in advance through real-time monitoring by connecting various data gathering and communications in a general factory or HACCP factory according to one embodiment, and supports alarm alerts through data analysis. In other words, it collects various data from a general factory or HACCP site and comprehensively collates and analyzes various heterogeneous protocols with a data collation and analysis device to provide a remote monitoring support hub.

The above database stores various registration information (e.g., device and situation, user information, etc.) and data used to service the data logger function according to one embodiment.

The above display unit displays various UIs used to service such data logger functions. In this case, the display unit provides a GUI for monitoring and managing the sensor node and gateway network status. In addition, in this case, the GUI function is implemented in three parts. First, the Connect function is configured as a basic Ethernet with the GUI and gateway connection information settings, and the Sensor Data function implements a periodic upload display of the sensing value transmitted from the sensor module. In addition, the connection data collection function (system) provides a universal serial bus (USB) to connect a PC to a printer, monitor, modem, or peripheral device such as a DAQ.

Figure 6 is a flow chart illustrating the operation of an AI-based data logger system according to an embodiment of the present invention in sequence.

As illustrated in Fig. 6, the AI-based data logger system according to one embodiment is first installed in the field for each of a number of different factory facilities and situations, and is connected to an external sensor that monitors each status and an operating unit that controls it. Accordingly, the field data is detected using the external sensor, the current field situation is monitored, and the operating unit is controlled.

In addition, the AI processing unit collects heterogeneous field data monitored by the control panel in batches (S601), transmits and relays them to the registration management information processing device (S602), and classifies and processes the heterogeneous field data by different factory facilities and situations to provide a data logger function (S605).

Therefore, the management information processing device uses the heterogeneous field data received by the AI processing unit to determine whether an abnormality has occurred for each factory facility and situation, and controls the adjustment operation of the operating unit differently through the corresponding control panel depending on whether an abnormality has occurred (S603-S604).

In this state, the data logger system performs the following operations in the AI processing unit.

a) First, when servicing the above data logger function, heterogeneous field data is gathered in a corresponding communication format for each type of control panel device (including PLC, DCS, OPC, DBMS, and device) according to different factory facilities and situations, thereby processing heterogeneous signals.

b) Next, the heterogeneous field data is compared with the set reference alarm value (S606), and if the heterogeneous field data falls within the reference alarm value as a result of the comparison, it is determined to be normal. On the other hand, if the heterogeneous field data deviates from the reference alarm value, it is determined to be abnormal and the data is analyzed.

In addition, information on the above data analysis is delivered differently to different registration upper systems in different interface formats (including XML, file, and DB access).

In the case where the above is determined to be abnormal, the control information of the operating part is provided to the corresponding control panel in a communication format for each control panel device type according to the current abnormal state (S607).

c) In addition, when the above heterogeneous field data is collected, the pre-alarm modeling value corresponding to the current heterogeneous field data is extracted according to the pre-failure processing format below and compared with the set pre-alarm value.

As a result of the above comparison, if the pre-alert modeling value corresponding to the current heterogeneous field data falls within the pre-alert value, no notification is given, and if the pre-alert modeling value corresponding to the current heterogeneous field data falls outside the pre-alert value, a notification is given to perform pre-failure processing.

And also, in this case, the pre-failure processing format is as follows.

a) First, we classify and learn heterogeneous field data by different control panel device types and monitoring service types, and define a format for modeling standardized and repetitive pre-alarm values.

b) Additionally, datasets for heterogeneous field data from a number of different control panel device types and monitoring service types are extracted at different locations and time zones.

c) Then, the dataset is characterized to reflect the different usage locations, time zones, and ambient condition types that provide the service.

d) Next, the attributes of heterogeneous field data are determined by using the above-described information into the control panel device type and the monitoring service type.

e) Then, normalize the above decision information.

f) Therefore, using the above decision information, pre-alarm values are modeled according to heterogeneous field data by control panel device type and monitoring service type, and each pre-alarm value is set in advance, and independent (pre-alarm modeling value) and dependent (heterogeneous field data) variables are also set.

g) Then, learning and training data are configured using the above setting information.

h) Therefore, a pre-failure processing format is configured through the above-mentioned generation information.

Accordingly, through this, a separate AI processing unit is provided so that an AI-based data logger function can be added in addition to the existing input/output information. In addition, an AI-based data logger system is provided that has the function of checking the data at the time of the problem through the AI processing unit and restoring it to the computer of the central control unit in the event of a power outage or communication failure of the central control unit.

In addition, these data logger systems collect and process signals acquired from various sensors in the AI processing unit, and provide data processing and analysis to enable analysis and diagnosis.

In addition, these data collection and analysis devices solve the problem of complicated measurements and many connecting devices, which are inconvenient to use. Therefore, a signal collection device that acts as a comprehensive protocol gateway that can collect and analyze various signals such as temperature/pressure/noise/vibration by connecting to a portable PC or laptop is provided.

As described above, one embodiment provides a separate AI processing unit so that an AI-based data logger function can be added in addition to existing input/output information. In addition, in the event of a power outage or communication failure in the central control unit, the AI processing unit has a function to check the data at the time of the problem in the future and restore it to the central control unit's computer.

And also, in these cases, the AI processing unit connects various data gathering and communication in general factories or HACCP factories, etc., through real-time monitoring to detect and respond to process abnormalities in advance. In addition, it provides a data analysis solution that can support alarm alerts through data analysis.

That is, it provides a remote monitoring support hub role by collecting various data from general factories or HACCP sites and comprehensively collecting and analyzing heterogeneous protocols using a data collection and analysis tool.

In addition, the AI processing unit provides services for implementing device hardware based on the SBC (Single Board Computing) architecture and collecting/processing/transmitting data from various general-purpose facilities, and provides various I/Fs that can collect data from manufacturing facilities and devices (sensors).

Accordingly, through this, a separate AI processing unit is provided so that an AI-based data logger function can be added in addition to the existing input/output information. In addition, an AI-based data logger system is provided that has the function of checking the data at the time of the problem through the AI processing unit and restoring it to the computer of the central control unit in the event of a power outage or communication failure of the central control unit.

In addition, these data logger systems collect and process signals acquired from various sensors in the AI processing unit, and provide data processing and analysis to enable analysis and diagnosis.

In addition, these data collection and analysis devices solve the problem of complicated measurements and many connecting devices, which are inconvenient to use. Therefore, a signal collection device that acts as a comprehensive protocol gateway that can collect and analyze various signals such as temperature/pressure/noise/vibration by connecting to a portable PC or laptop is provided.

On the one hand, these data logger systems, when servicing these data logger functions, provide slightly better services by matching and maintaining the database from the configuration below and also securing a connection with the administrator mobile terminal, i.e., the administrator device.

To this end, the AI processing unit first performs the following operations to match and maintain the database.

a) That is, when servicing the above data logger function, first, the table storing device registration information/data is identically prepared between the registration control panel or the registration management information processing device or the registration manager mobile terminal, and the matching relationship for the above table is set and registered in advance.

b) Then, when the information in the table is changed in the control panel, the management information processing device, or the manager mobile terminal, the corresponding information in the table of the AI processing unit is also synchronized according to the matching relationship, and the databases are matched with each other.

c) Therefore, when synchronizing the table of the AI processing unit through this, the information is diversified for each data type according to a number of different control panel device types and monitoring service types to match the database.

Next, the AI processing unit also performs the following operations to secure a connection with the administrator device.

a) First, when communicating with such an administrator device, i.e. an administrator mobile terminal, the connection to the registered local communication network is checked first, and if the local communication network is connected as a result of the check, a connection is made as a corresponding setting administrator common account for each administrator location.

b) If the local communication network is not connected as a result of the above verification, the connection to the registered wireless communication network is checked secondarily, and if the wireless communication network is connected as a result of the above verification, the connection is made using an individual IP address.

c) If the above verification result shows that the wireless communication network is not connected, the connection is secured with the administrator mobile terminal by connecting using the terminal identification number of the registered mobile communication network.

In this case, the AI processing unit also connects to a normal manager from the configuration below, thus providing slightly better connection security.

a) That is, first, for this purpose, an IP table that registers the administrator public account of the local communication network and the individual IP addresses of the wireless communication network is configured in advance.

b) Next, when providing information to the above manager mobile terminal, a HELLO message of the corresponding communication network is transmitted and the next hop switch IP address in the response result is extracted.

c) So, check the switch IP address that is identical to the next hop switch IP address in the switch neighbor connection relationship list.

d) If, as a result of the above verification, there is a switch IP address that is identical to the next hop switch IP address, check whether the corresponding administrator public account or individual IP address is also in the IP table to determine whether an unauthorized person is accessing it.

e) Therefore, if the corresponding administrator public account or individual IP address is also in the IP table, a JOIN/PRUNE message is sent to connect to the corresponding communication network.

On the other hand, when these systems monitor in this way, they learn various device and sensor information from the configuration below to suit the surrounding situation and enable smooth monitoring.

In addition, on the one hand, these systems create a monitoring format that takes into account the on-site situation, such as the surrounding conditions, and provide related monitoring and control information according to the on-site situation, thereby effectively processing related monitoring and providing good service.

In such cases, these learning models attribute data related to various locations, time zones, and surrounding conditions, thereby increasing processing rates.

a) First, a learning model for monitoring each control panel device and sensor, etc., is defined to learn status information by classifying it by time zone and location, by installation location, temperature, and humidity, when recommending control information, i.e., advance warning values.

b) Next, extract a basic dataset for a number of different state information.

c) Then, these datasets are characterized by a number of different installation locations, time zones, temperature, humidity, etc.

d) So, based on these attributed results, the attributes of the state information are determined for each of a number of different learning models.

e) After that, normalize the above determined results.

f) Then, based on these normalized results, state information is set for a number of different learning models. Therefore, a number of different state information is set as independent (recommendation information) and dependent (state information) variables to generate information that recommends the surrounding conditions and other field situations.

g) Next, the above-mentioned results are generated as learning and training data.

h) So, through this, we create a format for monitoring from these results.

Therefore, when these systems monitor as described above, they provide recommended information, that is, control information, in real time in a pre-registered format, thereby further enhancing the related services.

Additionally, let me explain a little more about how these formats are generated.

First, these formats generate formats for a number of different devices. Therefore, models can be newly generated for each device manufacturer in various fields, or models can be generated in a few batches based on a standard. This allows the appropriate method to be determined based on the characteristics of the data.

Next, in cases where a large amount of data is not collected due to errors in the device or other factors, or where there are unusually many outliers in reservations, the corresponding data files are removed.

And, in cases where some data is not collected due to data interruption, the corresponding data is removed.

Next, we determine the valid attributes for different formats, generate normal values, and then determine the independent and dependent variables.

Then, to create a format, learning and training data are created from the entire data. Typically, 70% of the entire dataset is used as learning data, and 30% is used as training data to test the model after creating the model.

Next, create a format. At this stage, decide which format to use. For example, it refers to a configuration that configures the input and output layers by configuring the necessary layers based on deep learning and sets the optimal number of outputs. Then, evaluate the model created in this way, and if the error rate of this model is satisfactory, simulate the model with new data, and if the model does not need to be updated, save the format and use it.

Figure 7 is a drawing showing an HMI monitoring screen in a factory applied to an AI-based data logger system according to an embodiment.

As illustrated in FIG. 7, HMI monitoring according to an embodiment collects factory site data accurately and without omission in real time, and is applied to various industries and sectors to provide information verified in the field, and also provides synchronization of production site information events (see a).

In addition, this HMI monitoring also provides a comprehensive interface for various heterogeneous signals, performs monitoring aggregation and management, and provides an HMI linkage interface for batch and immediate production facility status data. In addition, it also provides immediate alarm data for major alarm issues (see b).

100: (on-site) control panel
200: AI processing unit
300: Management information processing device

Claims (4)

A control panel that is installed at each site according to a number of different factory facilities and situations, and is connected to an external sensor that monitors each status and a moving part that controls it, detects on-site data using the external sensors, monitors the current on-site situation, and controls it with the moving part;
An AI processing unit that collects heterogeneous field data monitored by the above control panel in batches, transmits and relays it to a registration management information processing device, and classifies and processes the heterogeneous field data by different factory facilities and situations to provide a data logger function; and
It includes a management information processing device that determines whether an abnormality has occurred for each factory facility and situation using heterogeneous field data transmitted by the above AI processing unit, and controls the adjustment operation of the operating unit differently through the corresponding control panel depending on whether an abnormality has occurred.

The above AI processing unit,
a) When servicing the above data logger function, heterogeneous field data is gathered in a communication format corresponding to each type of control panel device according to different factory facilities and situations, thereby processing heterogeneous signals.
b) Compare the heterogeneous field data with the set standard alarm value, determine that the heterogeneous field data is normal if the comparison result falls within the standard alarm value, and analyze the data by determining that the heterogeneous field data is abnormal if the heterogeneous field data falls outside the standard alarm value.
In addition, information on the above data analysis is transmitted differently to each different registration upper system in different interface formats,
In the case of determining the above, the control information of the operating unit is provided to the corresponding control panel in a communication format for each control panel device type according to the current abnormal state.
c) When the above heterogeneous field data is collected, the pre-alarm modeling value corresponding to the current heterogeneous field data is extracted according to the pre-failure processing format below and compared with the set pre-alarm value.
If the pre-alert modeling value corresponding to the current heterogeneous field data as a result of the above comparison falls within the pre-alert value, no notification is given, and if the pre-alert modeling value corresponding to the current heterogeneous field data falls outside the pre-alert value, a notification is given to perform pre-failure processing.
And also, the above pre-failure processing format is,
a) First, we classify and learn heterogeneous field data by different control panel device types and monitoring service types, and define a format for modeling standardized and repetitive pre-alarm values.
b) In addition, data sets for heterogeneous field data by different types of control panel devices and types of monitoring services are extracted at different locations and time zones, respectively.
c) attribute the above dataset to reflect the different usage locations, time zones, and ambient condition types that provide the service;
d) Using the above-described information, determine the properties of heterogeneous field data into control panel device types and monitoring service types;
e) normalize the above decision information,
f) Using the above decision information, model the pre-alarm values according to heterogeneous field data by control panel device type and monitoring service type, and set each pre-alarm value in advance, and also set independent and dependent variables.
g) Generate learning and training data using the above setting information,
h) An AI-based data logger system characterized by configuring a pre-failure processing format using the above-mentioned generated information. ◈Claim 2 was abandoned upon payment of the registration fee.◈ In claim 1,
The above AI processing unit
a) When servicing the above data logger function, the same table storing device registration information/data is prepared between the registration control panel or the registration management information processing device or the registration manager mobile terminal, and the matching relationship for the above table is set and registered in advance.
b) When information in a table is changed in the control panel, management information processing device, or manager mobile terminal, the corresponding information is synchronized in the table of the AI processing unit according to the matching relationship, and the databases are matched with each other.
c) An AI-based data logger system characterized in that, when synchronizing the table of the AI processing unit, the database is matched by diversifying information by data type according to a plurality of different control panel device types and monitoring service types. ◈Claim 3 was waived upon payment of the registration fee.◈ In claim 2,
The above AI processing unit
a) When communicating with the above administrator mobile terminal, first, check whether the registered local communication network is connected, and if the local communication network is connected as a result of the above check, connect as a corresponding setting administrator common account for each administrator location.
b) If the local communication network is not connected as a result of the above verification, the connection to the registered wireless communication network is checked secondarily, and if the wireless communication network is connected as a result of the above verification, the connection is made using an individual IP address.
c) An AI-based data logger system characterized in that, if the wireless communication network is not connected as a result of the above verification, the connection is secured with the administrator mobile terminal by connecting using the terminal identification number of the registered mobile communication network; ◈Claim 4 was waived upon payment of the registration fee.◈ In claim 3,
The above AI processing unit
a) Pre-configure an IP table that registers the administrator's common account of the above local communication network and the individual IP addresses of the above wireless communication network,
b) When providing information to the above manager mobile terminal, a HELLO message of the corresponding communication network is transmitted and the next hop switch IP address in the response result is extracted,
c) Check the switch IP address that is identical to the next hop switch IP address in the switch neighbor connection relationship list.
d) If, as a result of the above verification, there is a switch IP address that is identical to the next hop switch IP address, check whether the corresponding administrator public account or individual IP address is also in the IP table, thereby checking whether an unauthorized person is accessing it.
e) An AI-based data logger system characterized in that, if the corresponding administrator public account or individual IP address is also in the IP table as a result of the above verification, a JOIN/PRUNE message is sent, thereby connecting to the corresponding communication network;

Images