Generator performance detection method, medium and equipment of multi-source heterogeneous big data platform

Abstract

The application provides a generator performance detection method, medium and equipment of a multi-source heterogeneous big data platform, relates to the technical field of performance detection, and comprises the following steps: constructing a multi-source heterogeneous monitoring network of a target generator set; reading predetermined unit operation performance indexes, including primary frequency modulation contribution rate, small disturbance qualified rate and automatic power generation control performance indexes; performing operation performance monitoring to obtain unit operation performance monitoring data; reading predetermined unit operation performance standards, inputting the unit operation performance standards into a unit performance analysis model to obtain unit performance analysis results; performing defect performance marking to generate performance defect early warning information containing a first early warning level; and performing operation parameter regulation and control when a preset early warning level is reached. The application solves the technical problem that the generator performance detection of the prior art is usually performed through offline calculation or evaluation based on historical data, real-time monitoring and real-time adjustment of unit performance are lacking, and system scheduling precision and flexibility are insufficient.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of performance detection, and in particular to a generator performance detection method, medium and equipment of a multi-source heterogeneous big data platform. BACKGROUND

[0002] At present, the traditional generator performance monitoring and evaluation system has a lag, that is, the operating personnel can only obtain the performance evaluation results of the unit on the next day. This time delay makes the unit unable to be immediately adjusted or optimized when performance problems occur. For example, when the AGC performance is lower than the requirement, feedback is often obtained on the next day and improvement measures are taken, which may cause the unit to lose the initiative to participate in bidding for on-grid power, thereby affecting the economic efficiency of the unit. Moreover, the traditional generator unit performance monitoring method usually relies on a single data source or a relatively isolated monitoring system, lacks the integration capability of multi-source heterogeneous data, and cannot realize cross-system data integration and intelligent analysis when facing complex requirements of power grid dispatching. The dispatching instructions between the power grid and the power plant often cannot fully consider the dynamic changes and real-time performance indicators of different units, resulting in insufficient flexibility and accuracy of dispatching decisions. SUMMARY

[0003] The present application provides a generator performance detection method, medium and equipment of a multi-source heterogeneous big data platform, aiming to solve the technical problem that the existing generator performance detection is usually performed through offline calculation or evaluation based on historical data, lacking real-time monitoring and real-time adjustment capability of unit performance, resulting in insufficient system scheduling precision and flexibility.

[0004] The first aspect of the present application provides a generator performance detection method of a multi-source heterogeneous big data platform, which comprises: constructing a multi-source heterogeneous monitoring network of a target generator unit, wherein the multi-source heterogeneous monitoring network comprises a remote terminal unit, a distributed control system and a primary frequency modulation homologous device; reading predetermined unit operation performance indicators, wherein the unit operation performance indicators comprise a primary frequency modulation contribution rate, a small disturbance qualification rate and an automatic generation control performance indicator; performing operation performance monitoring on the target generator unit according to the predetermined unit operation performance indicators through the multi-source heterogeneous monitoring network to obtain unit operation performance monitoring data; reading predetermined unit operation performance standards, inputting the predetermined unit operation performance standards and the unit operation performance monitoring data into a unit performance analysis model to obtain a unit performance analysis result; performing defect performance labeling based on the unit performance analysis result, generating performance defect early warning information according to the defect performance labeling result, wherein the performance defect early warning information contains a first early warning level; and performing operation parameter regulation and control on the target generator unit according to the defect performance labeling result when the first early warning level reaches a preset early warning level.

[0005] In a second aspect of the present disclosure, a storage medium is provided, and the storage medium has a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the generator performance detection method of the multi-source heterogeneous big data platform in the first aspect.

[0006] In a third aspect of the present disclosure, a computer device is provided, and the computer device comprises a memory and a processor, and the memory stores a computer program, and the processor executes the computer program to implement the steps of the generator performance detection method of the multi-source heterogeneous big data platform in the first aspect.

[0007] The one or more technical solutions provided in the present disclosure have at least the following beneficial effects:

[0008] By using the multi-source heterogeneous monitoring network including the remote terminal unit, the distributed control system and the primary frequency modulation homologous device, the operation performance of the target generator set can be monitored in real time. This real-time monitoring can timely capture the deviation of the unit in the primary frequency modulation contribution rate, the small disturbance qualified rate, the automatic power generation control and other performance indicators, so that the diagnosis and measures can be taken at the early stage of the problem, and the potential risks can be reduced. By inputting the unit operation performance monitoring data and the predetermined performance standard into the unit performance analysis model, accurate performance analysis results can be generated. This process can accurately evaluate the current operation state of the unit, timely find performance defects, and generate defect performance markers and warning information to help the operation and maintenance personnel to make timely adjustments. When the performance defects are found, the warning level is automatically generated based on the defect performance markers, and the operation parameters of the unit are regulated and controlled to optimize the operation state of the unit in real time. This optimization control improves the response speed and frequency modulation accuracy of the unit, and ensures that the unit always maintains the best working state. Through comprehensive data processing based on the big data platform, dynamic adjustment and optimization of the unit performance are supported, real-time data acquisition, statistical analysis and multi-source data fusion are performed to ensure the stability and reliability of the system. In addition, the big data platform can provide rich historical data support for future performance evaluation, and lay a foundation for continuous optimization and intelligent decision-making of the system.

[0009] The above description is only a summary of the technical solutions of the present disclosure. In order to more clearly understand the technical means of the present disclosure, the specific embodiments of the present disclosure can be implemented in accordance with the content of the description, and in order to make the above and other purposes, features and advantages of the present disclosure more obvious and easy to understand, the following specific embodiments of the present disclosure are described. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A generator performance detection method flowchart of a multi-source heterogeneous big data platform is provided for the embodiments of the present disclosure.

[0011] Figure 2This application provides a schematic diagram of the operation performance monitoring process in the generator performance testing method of a multi-source heterogeneous big data platform.

[0012] Figure 3 This is a schematic diagram of the structure of an exemplary computer device provided in an embodiment of this application. Detailed Implementation

[0013] This application provides a generator performance testing method, medium, and equipment based on a multi-source heterogeneous big data platform. It solves the technical problem that existing generator performance testing methods typically rely on offline calculations or historical data assessments, lacking real-time monitoring and adjustment capabilities for unit performance, resulting in insufficient system scheduling accuracy and flexibility.

[0014] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0015] Example 1, as Figure 1 As shown in the embodiment of this application, a generator performance testing method for a multi-source heterogeneous big data platform is provided, the method comprising:

[0016] A multi-source heterogeneous monitoring network for the target generator set is constructed, wherein the multi-source heterogeneous monitoring network includes a remote terminal unit, a distributed control system, and a primary frequency regulation device.

[0017] The remote terminal unit is used to collect various operating data of the generator set in real time. These data include important parameters such as temperature, pressure, current, and voltage. The remote terminal unit is responsible for transmitting this data to the control center or data analysis platform through the communication network. Specifically, remote terminal unit equipment is installed in key parts of the generator set (such as the generator and control device), and the equipment is configured to collect data from various parts of the generator set system. The equipment is also set to transmit and provide feedback data to other equipment to ensure efficient and reliable data transmission.

[0018] A distributed control system is a core platform for control and monitoring. It is responsible for receiving, processing, and storing data from remote terminal units. It can perform real-time monitoring and adjust the operating status of the unit according to preset strategies. Specifically, the distributed control system is configured to ensure that it can receive various sensor data from remote terminal units. The processing algorithms and control logic of the distributed control system are set, and it can judge various performance indicators of the unit in real time based on the received data. The distributed control system can make dynamic adjustments according to control requirements, such as adjusting the power output of the generator set and frequency regulation.

[0019] A primary frequency regulation device is used to support the frequency regulation function of generator sets to ensure the stability of the power grid frequency. This device can automatically adjust the output power of the generator set according to the fluctuation of the power grid frequency to respond to the power grid demand. Specifically, a primary frequency regulation device is installed in the generator set to ensure that the equipment can receive the power grid frequency signal in real time. The device is configured to communicate with other systems such as remote terminal units and distributed control systems to ensure the synchronization of data and frequency regulation requirements. A frequency regulation control strategy is set to automatically adjust the output power of the generator set according to the actual situation of power grid frequency fluctuations to achieve real-time frequency regulation.

[0020] Through the collaboration of the above multi-source heterogeneous monitoring networks, the operating status of generator sets can be monitored and adjusted in real time, and the needs of the power grid can be responded to in a timely manner, ensuring that the generator sets can efficiently and reliably support the operation of the power grid in terms of frequency regulation and peak shaving.

[0021] Read the predetermined unit operating performance indicators, which include the primary frequency regulation contribution rate, small disturbance qualification rate, and automatic generation control performance indicators.

[0022] The primary frequency regulation contribution rate refers to the ability of a generator set to respond to frequency regulation signals when the grid frequency fluctuates. Typically, changes in grid frequency can lead to a mismatch between power demand and power generation. The generator set needs to make real-time adjustments based on changes in grid frequency to maintain frequency stability. Specifically, the generator set's response capability under specific frequency fluctuations is calculated by collecting frequency regulation data. This data includes information such as frequency changes and power output changes of the generator set.

[0023] The small disturbance pass rate refers to the ability of a generator set to effectively cope with small disturbances in the power grid (such as load changes or slight frequency fluctuations). The higher the small disturbance pass rate, the more stable the generator set can operate under small power grid fluctuations without affecting the safety of the power grid. Specifically, the stability performance of the generator set under small disturbances is evaluated through historical data analysis and real-time data acquisition. These disturbances include small load fluctuations or slight frequency changes. The pass rate can be measured by the number of times the generator set can recover to a stable state within a certain period of time. A qualified generator set exhibits a high recovery speed and stability.

[0024] Automatic generation control performance indicators refer to the ability of a generator set to automatically adjust its output power to meet the grid load demand after receiving instructions from the grid dispatch center. These indicators cover aspects such as adjustment rate, response accuracy, and stability. Specifically, the performance of the generator set is evaluated by analyzing the data of its automatic generation control system. Typically, the automatic generation control system automatically adjusts the generator set's power output according to changes in grid load to ensure grid load balance. Specific indicators include adjustment accuracy (whether the target power can be accurately achieved), adjustment rate (whether the adjustment speed meets the requirements), and adjustment stability (whether it can maintain a long-term stable operating state).

[0025] By reading and processing the unit's operating data, we ensure that the generator set's performance in frequency regulation, load fluctuation response, and automatic adjustment meets predetermined standards, and continuously monitor and evaluate it. Through real-time analysis of these indicators, we can promptly identify problems in the unit's operation and optimize and adjust the unit based on performance changes.

[0026] According to the predetermined unit operating performance indicators, the target generator unit is monitored for operating performance through the multi-source heterogeneous monitoring network to obtain unit operating performance monitoring data.

[0027] The multi-source heterogeneous monitoring network includes data acquisition units of various sources and types. According to the predetermined operating performance indicators, the monitoring network collects the unit's operating data through various sensors and terminal units. It also performs data cleaning and standardization to eliminate noise, fill in missing data, and ensure the accuracy and reliability of the monitoring data. Finally, it generates unit operating performance monitoring data, which includes all real-time data related to the predetermined operating performance indicators, as the basis for the next step of analysis.

[0028] Read the predetermined unit operating performance standard, input the predetermined unit operating performance standard and the unit operating performance monitoring data into the unit performance analysis model, and obtain the unit performance analysis results.

[0029] Read and refer to the predetermined unit operation performance standards. These standards are set based on historical data analysis, technical specifications and grid requirements. The operation performance standards include reasonable ranges or qualified standards for various performance indicators, such as the standard value of primary frequency regulation contribution rate, the target value of small disturbance qualification rate, and the requirements for AGC regulation accuracy. These standards provide a basis for comparison for subsequent performance analysis and evaluation.

[0030] The unit's operational performance monitoring data and predetermined operational performance standards are input into the unit performance analysis model. This model analyzes the unit's current operating status and determines whether it meets the standards. The model comprises multiple layers: the input layer receives actual monitoring data and predetermined standards; the performance defect identification layer analyzes for performance defects, such as insufficient frequency regulation response or low regulation accuracy; the performance defect analysis layer further analyzes the causes of defects and provides solutions based on historical data and operational trends; and the output layer outputs the analysis results, including whether the standards are met, the severity of the performance defects, and possible improvement measures. Based on the performance analysis model, unit performance analysis results are generated, including a performance defect index to evaluate the unit's performance on various indicators.

[0031] Based on the unit performance analysis results, defect performance is marked, and performance defect early warning information is generated according to the defect performance marking results, wherein the performance defect early warning information includes a first early warning level.

[0032] Defect performance labeling refers to identifying and labeling defects in unit operation based on the results of unit performance analysis. These defects include deviations in performance indicators such as insufficient frequency regulation response, low regulation accuracy, and slow regulation rate. First, potential performance defects are identified through the unit performance analysis model, specifically manifested as certain operating indicators failing to meet predetermined performance standards. Based on these analysis results, a performance label is generated for each defect, such as low frequency contribution rate or small disturbance qualification rate not meeting the standard. Each defect is labeled as a certain type of performance problem, and these defects have different degrees of severity, affecting subsequent system decisions.

[0033] When a unit exhibits performance defects, a performance defect warning message is generated. This message helps maintenance personnel identify problems early and take appropriate measures. The warning message includes the defect type (such as poor frequency regulation response, low AGC adjustment accuracy, etc.), defect severity, and scope of impact (whether it is a local or global problem). The warning message also includes a first warning level, which is the initial warning threshold. This warning level indicates the severity of the problem and is derived from a comprehensive assessment of the defect type and severity. Warning levels are generally divided into multiple levels, such as no alarm, minor alarm, moderate alarm, and severe alarm, with each warning level corresponding to different response measures and control strategies.

[0034] When the first warning level reaches the preset warning level, the operating parameters of the target generator set are adjusted according to the defect performance marking results.

[0035] When the first warning level of performance defects reaches the predetermined warning threshold, the operation parameter adjustment is triggered. The predetermined warning threshold is a preset warning level that can be set according to historical data, operating experience and grid demand. This threshold corresponds to the degree of deviation of a certain performance indicator. If the deviation reaches this value, it means that the unit may affect the stability of the grid and measures need to be taken immediately. This setting ensures that the unit will only be intervened and the parameters adjusted when the problem is relatively serious.

[0036] When the warning level is met, the unit's operating parameters are automatically adjusted based on the defect performance marking results. The purpose of this adjustment is to improve the unit's frequency regulation capability, AGC performance, and load regulation rate. Specifically, based on different types of defects, the generator unit's frequency regulation control parameters are adjusted, such as adjusting the primary frequency regulation contribution rate, the regulation accuracy of the AGC system, and the regulation rate. If the defect is caused by slow unit response speed or low regulation accuracy, the AGC response speed can be optimized or the regulation flexibility increased. The unit's load regulation strategy may also be adjusted to reduce load fluctuations and ensure stable grid operation. In some cases, intelligent optimization algorithms, such as particle swarm optimization and genetic algorithms, are used to dynamically optimize the unit's operating parameters to achieve maximum performance improvement and ensure the unit operates in a more efficient and stable state.

[0037] Furthermore, such as Figure 2 As shown, the step of monitoring the operational performance of the target generator unit according to the predetermined unit operating performance indicators through the multi-source heterogeneous monitoring network to obtain unit operating performance monitoring data includes:

[0038] The target generator set's operational data is collected through the multi-source heterogeneous monitoring network to obtain multi-source heterogeneous operational monitoring data. The multi-source heterogeneous operational monitoring data is preprocessed through a data cleaning domain to obtain standard operational monitoring data. This data cleaning domain includes a data denoising channel, a missing data imputation channel, an anomaly data correction channel, and a data standardization channel. Based on the predetermined unit operational performance indicators, the standard operational monitoring data is used for performance monitoring to obtain the unit's operational performance monitoring data.

[0039] Multi-source heterogeneous monitoring networks achieve comprehensive monitoring of target generator sets by working together through multiple data acquisition systems of different sources and types. Specifically, various operating data of the generator sets are collected through remote terminal units (RTUs) and distributed control systems (DCS) to obtain key parameters such as power output, frequency regulation data, load fluctuation data, temperature, pressure, and humidity. Frequency regulation data collected through primary frequency regulation devices, as well as parameters related to the grid frequency, yield time-series data that records the changes in various operating parameters of the generator sets. This data provides the basis for subsequent analysis and processing.

[0040] The data cleaning domain is a module that performs a series of processing steps on the collected raw data. Its purpose is to remove noise, correct outliers, fill in missing data, and standardize the data to ensure data quality. The data cleaning domain includes multiple processing channels. The data denoising channel removes noise and interference from the data, typically using filtering algorithms such as low-pass and high-pass filters to remove irrelevant signals outside the frequency range. The missing data filling channel handles missing data, typically using interpolation methods such as linear interpolation and spline interpolation, or filling algorithms based on historical data. The outlier correction channel handles outliers caused by equipment failure or acquisition errors; for example, extreme values ​​in the data may not reflect reality and need to be identified and corrected by setting thresholds. The data standardization channel unifies the collected data to a standard range, ensuring that data from different sources and units can be directly compared. Standardization is usually performed by subtracting the mean and dividing by the standard deviation. Standardized operational monitoring data after data cleaning does not contain significant noise and errors, missing data has been appropriately filled, outliers have been corrected, and all data has been standardized on a uniform scale.

[0041] Standard operating monitoring data is compared with predetermined unit operating performance indicators. Performance monitoring is conducted based on this data. For example, the unit's regulation capability during grid frequency fluctuations is monitored based on the primary frequency regulation contribution rate standard; the unit's stability and responsiveness during small load fluctuations are assessed based on the small disturbance pass rate; and the unit's ability to respond quickly and accurately to grid dispatch commands is determined based on the AGC performance indicators. Through these performance monitoring methods, the unit's operating status is evaluated in real time, its performance condition is determined, and the obtained performance monitoring data provides real-time data on various unit performance indicators, laying the foundation for subsequent performance analysis and parameter control.

[0042] Furthermore, the reading of the predetermined unit operating performance standards includes:

[0043] Based on the target generator set, a generator set retrieval source is set, wherein the generator set retrieval source includes the target generator set and multiple similar generator sets of the target generator set; historical generator set operation records are retrieved according to the generator set retrieval source, and a first historical generator set operation record is extracted according to the retrieval results to identify a first generator set operation performance range, wherein the first generator set operation performance standard includes a primary frequency regulation contribution rate range, a small disturbance qualification rate range, and an automatic generation control performance range; the retrieval results are traversed until the identification of the Nth generator set operation performance range is completed, and the intersection of the first generator set operation performance range to the Nth generator set operation performance range is filtered to obtain the predetermined generator set operation performance standard.

[0044] The generator set retrieval source is a collection of historical data and performance records related to the target generator set. The target generator set is the specific generator set undergoing performance evaluation. Multiple similar generator sets are other generator sets that are similar to the target generator set in terms of technical specifications, performance requirements, and functional configuration. The data of these similar generator sets can be used as a control group to obtain a wider range of reference data.

[0045] The generator set retrieval source is searched according to the type of the unit, performance indicators, operating date and other conditions to obtain the historical operating records of the target unit and similar units. These records reflect the performance of the unit under different operating conditions and include the performance data of the unit in different operating stages (such as normal operation, frequency regulation, load fluctuation, etc.).

[0046] Based on the retrieved historical records, the first historical unit operation record is randomly selected as the analysis object. Performance-related indicator data related to the target unit are extracted from this record to identify the first unit's operating performance range. The performance range refers to the normal range of performance indicators under different operating conditions. For example, the primary frequency regulation contribution rate has a minimum and a maximum value; within this range, the unit can be considered to be performing normally. Based on historical records, the normal operating range for each performance indicator is identified. Specifically, the primary frequency regulation contribution rate range includes the unit's minimum and maximum frequency regulation contribution under frequency regulation response; the small disturbance pass rate range is the unit's performance range under small disturbance conditions; and the AGC performance range is the normal range for AGC regulation accuracy, response time, etc.

[0047] All historical unit operation records retrieved are iterated through, and the performance range of each unit is extracted. All retrieved historical operation data are analyzed to determine the range of performance indicators. Intersection filtering involves performing intersection analysis on the performance ranges of different units to ensure that the target unit's operating performance standard covers the normal operating ranges appearing in all historical data. Specifically, the operating performance range of the first unit is analyzed for intersection with the operating performance ranges of the other N-1 units, filtering out a set of ranges that include the normal performance ranges of all units. This intersection filtering process ensures that the target unit's operating performance standard covers the normal operating ranges of similar units, ensuring the standard's broad applicability. Through intersection filtering, the predetermined operating performance standard of the target unit is finally determined. This standard provides a basis for subsequent performance monitoring and evaluation, ensuring that the target unit meets the requirements of the power grid during operation and can be reasonably adjusted and optimized.

[0048] Furthermore, the step of performing intersection filtering on the operating performance ranges of the first unit to the Nth unit to obtain the predetermined unit operating performance standard includes:

[0049] When the intersection filtering result is empty, pairwise enumeration calculations are performed on the operating performance intervals of the first unit to the Nth unit to obtain the distance set of unit operating performance intervals; the unit operating performance interval distance set is traversed, and the interval outlier coefficient is evaluated to generate a set of interval outlier coefficient evaluation values; according to the set of interval outlier coefficient evaluation values, the unit operating performance intervals with interval outlier coefficient evaluation values ​​greater than or equal to the interval outlier coefficient threshold are removed to obtain a centralized set of unit operating performance intervals; based on the centralized set of unit operating performance intervals, intersection filtering is performed to obtain the predetermined unit operating performance standard.

[0050] When the performance range obtained by intersection filtering is empty, it means that the historical data of the target unit and similar units do not overlap in performance range. This situation may occur when the performance differences between the units are large or the data ranges are very inconsistent.

[0051] To address this issue, a pairwise enumeration approach is adopted. This involves performing paired analysis on the performance ranges of every two units to obtain a reasonable set of performance ranges. Pairwise enumeration refers to combining and calculating any two sets of operating performance ranges one by one to determine the distance and overlap between the unit operating performance ranges. Specifically, for each pair of units, their performance ranges are compared, and the distance between them is calculated. This distance represents the difference between the two units on the same performance indicators. Through pairwise enumeration, a distance set between the unit performance ranges is obtained. This distance set characterizes the differences in performance indicators of different units, providing a basis for subsequent outlier identification and performance range selection.

[0052] Outlier coefficient evaluation is used to assess whether there are significant deviations in the performance ranges of different generating units. The outlier coefficient determines whether a range is an outlier by comparing the differences in the performance ranges of different generating units. If the difference is significant, the range may be an outlier. Specifically, the outlier coefficient calculation formula is used to analyze the differences between the performance ranges of generating units. If the performance range of a certain generating unit differs greatly from that of other generating units and is not within the normal operating range of most generating units, then the range is considered an outlier. The outlier coefficient can be calculated based on indicators such as the standard deviation and mean distance between the performance ranges of various generating units. Usually, by comparing the deviations of the performance ranges of different generating units from the overall distribution, it is determined which ranges may be outliers.

[0053] The outlier coefficients of all calculated performance intervals are summarized, and the final set of outlier coefficient evaluation values ​​is a collection of outlier coefficient results for all performance intervals generated by the outlier coefficient evaluation. Each performance interval is assigned an outlier coefficient value, which indicates whether there is an anomaly in the interval. The outlier coefficient evaluation values ​​are usually sorted by numerical value. The higher the outlier coefficient, the less the interval conforms to the overall distribution and the more likely it is to be a potential outlier. The set of outlier coefficient evaluation values ​​is used to further optimize the performance standards to ensure that the selected intervals best represent the stable operating state of the unit.

[0054] An outlier threshold is set for each performance interval. If the outlier of an interval is greater than or equal to this threshold, the interval is considered potentially abnormal and needs to be removed. For each unit's performance interval in the outlier evaluation value set, its outlier is checked against the set threshold. If the outlier of a unit's performance interval exceeds the threshold, it indicates a significant difference in performance compared to other units, and is therefore considered abnormal and should be removed. After removing these performance intervals, the remaining set is a filtered set of centralized unit operating performance intervals. These intervals represent the typical performance of each unit under normal operating conditions, better reflecting the normal operating status of most units and providing a more reliable performance standard. By removing outlier data, the used unit performance intervals become more accurate and consistent, avoiding interference from outliers in performance standards.

[0055] Based on a centralized set of unit performance ranges, an intersection screening is performed. That is, all these performance ranges are subjected to intersection analysis to ensure that the resulting performance standards can cover the performance of all normal units. The result of the intersection screening provides a comprehensive set of performance ranges that cover the performance range of all units under normal operating conditions. The performance standards obtained in this way can be applied to different types of units simultaneously without deviation due to the special performance of certain units.

[0056] Furthermore, the method for inputting the predetermined unit operating performance standard and the unit operating performance monitoring data into the unit performance analysis model to obtain the unit performance analysis results includes:

[0057] The unit performance analysis model is constructed, comprising an input layer, a performance defect identification layer, a performance defect analysis layer, and an output layer. The predetermined unit operating performance standards and the unit operating performance monitoring data are input into the performance defect identification layer through the input layer to identify performance defects and obtain unit performance defect identification results. The unit performance defect identification results are input into the performance defect analysis layer to obtain unit performance analysis results, wherein the unit performance analysis results include a performance defect index. The unit performance analysis results are output through the output layer.

[0058] A unit performance analysis model is constructed. This multi-layered analysis system aims to process unit operating data and identify performance defects. The model consists of multiple layers, each responsible for different functions. The input layer receives data from external sources, including predetermined unit operating performance standards and actual unit operating monitoring data. Its task is to prepare the data and pass it to the next layer of the model. The performance defect identification layer analyzes the input data to identify potential performance defects, identifying anomalies or non-compliance with standards. The performance defect analysis layer further analyzes the identified defects, including diagnosing the causes and assessing the scope of impact. This layer uses sophisticated algorithms to analyze the data and find the root causes of performance problems. The output layer outputs the analysis results in an easily understandable way, including unit performance reports, defect diagnoses, and optimization suggestions. The results are typically presented numerically or visually to facilitate subsequent decision-making by users. This model can process large amounts of unit operating data, automatically identify and report performance problems, and greatly improve the efficiency and accuracy of performance monitoring.

[0059] The predetermined unit operating performance standards and unit operating performance monitoring data are input into the performance defect identification layer through the input layer. The input layer's role is to standardize or format the data to meet the processing requirements of the performance defect identification layer. In the performance defect identification layer, the differences between the predetermined performance standards and the actual monitoring data are analyzed, and potential performance defects are identified through comparison. After processing by the performance defect identification layer, the unit performance defect identification results are obtained, including defect type, such as insufficient primary frequency regulation contribution rate, slow AGC response, etc.; defect severity, that is, according to the degree of impact of the defect, a severity assessment is given, such as minor defect, severe defect, etc.; the performance range in which the defect is located, thereby clearly indicating which performance indicators of the unit have problems, such as frequency regulation, load response, etc.

[0060] The results of unit performance defect identification are passed to the performance defect analysis layer. The task of this layer is to further analyze the identified defects, diagnose the root causes of the problems, and quantify the impact of the defects. Specifically, defect types are classified and analyzed in depth. For example, it determines which factors (such as equipment failure, scheduling delay, environmental changes, etc.) may lead to a decrease in frequency regulation contribution rate. Data analysis or machine learning algorithms are used to further analyze the causes of defects, such as regression analysis, decision tree analysis, etc., to find the key factors that lead to defects. Finally, the performance defect analysis layer outputs detailed performance analysis results of the unit, including the performance defect index. This index quantifies the overall performance defect situation of the unit. It is usually calculated based on multiple performance indicators (such as frequency regulation contribution rate, AGC accuracy, etc.) and reflects the degree of deviation of the overall operating status of the unit. The performance defect index is used to measure the gap between the unit performance and the predetermined standard and provides a basis for subsequent adjustments and optimizations.

[0061] The output layer outputs the unit performance analysis results, such as presenting the results in the form of reports, including charts, data tables, etc., so that operators can view them in detail and make corresponding adjustments and optimizations. These output results provide an important basis for the maintenance and optimization of unit operation, ensuring that the unit can be kept in the best operating condition.

[0062] Furthermore, after obtaining the unit performance analysis results, the following also includes:

[0063] Read the predetermined performance defect constraint features and determine whether the performance defect index meets the predetermined performance defect constraint features; if the performance defect index does not meet the predetermined performance defect constraint features, generate a first warning level based on the degree of deviation of the performance defect index.

[0064] Predetermined performance defect constraints are a series of standards set during the design phase to measure the ideal or acceptable range of unit performance. These characteristics are based on historical data, industry standards, or grid requirements and cover the range of performance deviations the unit can tolerate during operation. For example, minimum and maximum values ​​are set for the primary frequency regulation contribution rate; when the unit's contribution rate exceeds this range, it indicates a performance defect. The performance defect index is compared with the predetermined performance defect constraints to determine whether the current unit performance meets these predetermined constraints. If the unit's performance defect index is within the constraint range, it indicates that the unit is operating normally; conversely, if it deviates from the preset constraints, it indicates that the unit has performance problems and requires further adjustment or optimization.

[0065] The degree of deviation represents the difference between the unit's performance defect index and the predetermined performance defect constraint characteristics. This deviation is obtained by calculating the difference between the value of the performance defect index and the standard value. If the unit's performance defect index significantly exceeds the predetermined constraint range, the degree of deviation is large; if the deviation from the constraint is small, the degree of deviation is small.

[0066] Warning levels are used to indicate the severity of defects. Warning levels can be divided into multiple levels, such as mild, moderate, and severe. Based on the degree of deviation from the unit's performance defect index, a first warning level is generated. For example, if the deviation is small (i.e., the performance defect index slightly exceeds the predetermined standard), a mild warning is generated, indicating that the unit's operation is still within acceptable limits but requires monitoring. If the deviation is large (i.e., the performance defect index far exceeds the predetermined standard), a higher warning level is generated, requiring immediate adjustments or optimizations. By generating a first warning level, potential problems in unit operation can be promptly identified, and maintenance personnel can be helped to determine whether more proactive intervention measures are needed to ensure the unit's operational performance returns to normal.

[0067] Furthermore, after constructing the unit performance analysis model, the process also includes:

[0068] The unit performance analysis model is evaluated to obtain the accuracy of the unit performance analysis. If the accuracy of the unit performance analysis does not reach the preset accuracy threshold, the accuracy deviation of the unit performance analysis is calculated as the model optimization deviation. The unit performance analysis model is then optimized and trained based on the model optimization deviation.

[0069] The purpose of evaluating the performance of a generator unit performance analysis model is to verify its performance on real-world data and determine its accuracy, reliability, and effectiveness. Model evaluation involves comparing the model's output with actual conditions. Specifically, a set of real generator unit data is collected, which can be performance data collected during the unit's past operation. This real-world data is input into the generator unit performance analysis model to generate analysis results. These results are then compared with the actual operating performance of the generator unit to determine the deviation between the model's output and the actual performance. During the evaluation process, standardized metrics such as accuracy, precision, and recall are used to quantify the model's performance. Accuracy is the most commonly used evaluation criterion, representing the consistency between the model's analysis results and the actual values. For example, if the type of defect predicted by the model matches the type of defect actually monitored, it is evaluated as accurate. Through evaluation, it can be determined whether the current performance analysis model is accurate enough and can correctly identify the generator unit's performance problems. If the evaluation results indicate that the model's accuracy is not high enough, the next step is the optimization process.

[0070] Accuracy deviation represents the difference between the output of the unit performance analysis model and the actual performance standard. If the performance analysis results output by the model deviate significantly from the actual monitored values, the accuracy deviation is large, and vice versa. The accuracy deviation of the model is calculated by comparing the gaps between the defect index and performance standard analyzed by the model and the actual unit performance data. Specifically, it can be calculated using methods such as absolute error and relative error. The calculated accuracy deviation is used as the model optimization deviation, providing data support for subsequent optimization training and helping to identify areas where the model needs improvement.

[0071] Based on the calculated model optimization deviation, the unit performance analysis model is optimized. For example, gradient descent is used to adjust the model parameters by calculating the error gradient, making the model output closer to the actual value. Through multiple iterations, the model is optimized until the accuracy deviation is reduced to below a preset threshold, ensuring that the model has sufficient accuracy. Optimization training also includes data augmentation, feature selection, or model structure adjustment to ensure that the model can adapt to various operating environments and unit types. Through optimization training, the accuracy of the unit performance analysis model is improved, ensuring that it can accurately reflect the unit's operating status and potential problems, thereby improving the efficiency of performance monitoring and defect early warning.

[0072] Furthermore, the method for adjusting the operating parameters of the target generator set based on the defect performance marking results includes:

[0073] Based on the defect performance marking results, the operating data of the target generator set is traced to obtain unit operating performance monitoring defect data; defect cause analysis is performed on the unit operating performance monitoring defect data to obtain unit defect cause information; using the particle swarm optimization algorithm, unit operating parameters are optimized based on the unit operating performance monitoring defect data and the unit defect cause information to obtain unit operating optimization parameters; and the operation control of the target generator set is performed according to the unit operating optimization parameters.

[0074] The defect performance labeling results are the unit performance problems identified in the aforementioned analysis steps based on the unit's performance monitoring data and predetermined operating standards. These labeling results indicate which performance indicators of the unit have defects. These labels include the type of defect (such as insufficient frequency regulation contribution rate), the severity of the defect (such as mild, moderate or severe defect), and the time or specific operating conditions under which the defect occurred.

[0075] Based on the defect marking results, the historical operating data of the unit is traced to identify the source of the defect. By tracing, it is determined in which operation or operating phases the unit deviated, and finally the unit operation performance monitoring defect data is obtained. This data reflects the specific context in which the defect occurred, the unit's performance, and possible abnormal factors.

[0076] Defect causal analysis aims to identify the root causes of unit performance defects. By deeply analyzing various data during unit operation, it identifies specific factors affecting unit performance, such as hardware failures, inappropriate control strategies, and environmental changes. Specifically, data mining and pattern recognition methods are employed. By analyzing historical data, these techniques identify potential factors related to defects. For example, under certain conditions (such as large load changes), the unit's response accuracy deteriorates, leading to performance defects. Regression analysis is used to determine the degree of influence of different factors (such as fuel quality and operating methods) on unit performance, identifying the key factors most likely to cause performance defects. Defect causal analysis generates information on the causes of unit defects, including defect type and cause, influencing factors, and potential risk assessment. It provides specific causal diagnosis, laying the foundation for subsequent optimization and adjustments.

[0077] Particle Swarm Optimization (PSO) is an optimization algorithm that simulates the foraging behavior of a swarm of particles in nature. It is used to solve complex optimization problems. The algorithm continuously updates the position of each particle in the swarm to find the global optimum. In optimizing generator operating parameters, PSO can be used to find the optimal operating parameters to improve generator performance. Specifically, multiple particles are first randomly initialized, each representing a possible solution—a combination of generator operating parameters. For each particle, its fitness is calculated. The fitness function is based on the generator's operational shortcomings and the optimization objective, such as maximizing the primary frequency regulation contribution rate or minimizing response delay. Based on the fitness of each particle and the current optimal solution, PSO updates the particle's position, i.e., the generator's operating parameters, thereby finding a better solution. Through multiple iterations, the generator operating parameters are gradually optimized until the optimal solution is found. By using PSO, the best generator operating parameters are found in a multi-dimensional space. These parameters can maximize generator performance, reduce shortcomings, and meet the needs of grid dispatching.

[0078] The optimized operating parameters are transmitted to the unit's control system, which then adjusts these parameters in real time. This includes adjusting the unit's response time and regulation accuracy to changes in grid frequency, optimizing the unit's load regulation strategy based on grid load variations to ensure smooth adaptation to load fluctuations, and adjusting the unit's automatic generation control parameters based on the optimization results to ensure accurate and rapid response to grid dispatch commands. By controlling the unit according to the optimized operating parameters, the system ensures optimal performance during actual operation, meeting the grid's frequency regulation and peak shaving requirements.

[0079] In summary, the generator performance testing method for a multi-source heterogeneous big data platform provided in this application has the following technical effects:

[0080] By utilizing a multi-source heterogeneous monitoring network, including remote terminal units, a distributed control system, and a primary frequency regulation co-source device, the operating performance of the target generator unit can be monitored in real time. This real-time monitoring can promptly capture deviations in performance indicators such as primary frequency regulation contribution rate, small disturbance qualification rate, and automatic generation control, ensuring early diagnosis and intervention to reduce potential risks. By inputting the unit's operating performance monitoring data and predetermined performance standards into the unit's performance analysis model, accurate performance analysis results are generated. This process can accurately assess the unit's current operating status, promptly identify performance defects, and generate defect performance markers and early warning signals. Information is provided to help maintenance personnel make timely adjustments; when performance defects are detected, warning levels are automatically generated based on defect performance tags, and the unit's operating status is optimized in real time by adjusting the unit's operating parameters. This optimization control improves the unit's response speed and frequency regulation accuracy, ensuring that the unit always remains in optimal working condition; through comprehensive data processing based on a big data platform, dynamic adjustment and optimization of unit performance are supported, and real-time data acquisition, statistical analysis, and multi-source data fusion ensure the system's stability and reliability. In addition, the big data platform can provide rich historical data support for future performance evaluation, laying the foundation for continuous system optimization and intelligent decision-making.

[0081] Example 2 provides a storage medium on which a computer program is stored, which, when executed by a processor, implements any step of Example 1.

[0082] Example 3, as Figure 3 The diagram shown is a structural schematic of an exemplary computer device according to this application. Figure 3 In this document, the bus architecture is represented by bus 300. Bus 300 may include any number of interconnected buses and bridges, and bus 300 connects various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, i.e., a transceiver, providing a unit for communicating with various other devices over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 can be used to store data used by processor 302 during operation.

[0083] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A generator performance testing method for a multi-source heterogeneous big data platform, characterized in that, The method includes: Construct a multi-source heterogeneous monitoring network for the target generator set, wherein the multi-source heterogeneous monitoring network includes a remote terminal unit, a distributed control system, and a primary frequency regulation device; The remote terminal unit is used to collect various operating data of the generator set in real time. The distributed control system is a core platform for control and monitoring, responsible for receiving, processing and storing data from remote terminal units. The distributed control system can perform real-time monitoring and adjust the operating status of the unit according to preset strategies. The primary frequency regulation device is used to support the frequency regulation function of the generator set to ensure the stability of the power grid frequency. The device can automatically adjust the output power of the generator set according to the fluctuation of the power grid frequency to respond to the power grid demand. Read the predetermined unit operating performance indicators, wherein the unit operating performance indicators include primary frequency regulation contribution rate, small disturbance qualification rate, and automatic generation control performance indicators; According to the predetermined unit operation performance indicators, the target generator unit is monitored for operation performance through the multi-source heterogeneous monitoring network to obtain unit operation performance monitoring data. Read the predetermined unit operating performance standard, input the predetermined unit operating performance standard and the unit operating performance monitoring data into the unit performance analysis model, and obtain the unit performance analysis results; Based on the unit performance analysis results, defective performance is marked, and performance defect early warning information is generated according to the defective performance marking results, wherein the performance defect early warning information includes a first early warning level; When the first warning level reaches the preset warning level, the operating parameters of the target generator set are adjusted according to the defect performance marking results.

2. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 1, characterized in that, The step of monitoring the operational performance of the target generator unit according to the predetermined unit operating performance indicators through the multi-source heterogeneous monitoring network to obtain unit operating performance monitoring data includes: The target generator set is collected through the multi-source heterogeneous monitoring network to obtain multi-source heterogeneous operation monitoring data. The multi-source heterogeneous operation monitoring data is preprocessed through a data cleaning domain to obtain standard operation monitoring data. The data cleaning domain includes a data denoising channel, a missing data imputation channel, an abnormal data correction channel, and a data standardization channel. Based on the predetermined unit operating performance indicators, the standard operating monitoring data is used to perform performance monitoring to obtain the unit operating performance monitoring data.

3. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 1, characterized in that, The reading of the predetermined unit operating performance standards includes: Based on the target generator set, a generator set retrieval source is set, wherein the generator set retrieval source includes the target generator set and multiple similar generator sets of the target generator set; Historical generator operation records are retrieved based on the generator set retrieval source. The first historical generator operation record is extracted based on the retrieval results, and the first generator operation performance range is identified. The first generator operation performance standard includes the primary frequency regulation contribution rate range, the small disturbance qualification rate range, and the automatic generation control performance range. The search results are iterated until the operating performance range of the Nth unit is identified. The intersection of the operating performance range of the first unit to the operating performance range of the Nth unit is filtered to obtain the predetermined operating performance standard of the unit.

4. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 3, characterized in that, The step of performing intersection filtering on the operating performance ranges of the first unit to the Nth unit to obtain the predetermined unit operating performance standard includes: When the intersection filtering result is empty, the pairwise enumeration calculation is performed from the first unit operating performance range to the Nth unit operating performance range to obtain the unit operating performance range distance set. Traverse the distance set of the unit's operating performance intervals, evaluate the interval outlier coefficient, and generate a set of interval outlier coefficient evaluation values; According to the set of interval outlier evaluation values, the unit operating performance intervals with interval outlier evaluation values ​​greater than or equal to the interval outlier threshold are removed to obtain a centralized set of unit operating performance intervals. Based on the set of centralized unit operating performance ranges, the intersection is filtered to obtain the predetermined unit operating performance standard.

5. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 1, characterized in that, The method for inputting the predetermined unit operating performance standard and the unit operating performance monitoring data into the unit performance analysis model to obtain the unit performance analysis results includes: The unit performance analysis model is constructed, wherein the unit performance analysis model includes an input layer, a performance defect identification layer, a performance defect analysis layer, and an output layer; The predetermined unit operating performance standard and the unit operating performance monitoring data are input to the performance defect identification layer through the input layer to identify performance defects and obtain the unit performance defect identification result. The unit performance defect identification results are input into the performance defect analysis layer to obtain the unit performance analysis results, wherein the unit performance analysis results include a performance defect index; The unit performance analysis results are output through the output layer.

6. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 5, characterized in that, After obtaining the unit performance analysis results, the following is also included: Read the predetermined performance defect constraint features and determine whether the performance defect index satisfies the predetermined performance defect constraint features; If the performance defect index does not meet the predetermined performance defect constraint characteristics, a first warning level is generated based on the degree of deviation of the performance defect index.

7. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 5, characterized in that, After constructing the unit performance analysis model, the following steps are also included: The performance analysis model of the unit is analyzed and its effectiveness is evaluated to obtain the accuracy of the unit performance analysis. If the accuracy of the unit performance analysis does not reach the preset accuracy threshold, the accuracy deviation of the unit performance analysis is calculated and used as the model optimization deviation. The unit performance analysis model is optimized and trained based on the model optimization deviation.

8. The generator performance testing method for a multi-source heterogeneous big data platform as described in claim 1, characterized in that, The method for adjusting the operating parameters of the target generator set based on the defect performance marking results includes: Based on the defect performance marking results, the operating data of the target generator set is traced to obtain defect data for unit operating performance monitoring. The defect data of the unit operation performance monitoring were analyzed to obtain the defect cause information of the unit; Using the particle swarm optimization algorithm, the unit operating parameters are optimized and analyzed based on the unit operation performance monitoring defect data and the unit defect cause information to obtain the unit operation optimization parameters. The target generator set is operated and controlled according to the optimized operating parameters.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the generator performance testing method of the multi-source heterogeneous big data platform according to any one of claims 1 to 8.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, The processor executes the computer program to implement the steps of the generator performance testing method of the multi-source heterogeneous big data platform according to any one of claims 1 to 8.

Images