An automatic control system and method based on big data analysis

Abstract

The present invention discloses an automatic control system and method based on big data analysis, and relates to the field of automatic control technology. The system of the present invention includes: a data acquisition and data processing module, a data analysis and feature extraction module, a fault analysis and isolation area identification module, a real-time data analysis and event identification module, and an automatic control strategy execution and fault response module; the data acquisition and historical data processing module collects historical data and log records, screens and constructs historical data sets; the historical data analysis and feature extraction module analyzes historical data trends and establishes feature-event mapping relationships; the fault diagnosis and isolation area identification module identifies fault isolation areas and key points through correlation analysis, and extracts corresponding automatic control strategies; the real-time data analysis and event identification module acquires data in real time and extracts features to determine event types; the automatic control strategy execution and fault response module generates notifications based on event identification results.

Description

Technical Field

[0001] The present invention relates to the field of automation control technology, and in particular to an automation control system and method based on big data analysis. Background Art

[0002] With the continuous improvement of industrial automation and intelligence, traditional equipment monitoring and fault diagnosis methods have gradually failed to meet the needs of modern production environments. Especially in complex production lines, the correlation between equipment is strong, and the coordination between various links is crucial. The occurrence of failures is often a gradual process and may cause a chain reaction to the entire production system. Traditional equipment monitoring systems mostly rely on preset thresholds to issue alarms or take repair measures, but this method has many limitations and cannot identify potential faults in a timely manner. In the context of diversified equipment and complex failure modes, it is impossible to fully and effectively guarantee the safe operation of the production line.

[0003] Current equipment monitoring and fault diagnosis systems mostly rely on the historical fault data and basic sensor information of a single device. Although there are predictive maintenance systems based on data analysis, the existing technologies still have obvious deficiencies in the following aspects: Most of the existing technologies monitor and predict faults for a single device, lack overall consideration of the interaction between devices, and are difficult to conduct comprehensive fault prediction and risk assessment; some existing systems fail to effectively isolate faults and identify key points during fault diagnosis, and are thus unable to accurately locate the specific area of ​​the fault, affecting the efficiency and accuracy of fault handling; many traditional monitoring systems do not have the ability to learn and self-optimize, making it difficult to adapt to changes in equipment operating modes in industrial environments and easily miss the detection of new fault modes; this results in poor prediction accuracy in the system when responding to new situations, reducing the reliability and timeliness of fault diagnosis. Summary of the invention

[0004] The purpose of the present invention is to provide an automated control system and method based on big data analysis to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] An automatic control method based on big data analysis comprises the following steps:

[0007] Step S100. Collect historical data and log records of all industrial equipment in the monitoring area, filter out historical data of industrial equipment under different events through log records, and construct a historical data set; the historical data refers to the production-related operating parameters collected by the corresponding sensor equipment when the industrial equipment is started; the log record refers to the event record corresponding to the industrial equipment;

[0008] Step S200. Analyze the historical data trends of each industrial device under different events according to the historical data set, and extract corresponding historical data features based on the historical data trends; for each historical data set, analyze the historical data features corresponding to the corresponding industrial device, and establish a mapping relationship between the historical data features and the corresponding events;

[0009] Step S300. Based on the mapping relationship between historical data features and corresponding events, combined with log records, the events corresponding to different industrial equipment are subjected to correlation analysis, and the fault isolation area and the key points of the fault isolation area are identified according to the correlation analysis results; the automation control strategy of the corresponding event is extracted from the log records, and the fault isolation area and the key points of the fault isolation area are matched with the corresponding automation control strategy;

[0010] Step S400. Acquire real-time data of all industrial equipment in the monitoring area, extract real-time data features from the real-time data, and analyze the correlation between the real-time data features and the historical data features; judge the current event according to the correlation analysis results, and perform corresponding processing based on the current event.

[0011] Furthermore, step S100 includes:

[0012] S101. Collect historical data and log records of all industrial equipment in the monitoring area according to the set time window T, and T=[t0,tn], where t0 represents the start time of data collection and tn represents the end time of data collection; the events recorded in the log include normal events and fault events, the normal event refers to the industrial equipment operating according to the predetermined operating parameters and specifications, and no abnormality or fault occurs; the fault event refers to the industrial equipment deviating from the normal working state;

[0013] S102. For the historical data and log records corresponding to each industrial equipment in the monitoring area, according to the event occurrence timestamp t_event in the log record, filter out the historical data occurring in the time period T1 before and after the timestamp t_event, and T1=[t_event-ΔT, t_event+ΔT], where ΔT represents the time span; for the historical data of each event in the time period T1, perform data preprocessing, and convert the historical data into a unified format to form a historical data set Li, and Li={li1,li2,...,lim}, where Li represents the historical data set of the i-th industrial equipment in the monitoring area, i represents the number of the industrial equipment in the monitoring area, li1 represents the historical data corresponding to the first event of the i-th industrial equipment in the monitoring area, li2 represents the historical data corresponding to the second event of the i-th industrial equipment in the monitoring area, and so on, lim represents the historical data corresponding to the m-th event of the i-th industrial equipment in the monitoring area, and m represents the total number of events in the log records of the i-th industrial equipment in the monitoring area; wherein the number of historical data sets is equal to the number of industrial equipment in the monitoring area.

[0014] Further, step S200 includes:

[0015] S201. For each element in the historical data set Li, the elements in the historical data set Li are divided into several categories according to the event code recorded in the log, and the event code corresponding to each category is the same, wherein there is only one normal event code in the event recorded in the log, and there is more than one code for the fault event in the event recorded in the log; trend analysis is performed on the historical data corresponding to the element of each category, and the specific analysis process is as follows:

[0016] Use polynomial fitting or curve fitting to capture the changing trend of historical data over time. Set the fitting function to f(t), which represents the changing trend of historical data over time: f(t)=a 0 +a 1 t+a 2 t 2 +...+a k t k , where a 0 , a 1 , a 2 , ..., a k represents the fitting parameter, t represents the time; the parameter a is solved by the least squares method 0 , a 1 , a 2 , ..., a k , so that the sum of squared errors is minimized: ∑t k ∈T1,[f(t k )-y k ]2 , where y k represents the actual observed data;

[0017] Calculate the derivative f'(t) of the fitting function to evaluate the rate of change of historical data over time, thereby obtaining the trend change speed of historical data:

[0018] ;

[0019] S202. Based on the trend analysis results of historical data, the historical data features of each element are extracted and normalized to establish a historical data feature vector V, and V=[v1,v2,...,vs], where v1 represents the first historical data feature of the element in the historical data set, v2 represents the second historical data feature of the element in the historical data set, and so on, vs represents the sth historical data feature of the element in the historical data set; according to the category to which each element belongs, the corresponding historical data feature vectors are summarized, and the historical data feature vectors of each category are analyzed to obtain the historical data feature vector Vcj of each category, where Vcj represents the historical data feature vector of the jth category, and the maximum value of j is equal to the number of categories of elements in the historical data set; according to the event code corresponding to each category, a mapping relationship between the historical data feature vector Vcj and the corresponding event is established, which is expressed as: M={(Vcj,Ej)|j=1,2,...,N}, where Ej represents the event code corresponding to the jth category, and N represents the number of categories.

[0020] Furthermore, the specific analysis process of the historical data feature vector Vcj of each category is as follows:

[0021] According to the historical data feature vector of each category cj, the average value μx of each eigenvalue of the historical data feature vector is calculated, where x represents the number of the average value of the eigenvalue, ranging from 1 to s; according to the average value μx of the eigenvalue of the historical data feature vector, a feature center vector Ucj is formed, and Ucj=[μ1,μ2,...,μs]j, where Ucj represents the feature center vector of the j-th category, and [μ1,μ2,...,μs]j represents the eigenvalue of the feature center vector of the j-th category;

[0022] Calculate the deviation value P between the eigenvalue vx of each historical data feature vector and the eigenvalue μx of the corresponding feature center vector Ucj, and P=|vx-μx|. Compare the deviation value P of the eigenvalue of each historical data feature vector with the preset deviation threshold P0, and filter the eigenvalues ​​that satisfy the deviation value P greater than the preset deviation threshold P0, so as to obtain the historical data feature vector Vcj of the corresponding category.

[0023] Furthermore, step S300 includes:

[0024] S301. Based on the mapping relationship between the historical data features and the corresponding events, combined with the log records, the time period T1 of the fault event of each industrial device is obtained, the industrial device numbers with overlapping parts in the time period T1 are extracted, and arranged in the order of the fault event occurrence timestamp t_event of the corresponding industrial device, with the fault event number of the first industrial device as the index, so as to construct the associated event sequence of the fault events of different industrial devices, and each component of the associated event sequence is represented as: (i, E, t_event), where i represents the industrial device number, E represents the fault event code, and t_event represents the fault event occurrence timestamp;

[0025] S302. For each category corresponding to the fault event of each industrial equipment, summarize the associated event sequences indexed by the event code of the corresponding category, and count the number N of the same associated event sequences and the total number N1 of fault events of the corresponding category, calculate the proportion Z of the associated event sequences, and Z=N / N1; filter the associated event sequences whose proportion Z is greater than or equal to the threshold Z0, find the corresponding industrial equipment number according to the filtered associated event sequence, and find the corresponding position coordinates of the industrial equipment in the monitoring area within the monitoring area based on the industrial equipment number, so as to identify the fault isolation area, and set the industrial equipment corresponding to the index of the filtered associated event sequence as the key point of the fault isolation area; according to the filtered associated event sequence, obtain the automation control strategy of the corresponding fault event in the log record, and correspond the fault isolation area and the key point of the fault isolation area to the corresponding automation control strategy.

[0026] Furthermore, step S400 includes:

[0027] S401. Obtain the real-time data of all industrial equipment in the monitoring area, analyze the real-time data according to the analysis method of historical data, so as to extract the real-time data features and form a real-time data feature vector; perform similarity calculation on the real-time data feature vector and the historical data feature vector of the corresponding industrial equipment in turn, obtain the historical data feature vector corresponding to the maximum value of the similarity calculation as the matching target, and compare the similarity calculation result with the matching target with the preset threshold; if the similarity calculation result is less than the preset threshold, obtain the event number corresponding to the matching target; if the event number corresponding to the matching target belongs to a normal event, output notification information to relevant personnel, and the relevant personnel determine whether there is a fault event. If the judgment result is a normal event, the real-time data features corresponding to the current event are stored in the log record and marked as a normal event; if the judgment result is a fault event, record the control strategy of the relevant personnel, and store the real-time data features corresponding to the current event in the log record, and mark it as a fault event;

[0028] S402. If the similarity calculation result is greater than the preset threshold, the event number corresponding to the matching target is obtained; if the event number corresponding to the matching target belongs to a normal event, no processing is performed; if the event number corresponding to the matching target belongs to a fault event, the fault warning information is output to relevant personnel, and the fault warning information includes the fault isolation area and the key points of the fault isolation area, and the corresponding automation control strategy is output according to the fault isolation area and the key points of the fault isolation area.

[0029] An automatic control system based on big data analysis, comprising: a data acquisition and data processing module, a data analysis and feature extraction module, a fault analysis and isolation area identification module, a real-time data analysis and event identification module, and an automatic control strategy execution and fault response module;

[0030] The data acquisition and data processing module collects the historical data and log records of all industrial equipment in the monitoring area, filters out the historical data of industrial equipment under different events through log records, and constructs a historical data set;

[0031] The data analysis and feature extraction module analyzes the historical data trends of each industrial device under different events based on the historical data set, and extracts the corresponding historical data features based on the historical data trends; for each historical data set, analyzes the historical data features corresponding to the corresponding industrial device, and establishes a mapping relationship between the historical data features and the corresponding events;

[0032] The fault analysis and isolation area identification module performs correlation analysis on the events corresponding to different industrial equipment based on the mapping relationship between historical data features and corresponding events, combined with log records, and identifies the fault isolation area and the key points of the fault isolation area according to the correlation analysis results; and extracts the automation control strategy of the corresponding event from the log records, and corresponds the fault isolation area and the key points of the fault isolation area to the corresponding automation control strategy;

[0033] The real-time data analysis and event recognition module obtains the real-time data of industrial equipment in the monitoring area in real time, and extracts the real-time data features according to the historical data analysis method to form a real-time data feature vector; the real-time data feature vector and the historical data feature vector are similarly calculated, and the current event is judged as a normal event or a fault event based on the similarity calculation result;

[0034] The automatic control strategy execution and fault response module generates corresponding notification information based on the recognition results of the real-time data analysis and event recognition module. If there is fault warning information, the system will execute the corresponding automatic control strategy.

[0035] Further, the data acquisition and data processing module includes a data collection unit and a data processing unit;

[0036] The data collection unit collects historical data and log records of all industrial equipment in the monitoring area; the data processing unit preprocesses and converts the format of the collected historical data, and constructs a historical data set based on the log records;

[0037] The data analysis and feature extraction module includes a trend analysis unit, a feature extraction unit, and a category division unit;

[0038] The trend analysis unit performs trend analysis on the historical data of each industrial equipment; the feature extraction unit extracts the corresponding historical data features based on the trend analysis results of the historical data, and normalizes the historical data features to form a historical data feature vector; the category classification unit classifies the historical data feature vector into different categories based on the event codes in the log records; and calculates the feature center vector of each category, extracts the historical data features of each category and forms the corresponding historical data feature vector.

[0039] Further, the fault analysis and isolation area identification module includes an event correlation analysis unit, a fault isolation area identification unit, and a control strategy extraction and mapping unit;

[0040] The event correlation analysis unit performs correlation analysis on fault events based on the mapping relationship between historical data features and events, combined with log records; the fault isolation area identification unit identifies the fault isolation areas through the correlation analysis results, and marks the key points of the fault isolation areas in these areas; the control strategy extraction and mapping unit extracts the automation control strategies related to the fault events from the log records, and maps the automation control strategies with the fault isolation areas and the key points of the fault isolation areas.

[0041] Furthermore, the real-time data analysis and event recognition module includes a real-time data acquisition unit, a real-time data feature extraction unit, and an event recognition unit;

[0042] The real-time data acquisition unit collects the real-time data of all industrial equipment; the real-time data feature extraction unit analyzes the real-time data of all industrial equipment according to the analysis process of historical data, thereby extracting the real-time data features; the event recognition unit determines the type of event currently occurring based on the similarity calculation between the real-time data features and the historical data features;

[0043] The automated control strategy execution and fault response module includes a control strategy execution unit and a fault warning and notification unit;

[0044] After identifying a fault event, the control strategy execution unit performs automated control operations according to the corresponding control strategy; the fault warning and notification unit sends fault warning information to relevant personnel based on the identification results of the event identification unit when a fault event is identified, and provides location information of the fault isolation area and key points.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] By analyzing the historical data and log records of industrial equipment, the interaction relationship between devices is identified and correlation analysis is performed; this can not only predict the failure of a single device, but also consider the interaction between devices, thereby improving the overall prediction ability of the occurrence of failures; compared with the traditional technology that mainly focuses on a single device and lacks the analysis of the linkage between devices, the present invention provides a more comprehensive and integrated fault diagnosis method. In the existing fault diagnosis system, there are certain limitations in fault isolation and key point identification, and the source of the fault is often not accurately located, resulting in low fault handling efficiency; the present invention effectively identifies the fault isolation area and key points through the analysis based on historical data and the correlation analysis of fault events, and effectively responds through the automatic control strategy, which greatly improves the accuracy of fault location and the timeliness of processing. Traditional monitoring systems usually lack the ability to self-optimize and are difficult to adapt to changes in equipment operation modes; the present invention can adapt to changes in the operating status of industrial equipment and discover new fault modes in a timely manner by continuously collecting and analyzing historical data and real-time data; this self-optimization ability enables the system to continuously adjust and optimize with changes in equipment and environment, ensuring the efficiency and accuracy of fault prediction and diagnosis. The present invention combines the characteristics of real-time data and historical data, and through similarity calculation matching analysis, it can quickly identify whether it is a fault in real-time events, and issue early warnings based on the matching and similarity of historical data; therefore, the system can not only make predictions based on historical experience, but also dynamically adjust fault warnings based on real-time conditions, significantly improving the timeliness and reliability of fault diagnosis. The present invention combines fault isolation areas, key points, and automated control strategies to form a complete set of fault warning and processing solutions. Through these strategies, relevant personnel can quickly obtain fault information and take corresponding measures, greatly improving the efficiency and accuracy of fault handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1 It is a schematic diagram of an automatic control system module based on big data analysis of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1 , the present invention provides a technical solution:

[0051] An automatic control system based on big data analysis, comprising: a data acquisition and data processing module, a data analysis and feature extraction module, a fault analysis and isolation area identification module, a real-time data analysis and event identification module, and an automatic control strategy execution and fault response module;

[0052] The data acquisition and data processing module collects the historical data and log records of all industrial equipment in the monitoring area, filters out the historical data of industrial equipment under different events through log records, and constructs a historical data set;

[0053] The data analysis and feature extraction module analyzes the historical data trends of each industrial device under different events based on the historical data set, and extracts the corresponding historical data features based on the historical data trends; for each historical data set, analyzes the historical data features corresponding to the corresponding industrial device, and establishes a mapping relationship between the historical data features and the corresponding events;

[0054] The fault analysis and isolation area identification module performs correlation analysis on the events corresponding to different industrial equipment based on the mapping relationship between historical data features and corresponding events, combined with log records, and identifies the fault isolation area and the key points of the fault isolation area according to the correlation analysis results; and extracts the automation control strategy of the corresponding event from the log records, and corresponds the fault isolation area and the key points of the fault isolation area to the corresponding automation control strategy;

[0055] The real-time data analysis and event recognition module obtains the real-time data of industrial equipment in the monitoring area in real time, and extracts the real-time data features according to the historical data analysis method to form a real-time data feature vector; the real-time data feature vector and the historical data feature vector are similarly calculated, and the current event is judged as a normal event or a fault event based on the similarity calculation result;

[0056] The automatic control strategy execution and fault response module generates corresponding notification information based on the recognition results of the real-time data analysis and event recognition module. If there is fault warning information, the system will execute the corresponding automatic control strategy.

[0057] The data acquisition and data processing module includes a data collection unit and a data processing unit;

[0058] The data collection unit collects historical data and log records of all industrial equipment in the monitoring area; the data processing unit preprocesses and converts the format of the collected historical data, and constructs a historical data set based on the log records;

[0059] The data analysis and feature extraction module includes a trend analysis unit, a feature extraction unit, and a category division unit;

[0060] The trend analysis unit performs trend analysis on the historical data of each industrial equipment; the feature extraction unit extracts the corresponding historical data features based on the trend analysis results of the historical data, and normalizes the historical data features to form a historical data feature vector; the category classification unit classifies the historical data feature vector into different categories based on the event codes in the log records; and calculates the feature center vector of each category, extracts the historical data features of each category and forms the corresponding historical data feature vector.

[0061] The fault analysis and isolation area identification module includes an event correlation analysis unit, a fault isolation area identification unit, and a control strategy extraction and mapping unit;

[0062] The event correlation analysis unit performs correlation analysis on fault events based on the mapping relationship between historical data features and events, combined with log records; the fault isolation area identification unit identifies the fault isolation areas through the correlation analysis results, and marks the key points of the fault isolation areas in these areas; the control strategy extraction and mapping unit extracts the automation control strategies related to the fault events from the log records, and maps the automation control strategies with the fault isolation areas and the key points of the fault isolation areas.

[0063] The real-time data analysis and event recognition module includes a real-time data acquisition unit, a real-time data feature extraction unit and an event recognition unit;

[0064] The real-time data acquisition unit collects the real-time data of all industrial equipment; the real-time data feature extraction unit analyzes the real-time data of all industrial equipment according to the analysis process of historical data, thereby extracting the real-time data features; the event recognition unit determines the type of event currently occurring based on the similarity calculation between the real-time data features and the historical data features;

[0065] The automated control strategy execution and fault response module includes a control strategy execution unit and a fault warning and notification unit;

[0066] After identifying a fault event, the control strategy execution unit performs automated control operations according to the corresponding control strategy; the fault warning and notification unit sends fault warning information to relevant personnel based on the identification results of the event identification unit when a fault event is identified, and provides location information of the fault isolation area and key points.

[0067] An automatic control method based on big data analysis comprises the following steps:

[0068] Step S100. Collect historical data and log records of all industrial equipment in the monitoring area, filter out historical data of industrial equipment under different events through log records, and construct a historical data set; the historical data refers to the production-related operating parameters collected by the corresponding sensor equipment when the industrial equipment is started; the log record refers to the event record corresponding to the industrial equipment;

[0069] Step S200. Analyze the historical data trends of each industrial device under different events according to the historical data set, and extract corresponding historical data features based on the historical data trends; for each historical data set, analyze the historical data features corresponding to the corresponding industrial device, and establish a mapping relationship between the historical data features and the corresponding events;

[0070] Step S300. Based on the mapping relationship between historical data features and corresponding events, combined with log records, the events corresponding to different industrial equipment are subjected to correlation analysis, and the fault isolation area and the key points of the fault isolation area are identified according to the correlation analysis results; the automation control strategy of the corresponding event is extracted from the log records, and the fault isolation area and the key points of the fault isolation area are matched with the corresponding automation control strategy;

[0071] Step S400. Acquire real-time data of all industrial equipment in the monitoring area, extract real-time data features from the real-time data, and analyze the correlation between the real-time data features and the historical data features; judge the current event according to the correlation analysis results, and perform corresponding processing based on the current event.

[0072] Step S100 includes:

[0073] S101. Collect historical data and log records of all industrial equipment in the monitoring area according to the set time window T, and T=[t0,tn], where t0 represents the start time of data collection and tn represents the end time of data collection; the events recorded in the log include normal events and fault events, the normal event refers to the industrial equipment operating according to the predetermined operating parameters and specifications, and no abnormality or fault occurs; the fault event refers to the industrial equipment deviating from the normal working state;

[0074] S102. For the historical data and log records corresponding to each industrial equipment in the monitoring area, according to the event occurrence timestamp t_event in the log record, filter out the historical data occurring in the time period T1 before and after the timestamp t_event, and T1=[t_event-ΔT, t_event+ΔT], where ΔT represents the time span; for the historical data of each event in the time period T1, perform data preprocessing, and convert the historical data into a unified format to form a historical data set Li, and Li={li1,li2,...,lim}, where Li represents the historical data set of the i-th industrial equipment in the monitoring area, i represents the number of the industrial equipment in the monitoring area, li1 represents the historical data corresponding to the first event of the i-th industrial equipment in the monitoring area, li2 represents the historical data corresponding to the second event of the i-th industrial equipment in the monitoring area, and so on, lim represents the historical data corresponding to the m-th event of the i-th industrial equipment in the monitoring area, and m represents the total number of events in the log records of the i-th industrial equipment in the monitoring area; wherein the number of historical data sets is equal to the number of industrial equipment in the monitoring area.

[0075] Step S200 includes:

[0076] S201. For each element in the historical data set Li, the elements in the historical data set Li are divided into several categories according to the event code recorded in the log, and the event code corresponding to each category is the same, wherein there is only one normal event code in the event recorded in the log, and there is more than one code for the fault event in the event recorded in the log; trend analysis is performed on the historical data corresponding to the element of each category, and the specific analysis process is as follows:

[0077] Use polynomial fitting or curve fitting to capture the changing trend of historical data over time. Set the fitting function to f(t), which represents the changing trend of historical data over time: f(t)=a 0 +a 1 t+a 2 t 2 +...+a k t k , where a 0 , a 1 , a 2 , ..., a k represents the fitting parameter, t represents the time; the parameter a is solved by the least squares method 0 , a 1 , a 2 , ..., a k , so that the sum of squared errors is minimized: ∑t k ∈T1,[f(t k )-y k ] 2, where y k represents the actual observed data;

[0078] Calculate the derivative f'(t) of the fitting function to evaluate the rate of change of historical data over time, thereby obtaining the trend change speed of historical data: ;

[0079] In this embodiment, based on the trend analysis results, some important historical data features are extracted to characterize the operating status of industrial equipment under different events. The feature extraction process usually includes the following aspects:

[0080] Mean feature: The mean of the historical data of each industrial equipment in a period of time before and after the event represents the average operating status of the equipment during the event;

[0081] Variance feature: By calculating the variance of historical data, the magnitude of data fluctuation is measured, thereby capturing the stability or abnormal fluctuation of the device during the event;

[0082] Maximum and minimum value features: By calculating the maximum and minimum values ​​of historical data, you can identify the extreme value conditions of the equipment during a specific event. Especially for fault events, sometimes the changes in the maximum and minimum values ​​can reflect abnormalities.

[0083] Slope feature: The slope reflects the rate of change of equipment data. The trend change of industrial equipment during the event is characterized by calculating the average slope of the trend fitting function within the time window;

[0084] Periodic characteristics: If the operating status of the equipment shows periodic changes, the periodic components can be extracted through methods such as Fourier transform or wavelet transform.

[0085] S202. Based on the trend analysis results of historical data, the historical data features of each element are extracted and normalized to establish a historical data feature vector V, and V=[v1,v2,...,vs], where v1 represents the first historical data feature of the element in the historical data set, v2 represents the second historical data feature of the element in the historical data set, and so on, vs represents the sth historical data feature of the element in the historical data set; according to the category to which each element belongs, the corresponding historical data feature vectors are summarized, and the historical data feature vectors of each category are analyzed to obtain the historical data feature vector Vcj of each category, where Vcj represents the historical data feature vector of the jth category, and the maximum value of j is equal to the number of categories of elements in the historical data set; according to the event code corresponding to each category, a mapping relationship between the historical data feature vector Vcj and the corresponding event is established, which is expressed as: M={(Vcj,Ej)|j=1,2,...,N}, where Ej represents the event code corresponding to the jth category, and N represents the number of categories.

[0086] The specific analysis process of the historical data feature vector Vcj of each category is as follows:

[0087] According to the historical data feature vector of each category cj, the average value μx of each eigenvalue of the historical data feature vector is calculated, where x represents the number of the average value of the eigenvalue, ranging from 1 to s; according to the average value μx of the eigenvalue of the historical data feature vector, a feature center vector Ucj is formed, and Ucj=[μ1,μ2,...,μs]j, where Ucj represents the feature center vector of the j-th category, and [μ1,μ2,...,μs]j represents the eigenvalue of the feature center vector of the j-th category;

[0088] Calculate the deviation value P between the eigenvalue vx of each historical data feature vector and the eigenvalue μx of the corresponding feature center vector Ucj, and P=|vx-μx|. Compare the deviation value P of the eigenvalue of each historical data feature vector with the preset deviation threshold P0, and filter the eigenvalues ​​that satisfy the deviation value P greater than the preset deviation threshold P0, so as to obtain the historical data feature vector Vcj of the corresponding category.

[0089] Step S300 includes:

[0090] S301. Based on the mapping relationship between the historical data features and the corresponding events, combined with the log records, the time period T1 of the fault event of each industrial device is obtained, the industrial device numbers with overlapping parts in the time period T1 are extracted, and arranged in the order of the fault event occurrence timestamp t_event of the corresponding industrial device, with the fault event number of the first industrial device as the index, so as to construct the associated event sequence of the fault events of different industrial devices, and each component of the associated event sequence is represented as: (i, E, t_event), where i represents the industrial device number, E represents the fault event code, and t_event represents the fault event occurrence timestamp;

[0091] S302. For each category corresponding to the fault event of each industrial equipment, summarize the associated event sequences indexed by the event code of the corresponding category, and count the number N of the same associated event sequences and the total number N1 of fault events of the corresponding category, calculate the proportion Z of the associated event sequences, and Z=N / N1; filter the associated event sequences whose proportion Z is greater than or equal to the threshold Z0, find the corresponding industrial equipment number according to the filtered associated event sequence, and find the corresponding position coordinates of the industrial equipment in the monitoring area within the monitoring area based on the industrial equipment number, so as to identify the fault isolation area, and set the industrial equipment corresponding to the index of the filtered associated event sequence as the key point of the fault isolation area; according to the filtered associated event sequence, obtain the automation control strategy of the corresponding fault event in the log record, and correspond the fault isolation area and the key point of the fault isolation area to the corresponding automation control strategy.

[0092] Step S400 includes:

[0093] S401. Obtain the real-time data of all industrial equipment in the monitoring area, analyze the real-time data according to the analysis method of historical data, so as to extract the real-time data features and form a real-time data feature vector; perform similarity calculation on the real-time data feature vector and the historical data feature vector of the corresponding industrial equipment in turn, obtain the historical data feature vector corresponding to the maximum value of the similarity calculation as the matching target, and compare the similarity calculation result with the matching target with the preset threshold; if the similarity calculation result is less than the preset threshold, obtain the event number corresponding to the matching target; if the event number corresponding to the matching target belongs to a normal event, output notification information to relevant personnel, and the relevant personnel determine whether there is a fault event. If the judgment result is a normal event, the real-time data features corresponding to the current event are stored in the log record and marked as a normal event; if the judgment result is a fault event, record the control strategy of the relevant personnel, and store the real-time data features corresponding to the current event in the log record, and mark it as a fault event;

[0094] S402. If the similarity calculation result is greater than the preset threshold, the event number corresponding to the matching target is obtained; if the event number corresponding to the matching target belongs to a normal event, no processing is performed; if the event number corresponding to the matching target belongs to a fault event, the fault warning information is output to relevant personnel, and the fault warning information includes the fault isolation area and the key points of the fault isolation area, and the corresponding automation control strategy is output according to the fault isolation area and the key points of the fault isolation area.

[0095] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0096] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An automated control method based on big data analysis, characterized in that: The method comprises the following steps: Step S100. Collect historical data and log records of all industrial equipment in the monitoring area, filter out historical data of industrial equipment under different events through log records, and construct a historical data set; the historical data refers to the production-related operating parameters collected by the corresponding sensor equipment when the industrial equipment is started; the log record refers to the event record corresponding to the industrial equipment; Step S200. Analyze the historical data trends of each industrial device under different events according to the historical data set, and extract corresponding historical data features based on the historical data trends; for each historical data set, analyze the historical data features corresponding to the corresponding industrial device, and establish a mapping relationship between the historical data features and the corresponding events; Step S300. Based on the mapping relationship between historical data features and corresponding events, combined with log records, the events corresponding to different industrial equipment are subjected to correlation analysis, and the fault isolation area and the key points of the fault isolation area are identified according to the correlation analysis results; the automation control strategy of the corresponding event is extracted from the log records, and the fault isolation area and the key points of the fault isolation area are matched with the corresponding automation control strategy; Step S400. Acquire real-time data of all industrial equipment in the monitoring area, extract real-time data features from the real-time data, and analyze the correlation between the real-time data features and the historical data features; determine the current event based on the correlation analysis results, and perform corresponding processing based on the current event; The step S300 includes: S301. Based on the mapping relationship between the historical data features and the corresponding events, combined with the log records, the time period T1 of the fault event of each industrial device is obtained, the industrial device numbers with overlapping parts in the time period T1 are extracted, and arranged in the order of the fault event occurrence timestamp t_event of the corresponding industrial device, with the fault event number of the first industrial device as the index, so as to construct the associated event sequence of the fault events of different industrial devices, and each component of the associated event sequence is represented as: (i, E, t_event), where i represents the industrial device number, E represents the fault event code, and t_event represents the fault event occurrence timestamp; S302. For each category corresponding to the fault event of each industrial equipment, summarize the associated event sequences indexed by the event code of the corresponding category, and count the number N of the same associated event sequences and the total number N1 of fault events of the corresponding category, calculate the proportion Z of the associated event sequences, and Z=N / N1; filter the associated event sequences whose proportion Z is greater than or equal to the threshold Z0, find the corresponding industrial equipment number according to the filtered associated event sequence, and find the corresponding position coordinates of the industrial equipment in the monitoring area within the monitoring area based on the industrial equipment number, so as to identify the fault isolation area, and set the industrial equipment corresponding to the index of the filtered associated event sequence as the key point of the fault isolation area; according to the filtered associated event sequence, obtain the automation control strategy of the corresponding fault event in the log record, and correspond the fault isolation area and the key point of the fault isolation area to the corresponding automation control strategy.

2. The automatic control method based on big data analysis according to claim 1, characterized in that: The step S100 includes: S101. According to the set time window T, collect historical data and log records of all industrial equipment in the monitoring area, and T = [t0, tn], where t0 represents the start time of data collection and tn represents the end time of data collection; the events recorded in the log include normal events and fault events, the normal event means that the industrial equipment operates according to the predetermined operating parameters and specifications, and there is no abnormality or fault; the fault event means that the industrial equipment deviates from the normal working state; S102. For the historical data and log records corresponding to each industrial equipment in the monitoring area, according to the event occurrence timestamp t_event in the log record, filter out the historical data occurring in the time period T1 before and after the timestamp t_event, and T1 = [t_event-ΔT, t_event+ΔT], where ΔT represents the time span; for the historical data of each event in the time period T1, perform data preprocessing, and convert the historical data into a unified format to form a historical data set Li, and Li = {li1, li2, ..., lim}, where Li represents the historical data set of the i-th industrial equipment in the monitoring area, i represents the number of the industrial equipment in the monitoring area, li1 represents the historical data corresponding to the first event of the i-th industrial equipment in the monitoring area, li2 represents the historical data corresponding to the second event of the i-th industrial equipment in the monitoring area, and so on, lim represents the historical data corresponding to the m-th event of the i-th industrial equipment in the monitoring area, and m represents the total number of events in the log records of the i-th industrial equipment in the monitoring area; wherein the number of historical data sets is equal to the number of industrial equipment in the monitoring area.

3. The automatic control method based on big data analysis according to claim 2 is characterized in that: The step S200 includes: S201. For each element in the historical data set Li, the elements in the historical data set Li are divided into several categories according to the event code recorded in the log, and the event code corresponding to each category is the same, wherein there is only one normal event code in the event recorded in the log, and there is more than one code for the fault event in the event recorded in the log; trend analysis is performed on the historical data corresponding to the element of each category, and the specific analysis process is as follows: Set the fitting function to f(t), which represents the trend of historical data over time: f(t) = a0+a1t+a2t 2 +...+a k t k , where a0, a1, a2, ..., a k represents the fitting parameters, t represents the time; the least squares method is used to solve the parameters a0, a1, a2, ..., a k , so that the sum of squared errors is minimized: ∑t k ∈T1,[f(t k )-y k ] 2 , where y k represents the actual observed data; Calculate the derivative f'(t) of the fitting function to evaluate the rate of change of historical data over time, thereby obtaining the trend change speed of historical data: S202. Based on the trend analysis results of historical data, the historical data features of each element are extracted and normalized to establish a historical data feature vector V, and V = [v1, v2, ..., vs], where v1 represents the first historical data feature of the element in the historical data set, v2 represents the second historical data feature of the element in the historical data set, and so on, vs represents the sth historical data feature of the element in the historical data set; according to the category to which each element belongs, the corresponding historical data feature vectors are summarized, and the historical data feature vectors of each category are analyzed to obtain the historical data feature vector Vcj of each category, where Vcj represents the historical data feature vector of the jth category, and the maximum value of j is equal to the number of categories of elements in the historical data set; according to the event code corresponding to each category, a mapping relationship between the historical data feature vector Vcj and the corresponding event is established, which is expressed as: M = {(Vcj, Ej) | j = 1, 2, ..., N}, where Ej represents the event code corresponding to the jth category, and N represents the number of categories.

4. The automatic control method based on big data analysis according to claim 3 is characterized in that: The specific analysis process of the historical data feature vector Vcj of each category is as follows: According to the historical data feature vector of each category cj, the average value μx of each eigenvalue of the historical data feature vector is calculated, where x represents the number of the average value of the eigenvalue, ranging from 1 to s; according to the average value μx of the eigenvalue of the historical data feature vector, a feature center vector Ucj is formed, and Ucj=[μ1,μ2,...,μs]j, where Ucj represents the feature center vector of the jth category, and [μ1,μ2,...,μs]j represents the eigenvalue of the feature center vector of the jth category; Calculate the deviation value P between the eigenvalue vx of each historical data feature vector and the eigenvalue μx of the corresponding feature center vector Ucj, and P = |vx-μx|. Compare the deviation value P of the eigenvalue of each historical data feature vector with the preset deviation threshold P0, and filter the eigenvalues ​​that satisfy the deviation value P greater than the preset deviation threshold P0, so as to obtain the historical data feature vector Vcj of the corresponding category.

5. The automatic control method based on big data analysis according to claim 1, characterized in that: The step S400 includes: S401. Obtain the real-time data of all industrial equipment in the monitoring area, analyze the real-time data according to the analysis method of historical data, so as to extract the real-time data features and form a real-time data feature vector; perform similarity calculation on the real-time data feature vector and the historical data feature vector of the corresponding industrial equipment in turn, obtain the historical data feature vector corresponding to the maximum value of the similarity calculation as the matching target, and compare the similarity calculation result with the matching target with the preset threshold; if the similarity calculation result is less than the preset threshold, obtain the event number corresponding to the matching target; if the event number corresponding to the matching target belongs to a normal event, output notification information to relevant personnel, and the relevant personnel determine whether there is a fault event. If the judgment result is a normal event, the real-time data features corresponding to the current event are stored in the log record and marked as a normal event; if the judgment result is a fault event, record the control strategy of the relevant personnel, and store the real-time data features corresponding to the current event in the log record, and mark it as a fault event; S402. If the similarity calculation result is greater than the preset threshold, the event number corresponding to the matching target is obtained; if the event number corresponding to the matching target belongs to a normal event, no processing is performed; if the event number corresponding to the matching target belongs to a fault event, the fault warning information is output to relevant personnel, and the fault warning information includes the fault isolation area and the key points of the fault isolation area, and the corresponding automation control strategy is output according to the fault isolation area and the key points of the fault isolation area.

6. An automated control system based on big data analysis, applied to an automated control method based on big data analysis as claimed in any one of claims 1 to 5, characterized in that: The system includes: a data acquisition and data processing module, a data analysis and feature extraction module, a fault analysis and isolation area identification module, a real-time data analysis and event identification module, and an automatic control strategy execution and fault response module; The data acquisition and data processing module collects historical data and log records of all industrial equipment in the monitoring area, filters out historical data of industrial equipment under different events through log records, and constructs a historical data set; The data analysis and feature extraction module analyzes the historical data trends of each industrial device under different events according to the historical data set, and extracts the corresponding historical data features based on the historical data trends; for each historical data set, analyzes the historical data features corresponding to the corresponding industrial device, and establishes a mapping relationship between the historical data features and the corresponding events; The fault analysis and isolation area identification module performs correlation analysis on events corresponding to different industrial equipment based on the mapping relationship between historical data features and corresponding events in combination with log records, and identifies fault isolation areas and key points of fault isolation areas according to the correlation analysis results; and extracts the automation control strategy of the corresponding event from the log records, and corresponds the fault isolation area and key points of the fault isolation area to the corresponding automation control strategy; The real-time data analysis and event recognition module acquires the real-time data of the industrial equipment in the monitoring area in real time, and extracts the real-time data features according to the historical data analysis method to form a real-time data feature vector; performs similarity calculation on the real-time data feature vector and the historical data feature vector, and determines whether the current event is a normal event or a fault event based on the similarity calculation result; The automatic control strategy execution and fault response module generates corresponding notification information according to the identification results of the real-time data analysis and event identification module. If there is fault warning information, the system will execute the corresponding automatic control strategy.

7. The automatic control system based on big data analysis according to claim 6 is characterized in that: The data acquisition and data processing module includes a data collection unit and a data processing unit; The data collection unit collects historical data and log records of all industrial equipment in the monitoring area; the data processing unit preprocesses and converts the format of the collected historical data, and constructs a historical data set based on the log records; The data analysis and feature extraction module includes a trend analysis unit, a feature extraction unit and a category classification unit; The trend analysis unit performs trend analysis on the historical data of each industrial device; The feature extraction unit extracts corresponding historical data features according to the trend analysis results of the historical data, and normalizes the historical data features to form a historical data feature vector; The category classification unit classifies the historical data feature vectors into different categories according to the event codes in the log records; and calculates the feature center vector of each category, extracts the historical data features of each category and forms a corresponding historical data feature vector.

8. The automatic control system based on big data analysis according to claim 6, characterized in that: The fault analysis and isolation area identification module includes an event correlation analysis unit, a fault isolation area identification unit, and a control strategy extraction and mapping unit; The event correlation analysis unit performs correlation analysis on fault events based on the mapping relationship between historical data features and events in combination with log records; the fault isolation area identification unit identifies fault isolation areas through correlation analysis results and marks the key points of the fault isolation areas of these areas; the control strategy extraction and mapping unit extracts the automation control strategy related to the fault event from the log records, and maps the automation control strategy with the fault isolation area and the key points of the fault isolation area.

9. The automatic control system based on big data analysis according to claim 6, characterized in that: The real-time data analysis and event recognition module includes a real-time data acquisition unit, a real-time data feature extraction unit and an event recognition unit; The real-time data acquisition unit collects the real-time data of all industrial equipment; the real-time data feature extraction unit analyzes the real-time data of all industrial equipment according to the analysis process of historical data, thereby extracting real-time data features; the event recognition unit determines the type of event currently occurring based on the similarity calculation between the real-time data features and the historical data features; The automated control strategy execution and fault response module includes a control strategy execution unit and a fault warning and notification unit; The control strategy execution unit executes the automated control operation according to the corresponding control strategy after identifying the fault event; The fault warning and notification unit sends fault warning information to relevant personnel and provides location information of the fault isolation area and key points when a fault event is identified based on the identification result of the event identification unit.

Images