KR102402653B1 - AI-based SCADA system and operating method thereof - Google Patents

Abstract

The present invention relates to an artificial intelligence (AI)-based supervisory control and data acquisition (SCADA) system implemented to monitor system operation integrity and human operation errors and an operation method thereof. According to the present invention, a plurality of remote terminal devices are connected to each facility to be monitored, collect data of each facility to be monitored, and transmit the collected data. A communication control device collects data from the remote terminal device and performing data preprocessing with respect to the collected data to reduce time required for subsequent data processing to transmit the preprocessed data. A central server collects the preprocessing data from the communication control device to process and store the collected preprocessed data, monitor abnormal data, and transmits the processed data, and perform monitoring of operational integrity related to systemic factors and human operational errors caused by human factors on the basis of AI. A human-machine interface (HMI) displays processing data received from the central server on a display screen and performs system control through the central server according to an operator's operation command.

Description

AI-based SCADA system and operating method thereof

The technical field of the present invention relates to an AI-based SCADA system and an operating method thereof, and more particularly, to an AI-based SCADA system implemented to enable monitoring of system operation integrity and human operation errors and an operating method thereof.

The industry is becoming increasingly large-capacity and complex with the rapid development of an industrial society. As a method of artificially and manually managing industrial facilities and on-site management, efficient management has reached the limit. As computer and information and communication technologies develop rapidly, information collection, processing analysis, and facility control are rapidly increasing. By applying and applying computer and information communication technology to operation, the SCADA (Supervisory Control And Data Acquisition) system is more necessary than ever for the purpose of efficient and economical on-site power supply control and facility operation.

The SCADA system collects, receives, records, and displays the status information data of the remote device to a remote terminal unit using analog or digital signals on the communication path, so that the central control system monitors and controls the remote device. As a system, it is a system that centrally monitors and controls various types of remote facility devices such as power generation, transmission and distribution facilities, petrochemical plants, steel processing facilities, and factory automation facilities. Such a SCADA system, called an industrial control system, is a computer system for monitoring and controlling a work process based on industrial processes and equipment.

In the SCADA system, various communication protocols such as MODBUS, Profibus, and DNP have been conventionally used to monitor and control numerous remote devices, and standardized protocols (ie, IEC61850, IEC60870) are being applied in recent years. Since various remote devices that interoperate with the SCADA system apply standardized protocols, interoperability and operability tests between each device and system are essential. Here, IEC 61850 is an international standard for a power facility communication interface, and is an international standard in the field of power system communication to ensure smooth and reliable activity in a network. And the IEC 60870 standard defines systems used for remote control, which are used to control power transmission grids and other geographically prevalent control systems. And expensive switches that provide standardized protocols are used to test and verify these protocols.

Korea Patent Registration No. 10-1906191 (registered on October 2, 2018) discloses a SCADA system and an operating method thereof, including a remote terminal for measuring state information of field devices; a master server for collecting status information from remote terminals; It includes a remote camera for collecting an image of a remote terminal according to a control signal transmitted from a master server, the master server includes: an image collecting unit for collecting an image of the remote terminal from the remote camera; an extractor for extracting image feature information from an image; a search unit searching for remote terminal information matching the image feature information extracted from the remote terminal DB; By extracting feature points corresponding to the image elements from the image using tag information for each image element included in the remote terminal information, a preset number of feature points are continuously extracted from a plurality of frames included in the image, and the extraction position of the feature points is constant When a background tag is attached to an area including the keypoint, if a preset number of keypoints are successively extracted from a plurality of frames included in the image and the extraction position of the keypoint is not constant, a data tag is applied to an area including the keypoint a tag attachment unit to be attached; and a status data generator for generating status data of the remote terminal by using the image collection time, location information of the background tag and data tag, and tag information of a background tag and data tag corresponding to the location information, the background tag and data tag The tag information of is identification information describing the contents of an image element, and is characterized by being learned for each image element through machine learning. According to the disclosed technology, it is possible to effectively monitor the status information data of the passive remote terminal in real time, and also, when the connection between the important remote terminal and the server fails, the information displayed on the remote terminal is received without loss, thereby preventing the loss of the status information data. It can be minimized and the fault tolerance can be strengthened.

Korean Patent Registration No. 10-1906192 (registered on October 2, 2018) discloses a SCADA system capable of monitoring status information of a field device measured by a remote terminal without loss even in an emergency situation and an operating method thereof. According to the disclosed technology, a remote terminal that measures and transmits status information of field devices to the master server, a remote camera installed near the remote terminal, a master server that receives status information from a remote terminal, and a method for generating status information of field devices in case of emergency In the method of operating a SCADA system including a state information generating device, if the state information is not received from the remote terminal for a preset time, the master server transmits a control signal for controlling the remote camera to take an image of the remote terminal to do; acquiring, by the remote camera receiving the control signal, an image of the remote terminal and transmitting the image to the state information generating device; The state information generating apparatus may include: extracting image characteristic information from an image; The state information generating apparatus may include: searching for remote terminal information matching the image feature information extracted from the database; The state information generating apparatus may include: extracting a feature point corresponding to the image element from the image by using the image element included in the remote terminal information; attaching a background tag to the feature points if the feature points are continuously extracted from a plurality of frames included in the image by a preset number or more and the extraction positions of the feature points are constant; attaching a data tag to the feature points when the feature points are continuously extracted from a plurality of frames included in the image by a preset number or more and the extraction positions of the feature points are not constant; The apparatus for generating state information may include: identifying an image element corresponding to location information of a data tag in remote terminal information; The state information generating apparatus includes: inserting a measurement value into a data tag using tag information that is identification information describing the identified image element; The state information generating apparatus is characterized in that it includes the step of generating image state information of the field device by using the reception time for each frame and the measurement value inserted in the data tag.

1 is a view for explaining a conventional SCADA system.

Referring to FIG. 1, the conventional SCADA system is composed of many equipment and devices such as a terminal RTU, an FEP, a plurality of servers, a remote monitoring control system, and a fire controller. However, since it is a combination of several devices, it takes a lot of time, effort, and money to find the device that occurs when a problem occurs. There are also disadvantages in that it takes a lot of time and money to solve a problem because it takes a lot of time and money to solve the problem. In addition, when the operator is inexperienced and exercises wrong control, there is a high risk of creating a big threat to public health and safety. there was

In the prior art as described above, the protocol between the SCADA system and the intelligent electronic device (IED), the sensor, equipment or device, and the center is different, and it is prepared for an abnormal situation that may occur intermittently in the existing monitoring control and data collection operation. If an error occurs even after performing an interoperability test, the SCADA system stops or malfunctions for a long time, and a lot of time and manpower is invested in the process of finding the cause, and nevertheless, the exact cause of the abnormal situation may not be found. There was a downside. In particular, system problems caused by systemic causes such as defects or failures in hardware, software, and data of the system, and human causes such as inexperience of operation in which the operator fails to execute necessary measures or performs wrong operations It also had a limitation in that the SCADA system could not distinguish operational errors.

Korean Patent No. 10-1906191 Korean Patent Registration No. 10-1906192

The problem to be solved by the present invention is to solve the above-described shortcomings, and to provide an AI-based SCADA system implemented to enable monitoring of system operation integrity and human operation errors, and an operating method thereof.

As a means for solving the above problem, according to one aspect of the present invention, a plurality of remote terminal devices that are connected to each monitoring target facility, collect data of each monitoring target facility, and transmit the collected data; a communication control device that collects data from the remote terminal device and transmits the pre-processed data by performing pre-data processing on the collected data to reduce a time required for a subsequent data processing process; It collects pre-processing data from the communication control device, processes the collected pre-processing data, monitors abnormal data, stores data, and transmits operator control data received from HMI to the communication control device. Based on artificial intelligence that is learned and relearned from various data transmitted, received and stored through the SCADA unit, operational integrity related to systemic factors such as data loss due to device failure and errors, and incorrect operation commands due to inexperienced operation of the operator a central server composed of an artificial intelligence unit that monitors human operation errors caused by the same human factors; And it provides an AI-based SCADA system including an HMI for displaying the processing data from the central server on a display screen, and transmitting a system control command to the central server according to an operator's operating command.

In one embodiment, the remote terminal device, RTU for enabling mutual control with the equipment to be monitored; and an SDP for applying a plurality of communication methods and communication protocols to the RTU.

In an embodiment, the remote terminal device collects field data from a monitoring target facility, performs data processing, monitoring and control functions, and then transmits the data to the communication control device through a communication line.

In an embodiment, the remote terminal device is characterized in that it collects field data in conjunction with a power monitoring target facility, a power quality target facility, a lighting monitoring target facility, and a preventive diagnosis facility.

In an embodiment, the communication control device integrates, refines, reduces, and transforms data collected from the remote terminal device into data for assembling and dismantling packets, error detection and correction through preprocessing software, preprocessing operations It is characterized in that it comprises an FEP for transmitting the pre-processed data to the central server by performing the .

In an embodiment, the communication control device is configured in duplication of an active device and a standby device, and provides a continuous and continuous service by switching to a standby device according to an abnormality check result. characterized.

In one embodiment, the central server is configured as a duplication of a server in an active state and a server in a standby state, and provides continuous and continuous service by switching to a server in a standby state according to an abnormality check result do it with

In one embodiment, the central server learns the system abnormality prediction model with learning software, performs pre-processing and post-processing to form a model that is served in real time, improves the model with re-learning software, and through expert knowledge service Based on artificial intelligence that is learned in a way that has an infinite evolutionary structure that repeatedly performs transmission of the based reactive response to the operator by applying it to a predictive model with expert-level judgment and predicted knowledge, It is characterized by monitoring for operational integrity related to system elements of software and data, and human operation errors caused by human elements inexperienced in operation of the operator.

In one embodiment, the central server, the processing and storage of the pre-processing data collected from the communication control device, along with abnormal data monitoring and system control, along with the operator to monitor the current data to transmit the data; and an artificial intelligence unit for classifying, inferring, and monitoring the operational integrity of the system and the operator's operational errors based on artificial intelligence by utilizing all the stored data.

In one embodiment, the SCADA unit, the communication control device, the HMI, and the communication unit for transmitting and receiving data with the artificial intelligence unit; a determination unit for determining data to determine an on/off event of DI and an event of upper or lower limit of AI; a storage unit for storing alarm data and trends, reports and input data; a redundancy unit for checking data synchronization and program, communication abnormality, and DBMS abnormality, and performing redundancy switching according to the abnormality check result; It is characterized by having a module support unit for supporting each module management and utility.

In one embodiment, the artificial intelligence unit, a data collection / loading interworking platform for data collection and loading with a set of communication protocols and database processes; A data processing/model learning interworking platform for data processing and model learning as a set of processes for data processing and model learning required for artificial intelligence learning; A server interworking platform for interplatform interworking and artificial intelligence model service execution as a set of processes for interplatform interworking and artificial intelligence model service execution; a client infrastructure for allowing a client to use the AI unit as a set of processes for AI monitoring; a core for controlling the platform with a set of core processes; a domain platform for managing internal domains and domains of other servers as a set of processes for managing internal domains and domains of other servers; a core infrastructure for managing data processed by the data processing/model learning interworking platform as a set of processes for managing data; and a security interworking platform for in charge of security with a set of security-related processes.

In one embodiment, the artificial intelligence unit collects learning data and loads an input dataset for artificial intelligence learning, performs a preprocessing process of data integration, purification, reduction, and transformation, and uses this as artificial intelligence learning software. It is characterized in that a model of prediction, judgment, and response of abnormal causes is newly created or modified according to the learning result by learning.

In one embodiment, the artificial intelligence unit, but performs monitoring of the equipment integrity through the HMI; The data integrity is checked by monitoring the signals and data from the central server, the HMI, and FEP, and when there is no data reception through the FEP, the SCADA system is monitored as a whole to diagnose FEP module abnormalities or RTU failures. It is characterized in that it is presented to the operator.

In one embodiment, the artificial intelligence unit, but performs monitoring of the equipment integrity through the HMI; There is no problem in data transmission within the SCADA system, but when an error in data processing occurs, the operator is not aware of it, it is characterized by diagnosing the location of data processing and displaying a notification to the operator.

In one embodiment, the artificial intelligence unit, but performs monitoring for human error through the HMI; In the case of learning the values of AI data associated with a specific facility at the time of input, a notification is displayed to the operator if the AI data is 0 at the time of input of the specific facility or data within the normal range is not expressed.

In one embodiment, the artificial intelligence unit, but performs monitoring for human error through the HMI; It is characterized in that it learns the generation of related DI and AI events when a specific facility is turned on or off, and displays a notification to the operator when a preset time elapses when the specific facility is turned on or off.

In one embodiment, the artificial intelligence unit, but performs monitoring for human error through the HMI; It is characterized in that by learning the actions of the operator after the occurrence of the event, when the same event occurs, a notification is displayed to the operator when the operator proceeds with a different action than before.

In an embodiment, the artificial intelligence unit generates a model when starting the system and performs supervised learning, but performs model supervised learning through a dataset; performing unsupervised learning by extracting and serving the supervised model in the first step, and performing model unsupervised learning through operational data and datasets based on the model extracted and served in the first step; Perform automatic machine learning by extracting and serving the unsupervised model in two steps; Performs automatic machine learning through operation data and datasets for each preset cycle based on the two-step extraction and serving model; performing expert knowledge learning by extracting and serving the machine-learned model in three steps, and performing expert knowledge learning by using an expert knowledge service based on the three-step extraction and serving model; After extracting and serving the expert knowledge-learned model in four steps, it is characterized in that the process of performing supervised learning again based on the four-step extraction and serving model is repeatedly performed.

In one embodiment, the artificial intelligence unit, but performs the first step of predicting the system abnormal situation; To develop an anomaly prediction model, it receives big data in a file format that can be recognized in artificial intelligence learning, such as Excel or image file, from normal and abnormal sensor data or labeled dataset acquired in advance, and uses artificial intelligence learning software to predict anomalies. It is characterized in that, in the case of the learned model, after verification through model verification, the learned model is distributed, and the distributed model is served to transmit an abnormal pattern notification to the HMI.

In one embodiment, the artificial intelligence unit, but performs the second step of predicting the system abnormal situation; The data observed in the SCADA system or facility is collected through streaming technology, the data extracted through the purification process is saved in the format of a file used by the AI learning software, and the extracted data stored in the AI learning software is read again It is characterized by forming a model that is pre-processed with data usable in the AI learning model and served in real time so that the operator can check the results predicted by the AI learning model.

In one embodiment, the artificial intelligence unit performs a system abnormal situation pre-prediction 3 step of adding re-learning software such as automatic machine learning software, but with respect to the operator's manual model learning and creation process, daily, weekly or monthly It is characterized in that it generates a model by self-learning using the extracted data set for each cycle designated by the operator, and distributes the generated model through self-verification and comparison with previous models.

In one embodiment, the artificial intelligence unit performs the 4 steps of predicting system anomalies in advance, but through a remote advisory system in which each relevant expert inquires event data and establishes an expert opinion on the data, a SCADA system expert The level of judgment and predictive knowledge is applied to the predictive model, and the post-response of the applied predictive model is transmitted to the HMI, and the predictive model of the 4th stage of the system abnormal situation prediction is again the initial prediction of the first system abnormal situation preliminary prediction stage It is characterized in that it is used as a model.

In one embodiment, the artificial intelligence unit collects data of the normal operation and operating state of the monitoring target facility or SCADA system, unsupervised or supervised learning the normal pattern of the normal operation and operating state, and deviates from the reference value. When an abnormality occurs, it is predicted as one of the causes of systemic operational integrity and human operation error and displayed to the operator through the HMI. It is characterized in that this is done.

In one embodiment, the artificial intelligence unit monitors the occurrence of an abnormal pattern before the occurrence of an abnormality or error, predicts the possibility of an abnormality in advance and displays it to the operator through the HMI, and the past normal operation of the SCADA system And it is characterized in that the data on the abnormal situation is learned in advance to determine and respond to the cause of the abnormal occurrence of the equipment to be monitored and the SCADA system from the initial stage when the system is installed.

In one embodiment, the HMI may include: an editing unit for editing DI, AI, DO, AO data and single-line diagrams and reports; a communication unit for transmitting and receiving data to and from the central server; Utility department in charge of utility programs of event search, trend search, TAG support, and report registration; and a monitoring unit for displaying DI, AI, DO, and AO data as symbols, pictures, and characters, and for outputting event data and an alarm sound.

In an embodiment, the HMI is characterized in that the overall schematic, control panel screen, PDS screen, communication, device abnormality screen, power quality screen, and fault point expression panel screen are displayed on the display.

As a means for solving the above problems, according to another feature of the present invention, a plurality of remote terminal devices are connected to each of the monitoring target facilities, collect data of each monitoring target facility, and transmit the collected data. step; transmitting the pre-processed data by a communication control device collecting data from the remote terminal device, performing pre-data processing on the collected data to reduce the time required for a subsequent data processing process; The central server collects pre-processing data from the communication control device and transmits the data processing, abnormal data monitoring and data storage, and operator control data received from the HMI to the communication control device; Based on artificial intelligence, monitoring for operational integrity related to systemic factors, such as data loss due to device failure and errors, and human operation errors caused by human factors, such as erroneous operating instructions due to inexperienced operation of the operator ; And HMI displays the processing data from the central server on a display screen, and provides an artificial intelligence-based SCADA system operating method comprising the step of performing system control through the central server according to an operator's operation command.

As an effect of the present invention, by providing an artificial intelligence-based SCADA system implemented to enable monitoring of system operation integrity and human operation errors and an operation method thereof, the artificial intelligence-based artificial intelligence unit of the operator console is a SCADA system system It is possible to separate and monitor operational integrity and human operational errors, and thus, when a SCADA system error occurs, the cause can be clearly identified.

According to the present invention, it is possible to perform unsupervised learning of normal operation and operation patterns by collecting data of the monitoring target facility and SCADA system in normal operation and operation state, using the artificial intelligence unit of the central server, and When an abnormality that deviates from the value occurs, an abnormal pattern different from the related normal pattern can be traced and displayed as the cause. As supervised learning is performed, it also has the effect of improving the accuracy of the cause analysis and judgment of anomalies.

According to the present invention, by utilizing the artificial intelligence unit of the central server for the monitoring of facilities and SCADA systems, the amount of manpower and time input to find the cause of the abnormality can be greatly reduced, and the accuracy of the determination and response of the cause It also has the effect of continuously increasing the value close to 100%.

According to the present invention, by monitoring the occurrence of the abnormal pattern before the occurrence of the abnormality, the possibility of the occurrence of the abnormality can be predicted in advance and displayed to the operator. By learning the data on the situation in advance, it has the effect of increasing the accuracy of the determination and response to the cause of the abnormal occurrence of the facility and the SCADA system from the initial stage when the system is installed.

1 is a view for explaining a conventional SCADA system.
2 is a diagram illustrating an AI-based SCADA system according to an embodiment of the present invention.
FIG. 3 is a diagram for explaining the RTU in FIG. 2 .
FIG. 4 is a view for explaining the scada unit in FIG. 2 .
FIG. 5 is a view for explaining the artificial intelligence unit shown in FIG. 2 .
6 is a view for explaining the learning of the model of the artificial intelligence unit shown in FIG.
FIG. 7 is a view for explaining the first step of predicting the system abnormal situation of the artificial intelligence unit in FIG. 2 .
FIG. 8 is a view for explaining the second step of predicting system abnormality in the artificial intelligence unit in FIG. 2 .
9 is a view for explaining the third step of predicting the system abnormal situation of the artificial intelligence unit in FIG. 2 .
10 is a view for explaining the fourth step of predicting the system abnormal situation in the artificial intelligence unit in FIG. 2 .
FIG. 11 is a diagram for explaining the HMI in FIG. 2 .
12 is a view for explaining the monitoring unit of FIG. 11 .
13 is a view for explaining an AI-based SCADA system operating method according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those of ordinary skill in the art to which the present invention pertains can easily implement them. However, since the description of the present invention is merely an embodiment for structural or functional description, the scope of the present invention should not be construed as being limited by the embodiment described in the text. That is, since the embodiment is capable of various changes and may have various forms, it should be understood that the scope of the present invention includes equivalents capable of realizing the technical idea. In addition, since the object or effect presented in the present invention does not mean that a specific embodiment should include all of them or only such effects, it should not be understood that the scope of the present invention is limited thereby.

The meaning of the terms described in the present invention should be understood as follows.

Terms such as “first” and “second” are for distinguishing one component from another, and the scope of rights should not be limited by these terms. For example, a first component may be termed a second component, and similarly, a second component may also be termed a first component. When a component is referred to as being “connected to” another component, it may be directly connected to the other component, but it should be understood that other components may exist in between. On the other hand, when it is mentioned that a certain element is "directly connected" to another element, it should be understood that the other element does not exist in the middle. Meanwhile, other expressions describing the relationship between elements, that is, "between" and "between" or "neighboring to" and "directly adjacent to", etc., should be interpreted similarly.

The singular expression is to be understood to include the plural expression unless the context clearly dictates otherwise, and terms such as "comprise" or "have" are not intended to refer to the specified feature, number, step, action, component, part or any of them. It is intended to indicate that a combination exists, and it should be understood that it does not preclude the possibility of the existence or addition of one or more other features or numbers, steps, operations, components, parts, or combinations thereof.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs, unless otherwise defined. Terms defined in general used in the dictionary should be interpreted as having the same meaning in the context of the related art, and cannot be interpreted as having an ideal or excessively formal meaning unless explicitly defined in the present invention.

Now, an AI-based SCADA system and an operating method thereof according to an embodiment of the present invention will be described in detail with reference to the drawings.

2 is a view for explaining an AI-based SCADA system according to an embodiment of the present invention, FIG. 3 is a view for explaining the RTU in FIG. 2, and FIG. 4 is a view for explaining the SCADA part in FIG. , FIG. 5 is a view for explaining the artificial intelligence unit in FIG. 2, FIG. 6 is a view for explaining the learning of the model of the artificial intelligence unit in FIG. 2, and FIG. 7 is a system abnormal situation prediction of the artificial intelligence unit in FIG. It is a diagram for explaining the first step, and FIG. 8 is a diagram for explaining the second step of predicting the system abnormal situation of the artificial intelligence unit in FIG. 2, and FIG. FIG. 10 is a diagram for explaining the 4th step of predicting the system abnormal situation of the artificial intelligence unit in FIG. 2, FIG. 11 is a diagram for explaining the HMI in FIG. 2, and FIG. 12 is the monitoring unit in FIG. It is an explanatory drawing.

2 to 12 , the AI-based SCADA system 100 includes a plurality of remote terminal devices 110 , a communication control device 120 , a central server 130 , and an HMI 140 .

The remote terminal device 110 is connected to a monitoring target facility of a facility system in a remote location, collects data of the monitoring target facility, and converts the data into a format that can be transmitted to the communication control device 120 side and transmits the converted data to the communication control device 120 .

In one embodiment, the remote terminal device 110, RTU (Remote Terminal Unit) 111 for enabling mutual control with the equipment to be monitored; In addition, a smart digital processor (SDP) 112 for applying a plurality of various communication methods and communication protocols to the RTU 111 may be provided.

In one embodiment, the remote terminal device 110 collects field data from a monitoring target facility (ie, field devices (eg, thermometer or hydraulic meter, etc.) and performs data processing, monitoring and control functions, The collected field data may be transmitted to the upper central server 130 through the communication control device 120 through the communication line.

In one embodiment, the RTU 111 is, as shown in FIG. 3, a plurality of field devices or equipment (eg, power monitoring target facility, power quality target facility, lighting monitoring target facility, preventive diagnosis facility ( For example, in conjunction with GIS/LA, transformer, etc.), field data of field devices or equipment may be collected and transmitted to the communication control device 120 .

The communication control device 120 collects data transmitted from the remote terminal device 110 , and performs pre-processing on the collected data to reduce the time required for subsequent data processing to reduce the pre-processed data. is transmitted to the central server 130 .

In one embodiment, the communication control device 120, for the field data transmitted from the remote terminal device 110 (ie, the RTU 111), the assembly and disassembly of the packet through pre-processing software, error detection and correction, etc. In order to reduce the time required for processing the pre-processing data in the central server 130 by performing pre-processing operations such as integration, refining, reduction, and transformation into the data for pre-processing and transmitting it to the central server 130 A front-end processor (FEP) for Here, the pre-processing software may be included in the SCADA server application or included in the SCADA client application.

In an embodiment, the communication control device 120 is configured as a duplication of an active device and a standby device, and performs switching to a standby device according to an abnormality check result to provide continuous and continuous service. can be provided.

The central server 130 collects the pre-processing data transmitted from the communication control device 120 , processes the collected pre-processing data, stores the processed data, and transmits it to the HMI 140 .

In one embodiment, the central server 130 is configured in duplication of a server in an active state and a server in a standby state, and provides continuous and continuous service by switching to a server in a standby state according to the abnormality check result. can make it

In one embodiment, the central server 130 improves the prediction model in four steps, learns it, and repeats it to increase the accuracy and efficiency of prediction; In order to discover and respond to the causes of abnormalities occurring in the operation and operation of the SCADA system, which centrally monitors and controls the state data of major infrastructure such as electricity, railroads, and water supply, an anomaly prediction model with artificial intelligence learning software as the first step In step 2, pre-processing and post-processing are performed to form a model that is served in real time, in step 3, the model is improved with re-learning software, and in step 4, expert-level judgment and predicted knowledge through expert knowledge service. Based on artificial intelligence that is learned in such a way that it is applied to a predictive model with Operational integrity related to system elements such as hardware, software, and data of the system, and human operation errors caused by human factors such as operator inexperienced operation can be monitored.

In one embodiment, the central server 130 processes the pre-processing data collected from the communication control device 120 and transmits the data to the HMI 140 for monitoring and storing abnormal data, and monitoring the operator's system, the HMI SCADA unit 131 for transmitting and storing the control data received through the communication control device; In addition, an artificial intelligence unit 132 for classifying and monitoring the operational integrity of the system and human operation errors based on artificial intelligence by utilizing all the stored data may be provided.

In one embodiment, the scan unit 131, as shown in FIG. 4, includes a communication unit 131a, a determination unit 131b, a storage unit 131c, a redundancy unit 131d, and a module support unit 131e. Thus, the received data is processed, and the processed data is transmitted to the artificial intelligence unit 132 and the HMI 140 (ie, the operator console 141) or receives control data input from the HMI 140 and communicates. can be sent to the control unit.

The communication unit 131a transmits/receives data to and from the communication control device 120 , the HMI 140 (ie, the operator console 141 ), and the artificial intelligence unit 132 .

The determination unit 131b determines data to determine an on/off event of a digital input (DI) and an event of an upper limit or a lower limit of an analog input (AI).

The storage unit 131c stores alarm data, trends, reports (hourly reports, daily reports, weekly reports, monthly reports, etc.) and input data.

The duplication unit 131d checks data synchronization, program, communication error, DBMS error, and the like, and performs duplication switching according to the corresponding error check result.

The module support unit 131e supports each module management and utility.

In one embodiment, the artificial intelligence unit 132, as shown in Figure 5, data collection / loading interworking platform (132a), data processing / model learning interworking platform (132b), server interworking platform (132c), client A function to classify and monitor the operational integrity of the system and human operation errors based on artificial intelligence by providing the infrastructure 132d, the core 132e, the domain platform 132f, the core infrastructure 132g, and the security interworking platform 132h can be performed.

The data collection/load interworking platform 132a is a set of communication protocols and database processes, and performs data interworking for data collection and loading.

The data processing/model learning interworking platform 132b is a set of processes for data processing and AI model learning required for AI model learning, and performs data interworking for data processing and model learning.

The server interworking platform 132c is a set of processes for interworking between platforms within the artificial intelligence unit 132 and for service execution of the artificial intelligence model, and interworking between internal platforms and server interworking for service execution of the artificial intelligence model. carry out

The client infrastructure 132d is a set of processes for artificial intelligence monitoring for enabling the client to use the artificial intelligence unit 132 , and allows the client to use the artificial intelligence unit 132 .

The core 132e is a set of core processes of the artificial intelligence unit 132 and controls the platform inside the artificial intelligence unit 132 .

The domain platform 132f is a set of processes for managing the domains inside the artificial intelligence unit 132 and domains of other servers, and manages the internal domains and domains of other servers.

The core infrastructure 132g is a set of processes for managing data processed by the data processing/model learning interworking platform 132b, and manages data processed by the data processing/model learning interworking platform 132b.

The security interworking platform 132h is a set of security-related processes of the artificial intelligence unit 132 and takes charge of the security of the artificial intelligence unit 132 .

In one embodiment, the artificial intelligence unit 132 requires the collection and preprocessing of learning data (data integration, purification, reduction, transformation, etc.) for learning artificial intelligence, and learns it as an artificial intelligence learning model. According to the learning result, it is possible to newly create or modify a model for predicting, judging, and responding to an abnormal cause.

In one embodiment, the artificial intelligence unit 132, but performs monitoring of the facility integrity through the HMI (140); The data integrity can be checked by monitoring the signals and data from the central server 130 and the HMI 140 (that is, the operator console 141), the FEP, and when there is no data reception through the FEP, the SCADA system By monitoring the whole, it is possible to present to the operator through the HMI 140 an abnormality of the module of the FEP or a diagnosis of a failure of the lower terminal device 110 such as the RTU 111, and there is no abnormality in data transmission within the SCADA system. , when the operator does not recognize an error in data processing, it is possible to diagnose the location of data processing and display a notification to the operator through the HMI 140 .

In one embodiment, the artificial intelligence unit 132, but performs monitoring for human error through the HMI (140); If the AI data values associated with a specific facility are learned at the time when a specific facility is turned on, if the AI data is 0 or the data within the normal range is not displayed when the specific facility is put on, a notification is sent to the operator through the HMI 140 In addition, it learns to generate related DI and AI events when a specific facility is turned on or off, and when a preset time elapses when a specific facility is turned on or off, a notification can be displayed to the operator through the HMI 140. Moreover, by learning the actions of the operator after the event occurs, when the same event occurs, when the operator proceeds with actions different from before, a notification may be displayed to the operator through the HMI 140 .

In one embodiment, the artificial intelligence unit 132 learns a normal pattern, for example, in which the manager of a specific facility repeats the procedure of applying power to a specific facility around 9 am on weekdays and turning off the power around 6 pm on weekdays. In this case, when a certain period of time has elapsed without power being applied to a specific facility at around 9:00 am, the possibility of human operation error may be displayed as a notification to the operator of the SCADA system through the HMI 140 .

In one embodiment, the artificial intelligence unit 132 performs supervised learning in a first step by generating a learning model when starting the system, as shown in FIG. 6 , when performing artificial intelligence learning, but through model supervised learning; Extracting and serving the supervised model to perform unsupervised learning in two steps, but performing model unsupervised learning through operational data and datasets based on the model extracted and served in the first step; Extract and serve the unsupervised model to perform automatic machine learning in three steps; Performs automatic machine learning through operation data and datasets for each preset cycle based on the two-step extraction and serving model; extracting and serving the machine-learned model to perform expert knowledge learning in four steps, and performing expert knowledge learning by using an expert knowledge service based on the three-step extraction and serving model; After extracting and serving the model learned by the expert knowledge, the process of performing the first-stage supervised learning again based on the fourth-stage extraction and serving model can be infinitely repeated.

In one embodiment, the artificial intelligence unit 132, as shown in FIG. 7, performs the first step of predicting system anomalies in advance, normal and abnormal sensor data or labeled data acquired in advance to develop an anomaly prediction model The set receives big data in a file format that can be recognized in AI learning, such as Excel or image files, and learns an anomaly prediction model with artificial intelligence learning software. By distributing the model, serving the distributed model, and sending an abnormal pattern (eg, when the temperature or oil pressure of the facility is different from usual) to the HMI 140, notification of the abnormal pattern is sent to the operator through the HMI 140 can be displayed to

In one embodiment, the artificial intelligence unit 132, as shown in FIG. 8, performs the second step of predicting system anomalies in advance, but in order to supplement the first step of predicting system anomalies in advance, in a SCADA system or facility The observed data is collected through streaming technology, the data extracted through the refining process is saved in the format of a file used by the AI learning software, and the extracted data stored in the AI learning software is read again to be used in the AI learning model. By preprocessing data that can be used and forming a model that is served in real time so that the operator can check the results predicted by the artificial intelligence learning model, it is possible to increase the judgment and predictability of abnormal situations.

In one embodiment, the artificial intelligence unit 132, as shown in FIG. 9, performs the 3rd step of predicting system anomalies in advance, and to supplement the 2nd step of predicting system abnormalities in advance, such as automatic machine learning software In addition to re-learning software, for the operator's passive model learning and creation process in the second stage of predicting system anomalies, the model is developed by self-learning using the extracted data set by the operator-specified period by day, week, or month. By creating and distributing the generated model as a better model through self-verification and comparison with the previous model, judgment and predictability of abnormal situations can be increased.

In one embodiment, the artificial intelligence unit 132, as shown in FIG. 10, performs the 4 steps of predicting system anomalies in advance, but provides expert knowledge services such as Jena (ie, each relevant expert in the event of an event). (remote advisory system that establishes expert opinions on the data by inquiring data such as such) and the prediction model of the 4th stage of pre-predicting system abnormalities is used as the initial prediction model of the 1st stage of pre-predicting system abnormalities again. It is possible to have an infinite evolutionary structure that evolves into an initial predictive model in the first stage of system abnormal situation pre-prediction again.

In one embodiment, the artificial intelligence unit 132 repeats steps 1 to 4 of predicting system abnormalities in advance, and performs monitoring to distinguish and determine the operational integrity of the SCADA system and human operation errors with a success rate that gradually increases. Also, by performing monitoring of the facility system in the same way, it is also possible to classify and determine the cause when an abnormality occurs in the facility system by distinguishing between the operational integrity of the facility system and human operation error.

In one embodiment, the artificial intelligence unit 132 collects data of the normal operation and operating state of the monitoring target facility or SCADA system, and unguides (or maps) the normal pattern of the normal operation and operating state. Learning, if an abnormality that deviates from the reference value occurs, it can be predicted as one of the causes of systemic operational integrity and human operation error and displayed to the operator through the HMI 140 In the case of , learning based on the corresponding response can improve the accuracy of the cause analysis and judgment of the abnormality.

In one embodiment, the artificial intelligence unit 132 is input to find the cause of the abnormality by distinguishing and determining operational integrity and human operation error based on artificial intelligence in the monitoring of the monitoring target facility and the SCADA system. The amount of manpower and time required can be greatly reduced, and the accuracy of cause determination and response can be continuously increased close to 100%.

In one embodiment, the artificial intelligence unit 132 may monitor the occurrence of an abnormal pattern before the occurrence of an abnormality or error, predict the possibility of an abnormality in advance, and display it to the operator through the HMI 140, In addition, it is possible to learn in advance the data on the past normal operation and abnormal situation of the SCADA system to increase the accuracy of determining and responding to the cause of the occurrence of abnormalities in the equipment to be monitored and the SCADA system from the initial stage of installation of the system.

Human-machine interface (HMI) 140 provides software capable of human-machine interaction, and displays the processing data transmitted from the central server 130 on the display screen, and uses the operator console 141 to The system control is performed through the central server 130 according to the operation command of the multi-system operator.

In an embodiment, the HMI 140 includes an editing unit 141a, a communication unit 141b, a utility unit 141c, and a monitoring unit 141d, as shown in FIG. 11 , the operation of the SCADA system operator It is possible to provide software that enables human-machine interaction to control the SCADA system on command.

The editing unit 141a edits DI, AI, DO (Digital Output), AO (Analog Output) data, single-line diagrams, and reports.

The communication unit 141b transmits/receives data to and from the central server 130 (ie, the SCADA unit 131 and the artificial intelligence unit 132).

The utility unit 141c is in charge of utility programs such as event search, trend search, TAG support, and report registration.

The monitoring unit 141d displays DI, AI, DO, and AO data as symbols, pictures, characters, and the like, and outputs event data and an alarm sound.

In an embodiment, the monitoring unit 141d is, as shown in FIG. 12, the overall system diagram, (eg, train) control panel screen, PDS screen, communication, device abnormality screen, power quality screen, fault point expression panel screen By displaying the lights on the display, the operator can effectively monitor and manage the equipment through the SCADA system.

The AI-based SCADA system 100 having the configuration as described above is implemented to enable monitoring of system operation integrity and human operation errors, so that the AI-based AI unit 132 of the HMI is the SCADA system Systemic operational integrity and human operation error can be separately monitored, and when a SCADA system error occurs, it can be clearly identified whether the cause is a system error or human operation error, enabling prompt action. .

The artificial intelligence-based SCADA system 100 having the configuration as described above uses the artificial intelligence-based artificial intelligence unit 132 to collect data of the monitoring target facility and the SCADA system in normal operation and operation. Normal operation and operation patterns can be collected and supervised and unsupervised learning, and when an abnormality that deviates from the reference value occurs, an abnormal pattern different from the related normal pattern can be traced and displayed as the cause. When a system is repaired or an operational response action is taken, unsupervised learning or supervised learning is performed based on the corresponding response, thereby improving the accuracy of the cause analysis and judgment of anomalies.

The AI-based SCADA system 100 having the configuration as described above is input to find the cause of the abnormality by utilizing the AI-based AI unit 132 for monitoring facilities and the SCADA system. The amount of manpower and time required can be greatly reduced, and the accuracy of cause determination and response can be continuously increased close to 100%.

The AI-based SCADA system 100 having the configuration as described above, by monitoring the occurrence of an abnormal pattern before the occurrence of the abnormality, predicts the possibility of the occurrence of the abnormality in advance and displays the abnormality to the operator in advance. It is possible to prevent the pattern from occurring, and by the artificial intelligence-based artificial intelligence unit 132 learning in advance data about the past normal operation and abnormal situation of the SCADA system, the system (or the artificial intelligence unit 132) ) from the initial stage of installation, it is possible to increase the accuracy of determining and responding to the cause of abnormalities in facilities and SCADA systems.

13 is a view for explaining an AI-based SCADA system operating method according to an embodiment of the present invention.

Referring to FIG. 13 , the remote terminal device 110 is connected to a monitoring target facility of a facility system at a remote location, collects data of the monitoring target facility, and transmits the collected data to the communication control device 120 side. The data is converted into a possible format and the converted data is transmitted to the communication control device 120 (S301).

In transmitting the converted data in step S301 described above, the RTU 111 provided in the remote terminal device 110 enables mutual control with the equipment to be monitored; The SDP 112 provided in the remote terminal device 110 may apply a plurality of various communication methods and communication protocols to the RTU 111 .

In transmitting the converted data in the above-described step S301, the remote terminal device 110 collects field data from the monitoring target facility (ie, field devices (eg, thermometer or hydraulic meter) to process and monitor the data). and control function, but transmits the collected field data to the upper central server 130 through the communication control device 120 through the communication line.

In transmitting the converted data in step S301 described above, in the RTU 111, a plurality of field devices or equipment (eg, power monitoring target facility, power quality target facility, lighting monitoring target facility, preventive diagnosis facility (eg, , GIS/LA, transformer, etc.), collect field data of field devices or equipment and transmit it to the communication control device 120 .

When data is transmitted in the above-described step S301, the communication control device 120 collects data transmitted from the remote terminal device 110, and, for the collected data, reduces the time required for subsequent data processing. By performing pre-data processing, the pre-processed data is transmitted to the central server 130 (S302).

In transmitting the pre-processing data in the above-described step S302, in the FEP provided in the communication control device 120, for the field data transmitted from the remote terminal device 110) (ie, the RTU 111), pre-processing software By performing pre-processing operations such as integration, refining, reduction, and transformation into data for assembling and disassembling packets, error detection and correction, etc., and transmitting the pre-processed data to the central server 130 through It helps to reduce the time required to process the pre-machining data.

In the transmission of the pre-processing data in the above-described step S302, in the communication control device 120 consisting of a duplication of the active device and the standby device, the duplication unit in the SCADA unit 131 of the central server 130 According to the abnormality check result in (131d), it is possible to provide a continuous and continuous service by performing switching to a standby device.

When the pre-processing data is transmitted in the above-described step S302, the central server 130 collects the pre-processing data transmitted from the communication control device 120, processes the collected pre-processing data, and stores the processing data. It is transmitted to the HMI 140 (S303).

In transmitting the processing data in the above-described step S303, in the central server 130, which is configured as a duplication of an active server and a standby server, an abnormality check result in the duplication unit 131d in the SCADA unit 131 In accordance with this, it is possible to provide continuous and continuous service by switching to the server in the standby state.

In transmitting the processing data in step S303 described above, the central server 130 improves the prediction model in four steps and learns and repeats it to increase the accuracy and efficiency of prediction; In order to detect and respond to the causes of abnormalities occurring during the operation and operation of the SCADA system, which centrally monitors and controls the state data of major infrastructure such as electricity, railways, and water supply, an anomaly prediction model with artificial intelligence learning software is the first step. In step 2, pre-processing and post-processing are performed to form a model that is served in real time, in step 3, the model is improved with re-learning software, and in step 4, expert-level judgment and predicted knowledge through expert knowledge service. Based on artificial intelligence that is learned in such a way that it is applied to a predictive model with Operational integrity related to system elements such as hardware, software, and data, and operational errors caused by human factors such as operator inexperienced operation can be monitored.

In transmitting the processed data in the above-described step S303, the SCADA unit 131 provided in the central server 130 processes the pre-processed data collected from the communication control device 120, and among the processed data, the HMI 140 ) can transmit necessary data to be transmitted, and can store data input from the HMI 140; Based on such data, the artificial intelligence unit 132 provided in the central server 130 may classify and monitor the operational integrity of the system and human operation errors based on artificial intelligence.

In transmitting the processing data in the above-described step S303, the card unit 131 having the communication unit 131a, the determination unit 131b, the storage unit 131c, the redundancy unit 131d, and the module support unit 131e. , the received data may be processed, and the processed data may be transmitted to the artificial intelligence unit 132 and the HMI 140 (ie, the operator console 141).

In transmitting the processing data in step S303 described above, the communication unit 131a transmits and receives data with the communication control device 120 , the HMI 140 (that is, the operator console 141 ), and the artificial intelligence unit 132 , ; The determination unit 131b determines data to determine an on/off event of a digital input (DI) and an event of an upper limit or a lower limit of an analog input (AI); The storage unit 131c stores alarm data, trends, reports (hourly, daily, weekly, monthly, etc.) and input data; The duplication unit 131d checks data synchronization, program, communication error, DBMS error, etc., and performs duplication switching according to the corresponding error check result; The module support unit 131e supports each module management and utility.

In transmitting the processing data in step S303 described above, data collection/loading interworking platform 132a, data processing/model learning interworking platform 132b, server interworking platform 132c, client infrastructure 132d, core 132e) , in the artificial intelligence unit 132 having the domain platform 132f, the core infrastructure 132g, and the security interworking platform 132h, the function of classifying and monitoring the operational integrity of the system and human operation errors based on artificial intelligence. can be done

In transmitting the processing data in the above-described step S303, the data collection/loading interworking platform 132a performs data interworking for data collection and loading with a set of communication protocols and database processes; The data processing/model learning interworking platform 132b is a set of processes for data processing and model learning required for AI learning, and performs data interworking for data processing and model learning; In the server interworking platform 132c, as a set of processes for interworking between platforms in the artificial intelligence unit 132 and executing the service of the artificial intelligence model, server interworking for interworking between internal platforms and service execution of the artificial intelligence model is performed and; In the client infrastructure 132d, a set of processes for artificial intelligence monitoring for allowing the client to use the artificial intelligence unit 132, allowing the client to use the artificial intelligence unit 132; The core 132e is a set of core processes of the artificial intelligence unit 132, and controls the platform inside the artificial intelligence unit 132; The domain platform 132f is a set of processes for managing domains inside the artificial intelligence unit 132 and domains of other servers, and manages internal domains and domains of other servers; The core infrastructure 132g is a set of processes for managing data processed by the data processing/model learning interworking platform 132b, and managing data processed by the data processing/model learning interworking platform 132b; In the security interworking platform 132h, as a set of security-related processes of the artificial intelligence unit 132, the security of the artificial intelligence unit 132 is taken care of.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 requires collection of learning data and pre-processing (data integration, purification, reduction, transformation, etc.) for AI learning, By learning with artificial intelligence learning software, models such as prediction, judgment, and response of abnormal causes can be newly created or modified according to the learning results.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 performs monitoring for facility integrity through the HMI 140; The data integrity can be checked by monitoring the signals and data from the central server 130 and the HMI 140 (that is, the operator console 141), the FEP, and when there is no data reception through the FEP, the SCADA system By monitoring the whole, it is possible to present to the operator through the HMI 140 an abnormality of the module of the FEP or a diagnosis of a failure of the lower terminal device 110 such as the RTU 111, and there is no abnormality in data transmission within the SCADA system. , when the operator does not recognize an error in data processing, it is possible to diagnose the location of data processing and display a notification to the operator through the HMI 140 .

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 performs monitoring for human errors through the HMI 140; If the AI data values associated with a specific facility are learned at the time when a specific facility is turned on, if the AI data is 0 or the data within the normal range is not displayed when the specific facility is put on, a notification is sent to the operator through the HMI 140 In addition, it learns to generate related DI and AI events when a specific facility is turned on or off, and when a preset time elapses when a specific facility is turned on or off, a notification can be displayed to the operator through the HMI 140. Moreover, by learning the actions of the operator after the event occurs, when the same event occurs, when the operator proceeds with actions different from before, a notification may be displayed to the operator through the HMI 140 .

In transmitting the processing data in step S303 described above, the artificial intelligence unit 132 performs, for example, a procedure in which a manager of a specific facility applies power to a specific facility around 9:00 am on weekdays and turns off the power around 6 pm on weekdays. If a repeating normal pattern is learned, and a certain period of time elapses without power to a specific facility being applied around 9:00 am, the possibility of human operation error can be displayed as a notification to the operator of the SCADA system through the HMI 140. have.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 generates a model when starting the system when performing artificial intelligence learning and performs supervised learning in the first step, but the model through the dataset conduct supervised learning; Extracting and serving the supervised model to perform unsupervised learning in two steps, but performing model unsupervised learning through operational data and datasets based on the model extracted and served in the first step; Extract and serve the unsupervised model to perform automatic machine learning in three steps; Performs automatic machine learning through operation data and datasets for each preset cycle based on the two-step extraction and serving model; extracting and serving the machine-learned model to perform expert knowledge learning in four steps, and performing expert knowledge learning by using an expert knowledge service based on the three-step extraction and serving model; After extracting and serving the model learned by the expert knowledge, the process of performing the first-stage supervised learning again based on the fourth-stage extraction and serving model can be infinitely repeated.

In transmitting the processing data in step S303 described above, the artificial intelligence unit 132 performs the first step of predicting system anomalies in advance, but normal and abnormal sensor data or labeled datasets acquired in advance for developing an anomaly prediction model receives big data in a file format that can be recognized in AI learning, such as Excel or image file, and learns an anomaly prediction model with AI learning software. export) by distributing the trained model, serving the distributed model, and sending an abnormal pattern (for example, when the temperature or hydraulic pressure of the facility is different from usual) to the HMI 140, through the HMI 140 A notification of an abnormal pattern can be displayed to the operator.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 performs the 2nd step of predicting the system abnormal situation in advance, but in order to supplement the 1st step of the system abnormal situation prior prediction, it is observed in the SCADA system or facility It collects the data that is used through streaming technology, stores the data extracted through the refining process in the format of a file used by the artificial intelligence learning software, and reads the extracted data saved in the artificial intelligence learning software again to be used in the artificial intelligence model. By preprocessing data and forming a model that is served in real time so that the operator can check the results predicted by the artificial intelligence learning model, it is possible to increase the judgment and predictability of abnormal situations.

In transmitting the processing data in step S303 described above, the artificial intelligence unit 132 performs the 3rd step of predicting the system abnormal situation in advance, but in order to supplement the 2nd step of predicting the system abnormal situation in advance, As a learning software, for the operator's passive model learning and creation process in the second stage of predicting system anomalies in advance, it learns by itself using the extracted data stored at the cycle specified by the operator on a daily, weekly or monthly basis to create a model. , by allowing the generated model to be distributed as a better model through the verification process and comparison with the previous model by itself, it is possible to further increase the judgment and predictability of abnormal situations.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 performs the 4th step of predicting the system abnormal situation in advance, but an expert knowledge service such as Jena, etc. (that is, each relevant expert Through a remote advisory system that inquires the same data and establishes an expert opinion on the data), the judgment and predictive knowledge of the SCADA system expert level is applied to the predictive model, and the post-response of the applied predictive model is transferred to the HMI 140 It transmits and enables the prediction model of the 4th stage of pre-predicting system abnormalities to be used as the initial prediction model in the 1st stage of pre-predicting system abnormalities again, so that it is a predictive model of the 4th stage of pre-predicting system abnormalities through artificial intelligence learning It can evolve and have an infinite evolutionary structure that allows the evolved prediction model to become an initial prediction model in the first stage of system abnormal situation pre-prediction again.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 repeats steps 1 to 4 of predicting the system abnormal situation in advance, monitoring to distinguish and determine the operational integrity of the SCADA system and human operation error. It can be carried out with a progressively higher success rate, and by monitoring the facility system in the same way, it is also possible to classify and determine the cause when an abnormality occurs in the facility system by distinguishing between the operational integrity of the facility system and human operation error.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 collects the data of the normal operation and operating state of the monitoring target facility or SCADA system, and the normal pattern of the normal operation and the operating state In case of unsupervised or supervised learning, when an abnormality that deviates from the reference value occurs, it can be predicted as one of the causes of systemic operational integrity and human operation error and displayed to the operator through the HMI 140, In the case where the action is taken after the fact, learning is made based on the corresponding response, improving the accuracy of the cause analysis and judgment of anomalies.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 distinguishes between operational integrity and human operation error based on artificial intelligence in monitoring of the equipment to be monitored and the SCADA system. It is possible to significantly reduce the amount of manpower and time input to find the cause of the occurrence, and to continuously increase the accuracy of determining the cause and responding to close to 100%.

In transmitting the processing data in the above-described step S303, the artificial intelligence unit 132 monitors the occurrence of an abnormal pattern before the occurrence of an abnormality or error, predicts the possibility of occurrence of the abnormality in advance, and then through the HMI 140 It can be displayed to the operator, and by learning data on the past normal operation and abnormal situations of the SCADA system in advance, the accuracy of the judgment and response to the cause of the abnormal occurrence of the equipment to be monitored and the SCADA system from the beginning of the system installation is improved. may be raised

When the processing data is transmitted in step S303 described above, the HMI 140 provides software capable of human-machine interaction, and displays the processing data transmitted from the central server 130 on the display screen, and the operator console System control is performed through the central server 130 according to the operation command of the SCADA system operator through (141) (S304).

In performing the processing data display and system control in step S304 described above, in the HMI 140 including the editing unit 141a, the communication unit 141b, the utility unit 141c, and the monitoring unit 141d, the SCADA system Software capable of human-machine interaction can be provided to control the SCADA system according to the operator's operating instructions.

In performing the processing data display and system control in the above-described step S304, the editing unit 141a edits DI, AI, DO (Digital Output), AO (Analog Output) data and single-line diagrams and reports; The communication unit 141b transmits and receives data to and from the central server 130 (ie, the SCADA unit 131, the artificial intelligence unit 132); The utility unit 141c is in charge of utility programs such as event search, trend search, TAG support, and report registration; The monitoring unit 141d displays DI, AI, DO, and AO data as symbols, pictures, characters, etc., and outputs event data and an alarm sound.

In performing the processing data display and system control in step S304 described above, in the monitoring unit 141d, the overall system diagram, (eg, train) control panel screen, PDS screen, communication, device abnormality screen, power quality screen, fault point By displaying the expression panel screen as a display, the operator can effectively monitor and manage the facility through the SCADA system.

The AI-based SCADA system operation method having the above configuration enables monitoring of system operation integrity and human operation errors in a SCADA system that centrally monitors and controls the state data of major infrastructure, but a remote terminal The device 110 collects and transmits data of the equipment to be monitored from a remote location, and transmits it to the central server 130 through the communication control device 120 or RTU 111 capable of mutual control with the equipment to be monitored; and an SDP 112 that applies various communication schemes and communication protocols to the RTU; the communication control device 120 includes a front-end processor (FEP) that collects data from the remote terminal device 110, processes the data in advance, and transmits the data to the central server 130; The central server 130 includes the SCADA unit 131 that drives the SCADA system and the artificial intelligence unit 132 that classifies and monitors the operational integrity and human operation errors of the equipment system based on artificial intelligence, from the equipment to be monitored. collecting and storing data received through the remote terminal device 110 and the communication control device 120 , and transmitting abnormality detection and judgment prediction data to the HMI 140 ; The HMI 140 receives data from the central server 130 and displays it on the display screen, and provides software capable of human-machine interaction so that the SCADA system can be controlled according to the operation command of the SCADA system operator. ; Accordingly, the artificial intelligence unit 132 collects data of the state in which the monitoring target facility or SCADA system is operating as a normal operation and a set value by the operator, and teaches and unsupervised normal patterns of normal operation and operation, When an abnormality that deviates from the reference value occurs, the abnormal pattern can be predicted as one of the causes of systemic operational integrity and human operation error and displayed to the operator.

In the AI-based SCADA system operating method having the above-described configuration, the AI unit 132 receives normal and abnormal sensor data or labeled datasets acquired in advance in a file format to develop an anomaly prediction model, and learns artificial intelligence Learning an anomaly prediction model with software, and in the case of a trained model, deploying the trained model after verification through model validation and serving the deployed model to learn a one-step model that displays a notification to an operator when an anomaly pattern occurs; In order to supplement the first-stage model, data observed in the SCADA system or facility is collected through streaming technology, and the data extracted through the purification process is stored in the format of a file used by artificial intelligence learning software, and the stored extracted data Step 2, which reads again and pre-processes it into data usable in the AI learning model, forms a model that is served in real time so that the operator can check the results predicted by the AI learning model, and improves the judgment and predictability of abnormal situations train the model; In order to supplement the two-step model, Auto Machine Learning is added, and in the case of an automatic machine learning prediction model, the operator's instruction learning prediction model and creation process are stored at a cycle specified by the operator on a daily, weekly or monthly basis. Using the extracted data, unsupervised learning is used to generate a predictive model again, and the generated predictive model is self-validated and compared with the previous predictive model to distribute a better learning model to improve judgment and predictability of abnormal situations. trained a three-step model; Through the expert knowledge service, the SCADA system expert level judgment and predictive knowledge are applied to the predictive model, and a four-step model is trained to transmit the post-response of the applied predictive model to the operator; The prediction model of the 4th stage is used again as the initial prediction model of the 1st stage, and it evolves into the other 4th stage predictive model through artificial intelligence learning. It can be made to have a learning structure of infinite evolution.

Above, the embodiment of the present invention is not implemented only through the above-described apparatus and/or operation method, but through a program for realizing a function corresponding to the configuration of the embodiment of the present invention, a recording medium in which the program is recorded, etc. It may be implemented, and such an implementation can be easily implemented by an expert in the technical field to which the present invention pertains from the description of the above-described embodiments. Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improved forms of the present invention are also provided by those skilled in the art using the basic concept of the present invention as defined in the following claims. is within the scope of the right.

100: AI-based SCADA system
110: remote terminal device
111: RTU
112: SDP
120: communication control device
130: central server
131: Skadabu
131a: communication department
131b: judgment unit
131c: storage
131d: Redundancy
131e: Module Support
132: artificial intelligence department
132a: data collection / loading interlocking platform
132b: data processing / model learning interworking platform
132c: server interworking platform
132d: Client Infrastructure
132e: core
132f: domain platform
132g: Core Infrastructure
132h: security interworking platform
140: HMI
141: operator console
141a: Editorial Department
141b: communication department
141c: Utility Department
141d: monitoring unit

Claims (5)

In the AI-based SCADA system,
The AI-based SCADA system,
a plurality of remote terminal devices connected to each monitoring target facility, collecting data of each monitoring target facility, and transmitting data;
a communication control device that collects data from the remote terminal device and transmits the pre-processed data by performing pre-data processing on the collected data to reduce a time required for a subsequent data processing process;
By collecting pre-processing data from the communication control device, data processing, abnormal data monitoring, data storage, and operator control data received from the HMI are transmitted to the communication control device, and using the entire data stored and the SCADA unit to store the history Execute the learning of artificial intelligence, and through the learned artificial intelligence, operational integrity related to systemic factors such as data loss due to device failure and errors of the system, and incorrect operation commands caused by human factors due to inexperienced operation of the operator a central server composed of an artificial intelligence unit that monitors human operation errors; and
and an HMI for displaying processing data from the central server on a display screen and transmitting a system control command to the central server according to an operator's operating command,
The remote terminal device 110, RTU (Remote Terminal Unit) 111 for enabling mutual control with the equipment to be monitored; A smart digital processor (SDP) 112 for applying a plurality of communication methods and communication protocols to the RTU 111 is provided to collect field data of field devices or equipment and transmitted to the communication control device 120,
The communication control device,
With respect to the converted data collected from the remote terminal device, pre-processing of data for assembling and dismantling, error detection and correction of packets through pre-processing software, refining, reduction, and transformation is performed to convert the data into pre-processed data to the central server Equipped with an FEP that transmits to
The communication control device,
It consists of duplication of active state and standby state, and provides continuous and continuous service by switching to standby state according to the abnormality check result,
The central server,
It consists of a duplication of the active server and the standby server, and provides continuous and continuous service by switching to the standby server according to the abnormality check result.
The central server,
a scad unit for processing and storing the pre-processing data collected from the communication control device, and for transmitting the data of the current state so that the operator can monitor it together with abnormal data monitoring and system control; and an artificial intelligence unit for classifying, inferring, and monitoring the operational integrity of the system and the operator's operational errors based on artificial intelligence by utilizing all the stored data; and
The SCADA unit may include: a communication unit for transmitting and receiving data with the communication control device, the HMI, and the artificial intelligence unit; a determination unit for determining data to determine an on/off event of DI and an event of upper or lower limit of AI; a storage unit for storing alarm data and trends, reports and input data; a redundancy unit for checking data synchronization and program, communication abnormality, and DBMS abnormality, and performing redundancy switching according to the abnormality check result; Equipped with a module support unit to support each module management and utility,
The artificial intelligence unit, a data collection / loading interworking platform for data collection and loading with a set of communication protocols and database processes; A data processing/model learning interworking platform for data processing and model learning as a set of processes for data processing and model learning required for artificial intelligence learning; A server interworking platform for interplatform interworking and artificial intelligence model service execution as a set of processes for interplatform interworking and artificial intelligence model service execution; a client infrastructure for allowing a client to use the AI unit as a set of processes for AI monitoring; a core for controlling the platform with a set of core processes; a domain platform for managing internal domains and domains of other servers as a set of processes for managing internal domains and domains of other servers; a core infrastructure for managing data processed by the data processing/model learning interworking platform as a set of processes for managing data; and a security interworking platform for taking care of security with a set of security-related processes,
The artificial intelligence unit of the central server learns an anomaly prediction model with learning software, performs pre-processing and post-processing to form a model that is served in real time, improves the model with re-learning software, and provides expert knowledge service through expert knowledge service Based on the artificial intelligence that is learned in a way that has an infinite evolutionary structure that repeatedly performs the transmission of the based post-response to the operator by applying it to the prediction model with the judgment and predicted knowledge of the SCADA system, hardware or software, To monitor the operational integrity related to the systemic elements of the data, and human operational errors caused by the inexperienced human elements of the operator's operation;
The artificial intelligence unit, but performs monitoring of equipment integrity through the HMI; The data integrity is checked by monitoring the signals and data from the central server, the HMI, and FEP, and when there is no data reception through the FEP, the SCADA system is monitored as a whole to diagnose FEP module abnormalities or RTU failures. presented to the operator;
The artificial intelligence unit, but performs monitoring of equipment integrity through the HMI; There is no problem in data transmission within the SCADA system, but when an error in data processing occurs, the operator is unaware of the location of data processing and displays a notification to the operator,
The artificial intelligence unit, but performs monitoring for human error through the HMI; If the AI data values associated with the specific equipment are learned at the time the specific equipment is put in, the operator is notified if the AI data is 0 or the data within the normal range is not displayed at the time the specific equipment is put in.
The artificial intelligence unit, but performs monitoring for human error through the HMI; It learns to generate related DI and AI events when a specific facility is turned on or off, and displays a notification to the operator when a preset time elapses when the specific facility is turned on or off.
The artificial intelligence unit, but performs monitoring for human error through the HMI; By learning the actions of the operator after the event occurs, when the same event occurs, when the operator proceeds with a different action than before, a notification is displayed to the operator,
The artificial intelligence unit generates a model when starting the system and performs supervised learning, but performs model supervised learning through a dataset; performing unsupervised learning by extracting and serving the supervised model in the first step, and performing model unsupervised learning through operational data and datasets based on the model extracted and served in the first step; Perform automatic machine learning by extracting and serving the unsupervised model in two steps; Performs automatic machine learning through operation data and datasets for each preset cycle based on the two-step extraction and serving model; performing expert knowledge learning by extracting and serving the machine-learned model in three steps, and performing expert knowledge learning by using an expert knowledge service based on the three-step extraction and serving model; After extracting and serving the expert knowledge-learned model in 4 steps, repeat the process of performing supervised learning again based on the 4 step extraction and serving model,
The artificial intelligence unit, but performs the first step of predicting system anomalies in advance; In order to develop an anomaly prediction model, it receives big data in a file format that can be recognized in artificial intelligence learning, such as Excel or image file, from normal and abnormal sensor data or labeled dataset acquired in advance, and uses artificial intelligence learning software to predict anomalies. In the case of the learned model, after verification through model verification, the trained model is distributed, and the distributed model is served to transmit an abnormal pattern notification to the HMI, and the artificial intelligence unit, the system abnormal situation prediction 2 perform the steps; The data observed in the SCADA system or facility is collected through streaming technology, the data extracted through the refining process is stored in the file format used by the AI learning software, and the extracted data stored in the AI learning software is read again Pre-processing with data usable in the artificial intelligence learning model, forming a model that is served in real time so that an operator can check the results predicted by the artificial intelligence learning model, and the artificial intelligence unit, re-learning software such as automatic machine learning software Performs the 3rd step of predicting system anomalies in advance of adding , distributes the generated model through self-verification and comparison with previous models, and the artificial intelligence unit performs the 4 steps of predicting system abnormalities in advance, but each relevant expert inquires the event data to obtain expert opinion on the data Through the remote advisory system that establishes opinions, SCDA system expert level judgment and predictive knowledge are applied to the predictive model, and the post-response of the applied predictive model is transmitted to the HMI. to be used again as an initial prediction model in the first stage of system abnormality pre-prediction, and the artificial intelligence unit collects the data of the normal operation and operation state of the monitoring target facility or SCADA system, and the normal operation and operation state unsupervised or supervised learning of the normal pattern, and when an abnormality that deviates from the reference value occurs, it is predicted as one of the causes of systemic operational integrity and human operation error and displayed to the operator through the HMI, In case the action of
The artificial intelligence unit monitors the occurrence of an abnormal pattern before the occurrence of an abnormality or error, predicts the possibility of an abnormality in advance, and displays it to the operator through the HMI, By learning the data in advance, it is possible to determine and respond to the cause of the occurrence of abnormalities in the equipment to be monitored and the SCADA system from the initial stage of system installation,
The HMI may include: an editing unit for editing DI, AI, DO, AO data and single-line diagrams and reports; a communication unit for transmitting and receiving data to and from the central server; Utility department in charge of utility programs of event search, trend search, TAG support, and report registration; and a monitoring unit for displaying DI, AI, DO, and AO data as symbols, pictures, and characters, and outputting event data and an alarm sound, wherein the editing unit includes DI, AI, Digital Output (DO), and AO (Analog). Output) data, single-line diagram, and report editing are performed, and the communication unit transmits and receives data to and from the central server 130 (ie, SCADA unit 131, artificial intelligence unit 132), and the utility unit, event search , trend search, TAG support, and report registration utility program AI-based SCADA system, characterized in that it displays the overall schematic, control panel screen, PDS screen, communication, device abnormality screen, power quality screen, and fault point expression panel screen.

delete delete delete delete

Images