Intelligent early warning diagnosis method realized based on intelligent power plant

Abstract

The invention discloses an intelligent early warning diagnosis method based on an intelligent power plant, and the method comprises the steps: detecting an abnormality through employing an unsupervised learning method, and correcting a detection result in combination with a heuristic rule; adjusting the length of a dynamic time window according to the running state of the equipment, ensuring the requirement of data trend analysis, carrying out smoothing processing on trend data captured in the adjusted time window, and extracting fault trend characteristics; and performing abnormal distribution analysis on the fault trend characteristics, performing layered early warning on the faults according to abnormal levels, constructing a step-by-step troubleshooting model based on a Bayesian network, and analyzing fault causes step by step. According to the intelligent early warning diagnosis method, the influence of equipment faults on a power plant system can be effectively reduced, and the overall safety and stability of a power plant are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent early warning diagnosis of power plants, and in particular to an intelligent early warning diagnosis method based on a smart power plant. Background Art

[0002] In modern smart power plants, the continuous and efficient operation of equipment is crucial. However, during long-term operation, equipment is easily affected by environmental factors, load fluctuations, and aging of internal components, which can lead to abnormal operating parameters and possible failures. Traditional early warning diagnosis methods mainly rely on regular inspections and simple threshold alarms, which make it difficult to identify potential failures or abnormal trends in a timely and accurate manner. With the rapid development of big data and artificial intelligence technologies, data-driven intelligent early warning diagnosis systems have gradually become a key requirement for power plant equipment maintenance and safety management. Therefore, how to accurately identify abnormal operating conditions before significant equipment failures occur and quickly locate the source of the fault has become an important challenge in existing technologies. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides an intelligent early warning diagnosis method based on a smart power plant, aiming to solve the problem of insufficient real-time and accuracy in early warning diagnosis of smart power plant equipment. Traditional early warning methods based on fixed thresholds or regular inspections cannot flexibly respond to the complex operating conditions of equipment under different working conditions, and it is difficult to quickly lock the source of the fault. The present invention proposes an intelligent early warning diagnosis method based on a dynamic time window, unsupervised anomaly detection, heuristic rule correction and Bayesian network. By dynamically adjusting the analysis window, detecting abnormal distribution in real time and gradually troubleshooting the root cause of the fault, it can more accurately identify and diagnose equipment failures, ensure that the system can promptly detect potential risks under complex working conditions, and avoid the impact of sudden equipment failures on power plant operations.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions, an intelligent early warning diagnosis method based on a smart power plant, comprising:

[0006] Use unsupervised learning methods to detect anomalies and combine heuristic rules to correct the detection results;

[0007] Adjust the length of the dynamic time window according to the equipment operation status to ensure the needs of data trend analysis, smooth the trend data captured in the adjusted time window, and extract fault trend characteristics;

[0008] The abnormal distribution of fault trend characteristics is analyzed, and faults are warned in layers according to the abnormality level. At the same time, a step-by-step troubleshooting model is built based on the Bayesian network to gradually analyze the cause of the fault.

[0009] As a preferred solution of the intelligent early warning and diagnosis method based on a smart power plant described in the present invention, the use of an unsupervised learning method to detect anomalies includes establishing a normal behavior model in the equipment operation data through an unsupervised learning algorithm, comparing the degree of deviation between the equipment operation data and the normal behavior model, and detecting abnormal patterns.

[0010] As a preferred solution of the intelligent early warning diagnosis method based on the smart power plant described in the present invention, wherein: the correction of the detection result in combination with the heuristic rule includes, after the unsupervised learning algorithm detects the abnormal mode, comparing and analyzing the detection result with the known fault mode in the equipment historical data; the system establishes a heuristic rule base according to the characteristic conditions of the equipment historical faults, and the rule base contains the parameter range, correlation and characteristic change mode of the fault mode, that is, the equipment operation characteristics under different temperatures, pressures and vibration amplitudes;

[0011] When a new anomaly is detected, the system analyzes the parameter characteristics of the abnormal pattern one by one according to the heuristic rules. The parameter characteristics include the current fluctuation amplitude and the temperature rise rate, and determines whether they meet the characteristic conditions of a known fault. The parameter values ​​in the abnormal pattern are compared with the characteristic parameters of each known fault in the rule base one by one, and the similarity is calculated. If the similarity is high, the anomaly is classified into the corresponding known fault type; if the similarity is lower than the preset error standard, the system marks the abnormal pattern as "unknown anomaly" and keeps the record.

[0012] As a preferred solution of the intelligent early warning diagnosis method based on the smart power plant described in the present invention, wherein: the adjustment of the dynamic time window length includes dynamically adjusting the time window length of the data analysis according to the real-time operating status of the equipment and the load change;

[0013] The system continuously monitors the operating parameters of the equipment, including power, current, temperature, and vibration, and conducts real-time evaluation of the current load changes of the equipment. When the load changes exceed the preset range, the frequency of fluctuations in the equipment operating parameters increases sharply, and the system automatically shortens the time window. On the contrary, when the load change is lower than the preset range and the equipment is in normal operation, the system extends the time window, increases the time span of each data collection, smoothes out short-term fluctuations in the equipment data, and facilitates the extraction of long-term trend characteristics.

[0014] As a preferred solution of the intelligent early warning and diagnosis method based on a smart power plant described in the present invention, the smoothing processing includes using an improved smoothing algorithm that combines moving average and exponential smoothing to identify long-term trend characteristics in equipment data and sensitively detect changes in equipment operating status.

[0015] As a preferred solution of the intelligent early warning diagnosis method based on the smart power plant described in the present invention, the abnormal distribution analysis includes: if the frequency of the abnormal value is lower than the preset threshold, and the abnormal amplitude is within the preset range of the normal value, and the abnormal value has no obvious concentration in the time distribution, then the equipment operation is evaluated as a low-risk state; when the abnormal value appears in the data no more than 3 times in the current cycle, and the deviation amplitude is less than the early warning threshold, it is considered that the current abnormality is caused by environmental fluctuations or short-term load changes of the equipment, and no intervention is required;

[0016] If the frequency of abnormal values ​​reaches the preset medium threshold, and the amplitude deviates from the normal operating value, indicating that the equipment is unstable, it is classified as a medium risk level; when the abnormal values ​​are concentrated in time or appear frequently in a specific time period, it indicates that the equipment has intermittent failures or load fluctuations, and preventive inspections are required;

[0017] If the frequency of abnormal values ​​exceeds the high-risk preset threshold, the amplitude deviates seriously from the normal value, and the abnormal values ​​are highly concentrated on the time axis, the equipment status will be classified as a high-risk level; when the abnormal frequency increases significantly, that is, a sharp increase occurs, and abnormalities occur in each sampling cycle, indicating that the equipment has entered an unstable or out-of-control state, the system should trigger an alarm and recommend immediate inspection or shutdown maintenance.

[0018] As a preferred solution of the intelligent early warning diagnosis method based on the smart power plant described in the present invention, the hierarchical early warning includes: if the fault is assessed as low risk, the system starts a real-time monitoring program for this type of fault, continuously records low-frequency abnormal conditions and corresponding data parameters, and performs periodic inspections on them;

[0019] When a fault is classified as medium risk, the system automatically triggers the diagnostic program to conduct an in-depth analysis of the cause of the medium risk fault. The diagnostic program calls the historical data and current operation data of the equipment to check the possible sources of the fault, identify potential abnormal patterns, and generate a fault report.

[0020] If the fault reaches a high-risk level, the system will immediately trigger an alarm and upload the high-risk data and analysis results to the central system. After receiving the data, the central system will conduct an in-depth analysis based on the operating history data of the database and rule base, and provide fault handling suggestions. At the same time, the system will notify relevant operators to take emergency measures.

[0021] As a preferred solution of the intelligent early warning and diagnosis method based on a smart power plant described in the present invention, the step-by-step troubleshooting model constructed based on the Bayesian network includes using the main operating parameters of the equipment, including voltage, current, temperature, and pressure, as nodes. The edges between the nodes represent the causal relationship between different parameters. By setting the conditional probability relationship between the nodes, the failure probability of each node is calculated based on the observed data when a failure occurs, and the source of the failure is locked.

[0022] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, it implements the steps of an intelligent early warning diagnosis method based on a smart power plant.

[0023] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of an intelligent early warning diagnosis method based on a smart power plant are implemented.

[0024] Beneficial effects of the present invention: The present invention provides an intelligent early warning and diagnosis method that can dynamically adapt to the operating status of the equipment, realizes hierarchical early warning for different risk levels and gradual investigation and location of the source of the fault. By adopting the anomaly detection technology of dynamic time window adjustment, unsupervised learning combined with heuristic rules, the present invention significantly improves the accuracy of anomaly identification and diagnosis of operating data. At the same time, the adaptive conditional probability update algorithm based on the Bayesian network enables the system to continuously optimize the fault troubleshooting model and maintain efficient tracking of the root cause of the fault under different operating conditions. The intelligent early warning and diagnosis method of the present invention can effectively reduce the impact of equipment failures on the power plant system and improve the overall safety and stability of the power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0026] Figure 1 A flowchart of an intelligent early warning diagnosis method based on a smart power plant is provided as an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0028] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0030] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0031] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0032] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0033] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and provides an intelligent early warning diagnosis method based on a smart power plant, comprising:

[0034] S1: Use unsupervised learning methods to detect anomalies and combine heuristic rules to correct the detection results.

[0035] S2: Adjust the length of the dynamic time window according to the equipment operation status to ensure the needs of data trend analysis, smooth the trend data captured in the adjusted time window, and extract fault trend characteristics.

[0036] S3: Analyze the abnormal distribution of fault trend characteristics, and issue layered warnings for faults according to the abnormality level. At the same time, build a step-by-step troubleshooting model based on the Bayesian network to gradually analyze the cause of the fault.

[0037] The use of an unsupervised learning method to detect anomalies includes establishing a normal behavior model in the equipment operation data through an unsupervised learning algorithm, comparing the degree of deviation between the equipment operation data and the normal behavior model, and detecting abnormal patterns.

[0038] Assume that the historical normal data is a multidimensional vector X = {x1, x2, ..., x N},in Represents the value of d parameters. Modeling using GMM:

[0039]

[0040] Among them, π k is the weight of the kth Gaussian component, satisfying μ k and Σ k are the mean vector and covariance matrix of the kth Gaussian component respectively.

[0041] Assume the real-time data of the device is a vector The Mahalanobis distance relative to the normal behavior model is calculated to measure the degree of deviation of y from the normal behavior model. The improved Mahalanobis distance is defined as follows:

[0042]

[0043] Among them, μ k and Σ k are the mean and covariance matrix of the kth Gaussian component in the GMM model. By selecting the component closest to the real-time data This distance can more flexibly adapt to different working conditions.

[0044] Mahalanobis distance Compare with the dynamic threshold δ to determine whether an abnormality occurs:

[0045]

[0046] The dynamic threshold δ is automatically adjusted according to the operating data of the normal behavior model to adapt to changes in the equipment operating conditions. This dynamic adjustment can be set based on the standard deviation of the operating data or the quantile on a specific distribution.

[0047] The correction of the detection results by combining heuristic rules includes, after the unsupervised learning algorithm detects the abnormal mode, comparing and analyzing the detection results with the known failure modes in the equipment history data; the system establishes a heuristic rule base according to the characteristic conditions of the equipment history failure, and the rule base contains the parameter range, correlation and characteristic change mode of the failure mode, that is, the equipment operation characteristics under different temperatures, pressures and vibration amplitudes;

[0048] When a new anomaly is detected, the system analyzes the parameter characteristics of the abnormal pattern one by one according to the heuristic rules. The parameter characteristics include the current fluctuation amplitude and the temperature rise rate, and determines whether they meet the characteristic conditions of a known fault. The parameter values ​​in the abnormal pattern are compared with the characteristic parameters of each known fault in the rule base one by one, and the similarity is calculated. If the similarity is high, the anomaly is classified into the corresponding known fault type; if the similarity is lower than the preset error standard, the system marks the abnormal pattern as "unknown anomaly" and keeps the record.

[0049] The adjusting of the dynamic time window length includes dynamically adjusting the time window length of data analysis according to the real-time operation status of the equipment and the load change situation;

[0050] The system continuously monitors the operating parameters of the equipment, including power, current, temperature, and vibration, and conducts real-time evaluation of the current load changes of the equipment. When the load changes exceed the preset range, the frequency of fluctuations in the equipment operating parameters increases sharply, and the system automatically shortens the time window. On the contrary, when the load change is lower than the preset range and the equipment is in normal operation, the system extends the time window, increases the time span of each data collection, smoothes out short-term fluctuations in the equipment data, and facilitates the extraction of long-term trend characteristics.

[0051] The smoothing process includes using an improved smoothing algorithm combining moving average and exponential smoothing to identify long-term trend characteristics in equipment data and sensitively detect changes in equipment operating status.

[0052] Specifically, the "double smoothing" method is used, combining moving average and exponential smoothing to capture short-term changes in data and extract long-term trends. The specific model is as follows:

[0053] For the time series {x t}, first use the moving average with a window size of N to smooth the data and calculate the moving average

[0054]

[0055] in Represents smoothed data at time t. Moving average can reduce short-term data fluctuations and prepare for subsequent trend extraction.

[0056] Based on the moving average, Perform an exponential smoothing process to further extract long-term trend characteristics. Suppose the smoothing coefficient of exponential smoothing is α (0<α<1), and the value S after exponential smoothing is t Expressed as:

[0057]

[0058] Where S t It represents the smoothed trend value at time t, and α controls the smoothing speed. In high volatility periods, α is appropriately increased, and in stable periods, α is reduced.

[0059] In order to identify possible trend change points in the smoothed data, a trend deviation detection formula is further introduced. Define trend deviation D t is the difference between the current smoothed value and the previous trend value:

[0060]

[0061] If D t If it exceeds the set threshold δ, it is determined that a significant trend deviation occurs at this point, which may indicate a change in the equipment status.

[0062] The abnormal distribution analysis includes: if the frequency of abnormal values ​​is lower than the preset threshold, and the abnormal amplitude is within the preset range of normal values, and the abnormal values ​​have no obvious concentration in time distribution, then the equipment operation is evaluated as a low-risk state; when the abnormal value does not appear in the data more than 3 times in the current cycle, and the deviation amplitude is less than the warning threshold, then it is considered that the current abnormality is caused by environmental fluctuations or short-term load changes of the equipment, and no intervention is required;

[0063] If the frequency of abnormal values ​​reaches the preset medium threshold, and the amplitude deviates from the normal operating value, indicating that the equipment is unstable, it is classified as a medium risk level; when the abnormal values ​​are concentrated in time or appear frequently in a specific time period, it indicates that the equipment has intermittent failures or load fluctuations, and preventive inspections are required;

[0064] If the frequency of abnormal values ​​exceeds the high-risk preset threshold, the amplitude deviates seriously from the normal value, and the abnormal values ​​are highly concentrated on the time axis, the equipment status will be classified as a high-risk level; when the abnormal frequency increases significantly, that is, a sharp increase occurs, and abnormalities occur in each sampling cycle, indicating that the equipment has entered an unstable or out-of-control state, the system should trigger an alarm and recommend immediate inspection or shutdown maintenance.

[0065] The hierarchical warning includes that if a fault is assessed as low risk, the system starts a real-time monitoring program for this type of fault, continuously records low-frequency abnormal conditions and corresponding data parameters, and performs periodic checks on them;

[0066] When a fault is classified as medium risk, the system automatically triggers the diagnostic program to conduct an in-depth analysis of the cause of the medium risk fault. The diagnostic program calls the historical data and current operation data of the equipment to check the possible sources of the fault, identify potential abnormal patterns, and generate a fault report.

[0067] If the fault reaches a high-risk level, the system will immediately trigger an alarm and upload the high-risk data and analysis results to the central system. After receiving the data, the central system will conduct an in-depth analysis based on the operating history data of the database and rule base, and provide fault handling suggestions. At the same time, the system will notify relevant operators to take emergency measures.

[0068] The step-by-step troubleshooting model based on the Bayesian network includes using the main operating parameters of the equipment, including voltage, current, temperature, and pressure, as nodes. The edges between the nodes represent the causal relationship between different parameters. By setting the conditional probability relationship between the nodes, the failure probability of each node is calculated based on the observed data when a failure occurs, and the source of the failure is locked.

[0069] To achieve more effective troubleshooting, the model introduces an adaptive conditional probability update algorithm based on Bayesian networks. This algorithm can dynamically adjust the conditional probability of nodes when the system updates data, making the model more in line with the real-time operating status.

[0070] Assume that there are n nodes in the Bayesian network X={X1,X2,…,X n ], each node represents an operating parameter. The joint probability distribution P(X) of the nodes is expressed as:

[0071]

[0072] Among them, Pa(X i ) indicates X i The parent node set of is used to estimate each conditional probability through historical data.

[0073] In order to improve the adaptability of the model under different device states, an adaptive update formula is used to adjust the conditional probability according to the latest observation data. i The current conditional probability is P(X i ∣X j ), after observing the new data set D, the conditional probability is updated using Bayesian estimation:

[0074]

[0075] If the observed data frequently show some deviation, then gradually adjust P(X i ∣X j ) to make it consistent with the observed data.

[0076] When a fault occurs, suppose we observe that node X k An exception occurs at the node, and the posterior probability P(X i ∣X k ), check the fault source layer by layer. The posterior probability is calculated as follows:

[0077]

[0078] By calculating the posterior probability from the environmental nodes to the internal nodes layer by layer, the cause of the fault is determined, and irrelevant nodes are gradually eliminated until the source of the fault is locked.

[0079] Specifically, X in the formula represents all the node sets in the entire Bayesian network. i represents the ith node in the Bayesian network, representing an operating parameter of the device (such as voltage, current, temperature, etc.). i ) represents node X i The parent node set of X i There is a direct causal relationship between the parent node and the parent node. The state of the parent node will affect X i state. P(X) represents the joint probability distribution of the node set X, which represents the joint probability of all node states in the entire Bayesian network. i ∣Pa(X i )) indicates that given X i In the parent node state, X i The conditional probability of X is obtained through statistical calculation of historical data. i The probability relationship between P(X i ∣X j ,D) means that under the given condition X j And in the case of the new dataset D, node X i The conditional probability of X is obtained by updating the Bayesian estimation. i The latest probability distribution after observing D. P(X k ) represents node X k The prior probability of X before considering other node states k The probability of P(X j ) represents node X j The prior probability of P(X k ∣X i ) means that at a given node X i In the case of state, node Xk The conditional probability of P(X i ∣X k ) means that at a given node X k In the case of node X i The posterior probability is used to calculate other nodes in X when troubleshooting the source of the fault. k The possibility of abnormality.

[0080] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0081] Example 2

[0082] The second embodiment of the present invention is different from the previous embodiment in that:

[0083] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0084] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0085] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0087] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0088] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent early warning diagnosis method based on a smart power plant, characterized by: include, Use unsupervised learning methods to detect anomalies and combine heuristic rules to correct the detection results; Adjust the length of the dynamic time window according to the equipment operation status to ensure the needs of data trend analysis, smooth the trend data captured in the adjusted time window, and extract fault trend characteristics; The abnormal distribution of fault trend characteristics is analyzed, and faults are warned in layers according to the abnormality level. At the same time, a step-by-step troubleshooting model is built based on the Bayesian network to gradually analyze the cause of the fault.

2. The intelligent early warning diagnosis method based on a smart power plant according to claim 1, characterized in that: The use of an unsupervised learning method to detect anomalies includes establishing a normal behavior model in the equipment operation data through an unsupervised learning algorithm, comparing the degree of deviation between the equipment operation data and the normal behavior model, and detecting abnormal patterns.

3. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 2 is characterized by: The correction of the detection results by combining heuristic rules includes, after the unsupervised learning algorithm detects the abnormal mode, comparing and analyzing the detection results with the known failure modes in the equipment history data; the system establishes a heuristic rule base according to the characteristic conditions of the equipment history failure, and the rule base contains the parameter range, correlation and characteristic change mode of the failure mode, that is, the equipment operation characteristics under different temperatures, pressures and vibration amplitudes; When a new anomaly is detected, the system analyzes the parameter characteristics of the abnormal pattern one by one according to the heuristic rules. The parameter characteristics include the current fluctuation amplitude and the temperature rise rate, and determines whether it meets the characteristic conditions of a known fault. The parameter values ​​in the abnormal pattern are compared with the characteristic parameters of each known fault in the rule base one by one, and the similarity is calculated. If the similarity is high, the anomaly is classified into the corresponding known fault type; if the similarity is lower than the preset error standard, the system marks the abnormal pattern as "unknown anomaly" and keeps the record.

4. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 3 is characterized by: The adjusting of the dynamic time window length comprises: Dynamically adjust the time window length of data analysis according to the real-time operating status of the equipment and load changes; The system continuously monitors the operating parameters of the equipment, including power, current, temperature, and vibration, and conducts real-time evaluation of the current load changes of the equipment. When the load changes exceed the preset range, the frequency of fluctuations in the equipment operating parameters increases sharply, and the system automatically shortens the time window. On the contrary, when the load change is lower than the preset range and the equipment is in normal operation, the system extends the time window, increases the time span of each data collection, smoothes out short-term fluctuations in the equipment data, and facilitates the extraction of long-term trend characteristics.

5. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 4 is characterized by: The smoothing process includes using an improved smoothing algorithm combining moving average and exponential smoothing to identify long-term trend characteristics in equipment data and sensitively detect changes in equipment operating status.

6. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 5, characterized in that: The abnormal distribution analysis includes: if the frequency of abnormal values ​​is lower than the preset threshold, and the abnormal amplitude is within the preset range of normal values, and the abnormal values ​​have no obvious concentration in time distribution, then the equipment operation is evaluated as a low-risk state; when the abnormal value does not appear in the data more than 3 times in the current cycle, and the deviation amplitude is less than the warning threshold, then it is considered that the current abnormality is caused by environmental fluctuations or short-term load changes of the equipment, and no intervention is required; If the frequency of abnormal values ​​reaches the preset medium threshold and the amplitude deviates from the normal operating value, indicating that the equipment is unstable, it is classified as medium risk level; When abnormal values ​​have a certain concentration distribution in time, or appear frequently in a specific time period, it indicates that the equipment has intermittent failure or load fluctuation, and preventive inspection is required; If the frequency of abnormal values ​​exceeds the high-risk preset threshold, the amplitude deviates seriously from the normal value, and the abnormal values ​​are highly concentrated on the time axis, the device status is classified as a high-risk level; When the abnormal frequency increases significantly, that is, a sharp increase occurs, and abnormalities occur in each sampling cycle, it indicates that the equipment has entered an unstable or out-of-control state. The system should trigger an alarm and recommend immediate inspection or shutdown maintenance.

7. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 6, characterized in that: The hierarchical warning includes that if a fault is assessed as low risk, the system starts a real-time monitoring program for this type of fault, continuously records low-frequency abnormal conditions and corresponding data parameters, and performs periodic checks on them; When a fault is classified as medium risk, the system automatically triggers the diagnostic program to conduct an in-depth analysis of the cause of the medium risk fault. The diagnostic program calls the historical data and current operation data of the equipment to check the possible sources of the fault, identify potential abnormal patterns, and generate a fault report. If the fault reaches a high-risk level, the system will immediately trigger an alarm and upload the high-risk data and analysis results to the central system. After receiving the data, the central system will conduct an in-depth analysis based on the operating history data of the database and rule base, and provide fault handling suggestions. At the same time, the system will notify relevant operators to take emergency measures.

8. The intelligent early warning diagnosis method based on a smart power plant as claimed in claim 7, characterized in that: The step-by-step troubleshooting model based on the Bayesian network includes using the main operating parameters of the equipment, including voltage, current, temperature, and pressure, as nodes. The edges between the nodes represent the causal relationship between different parameters. By setting the conditional probability relationship between the nodes, the failure probability of each node is calculated based on the observed data when a failure occurs, and the source of the failure is locked.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images