Hydroelectric equipment intelligent monitoring system and method based on large model

Abstract

The invention discloses a hydroelectric equipment intelligent monitoring system and method based on a large model, and relates to the technical field of hydroelectric equipment intelligent monitoring, and the method comprises the steps: analyzing the sensitive degree of monitoring parameters in hydroelectric equipment to working condition changes, and obtaining characteristic hydroelectric equipment; analyzing the equipment state diagnosis accuracy of the threshold setting data on the characteristic hydroelectric equipment under the working condition to obtain a target working condition; determining a marked hydroelectric device and a marked working condition, analyzing a data influence scale of the marked hydroelectric device under the marked working condition on monitoring parameters in the characteristic hydroelectric device under the target working condition, and generating influence scale data of the characteristic hydroelectric device under the target working condition; according to the influence scale data, the abnormal monitoring model of the characteristic hydroelectric equipment is constructed, and the characteristic hydroelectric equipment is monitored, so that the state diagnosis of the hydroelectric equipment becomes accurate, the operation efficiency of the hydroelectric equipment is improved, accidental damage of the hydroelectric equipment is avoided, and safe operation of a hydropower station is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology for hydropower equipment, specifically an intelligent monitoring system and method for hydropower equipment based on a large model. Background Technology

[0002] Hydropower equipment is a general term encompassing various mechanical, electrical, and hydraulic equipment. It is mainly used in hydropower stations. Hydropower stations require hydropower equipment to work together to efficiently convert natural water energy into electrical energy and ensure its stable transmission, effective control, and adequate protection. Therefore, monitoring the status of hydropower equipment is particularly important. Traditional hydropower equipment monitoring relies heavily on periodic inspections and post-event handling. This method suffers from slow response, low efficiency, and a high dependence on human experience, resulting in many drawbacks. However, with the continuous development of technology, the use of the Internet of Things (IoT) to monitor the operating status of hydropower equipment has become increasingly common. This not only allows for real-time acquisition of operating data but also enables remote handling of anomalies via the network.

[0003] Due to their intermittent and fluctuating nature, wind and solar power generation place higher demands on the stability of the power grid. To ensure the real-time balance of the grid, flexible power sources like hydropower are needed for rapid adjustment, playing a role in "peak shaving and valley filling." Therefore, when wind and solar power generation is high, hydropower stations need to quickly reduce output or even temporarily shut down. Conversely, when wind and solar power generation is insufficient, hydropower stations need to quickly start up and increase load. Consequently, hydropower equipment in hydropower stations needs to frequently change operating conditions. However, common monitoring systems rely on stable thresholds for alarms on hydropower equipment, which is clearly unsuitable for the frequent changes in operating conditions. This can easily lead to inaccurate monitoring data of hydropower equipment, affecting not only its operating efficiency but also potentially causing damage and compromising the safety of the hydropower station. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent monitoring system and method for hydropower equipment based on a large model, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent monitoring of hydropower equipment based on a large model, the method comprising: Step S1: Obtain historical monitoring records of hydropower equipment in the hydropower station, analyze the sensitivity of monitoring parameters of hydropower equipment to changes in operating conditions, and obtain characteristic hydropower equipment; Step S2: Obtain threshold setting data for characteristic hydropower equipment from the platform, analyze the accuracy of the threshold setting data in diagnosing the equipment status of characteristic hydropower equipment under operating conditions, and obtain the target operating conditions; Step S3: Analyze the state correlation between other hydropower equipment under different operating conditions and the characteristic hydropower equipment under the target operating condition, determine the marked hydropower equipment and marked operating conditions, analyze the data influence scale of the marked hydropower equipment under the marked operating conditions on the monitoring parameters of the characteristic hydropower equipment under the target operating conditions, and generate the influence scale data of the characteristic hydropower equipment under the target operating conditions. Step S4: Based on the impact scale data, construct an anomaly monitoring model for the characteristic hydropower equipment and monitor the characteristic hydropower equipment.

[0006] Furthermore, step S3 includes: Obtain the target historical time period of the target operating condition of the characteristic hydropower equipment, and obtain the marked historical monitoring records of other hydropower equipment for the target operating condition; Obtain several operating conditions corresponding to the various markers of historical monitoring records of other hydropower equipment for the target operating condition, and obtain a set of markers of historical monitoring records for several operating conditions; Obtain the historical monitoring record set γ of the characteristic hydropower equipment under the target operating condition, and calculate the span value of the monitoring parameter in a certain historical monitoring record; Calculate the data status ratio H of the monitoring parameters in the characteristic hydroelectric equipment under the target operating condition; Obtain the marked operating conditions of the marked hydroelectric equipment; Acquire the characteristic hydropower equipment and mark the target historical records ρ´ and σ´ of the hydropower equipment for a certain target historical period; Calculate the data influence values ​​of the d-th monitoring parameter in the characteristic hydropower equipment under the target operating condition and the k-th monitoring parameter in the marked hydropower equipment under the marked operating condition in each target historical period, and aggregate them to obtain the data influence value set Q; Obtain the variance δ of the data impact value set Q, and set the variance threshold δ. ◇ When δ>δ ◇ If δ ≤ δ, then it is determined that the k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition has no effect on the data change of the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition. ◇ At that time, obtain the average value Q of the data influence value of each element in the data influence value set Q. μ ; Set the data influence threshold q´, when Q μ If Q > q´, then it is determined that the k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition has an impact on the data change of the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition. The k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition is denoted as the monitoring parameter affecting the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition, and the d-th monitoring parameter is marked. μ When ≤q´, the kth monitoring parameter in the marked hydroelectric equipment under the marked working condition will not be processed; Several influencing monitoring parameters that affect the monitoring parameters of the characteristic hydropower equipment under the marked working condition are obtained and collected to generate the influence scale data of the characteristic hydropower equipment under the target working condition. By generating impact scale data of characteristic hydropower equipment under target operating conditions in the above steps, the hidden impact of characteristic hydropower equipment under different operating conditions can be accurately captured. This solves the problem that traditional threshold range judgment cannot be adjusted according to the changes in the operating conditions of hydropower equipment itself, and also greatly improves the accuracy of subsequent judgment of hydropower equipment anomalies.

[0007] Furthermore, step S2 includes: Obtain the threshold ranges of various monitoring parameters preset by the platform for the characteristic hydropower equipment from the threshold setting data; Obtain the average value μ of the maximum and minimum values ​​of the monitored parameters in the characteristic hydropower equipment within the threshold range. Extract the total number B of historical monitoring records where the monitored parameters are greater than the average value μ from the historical monitoring record set of the characteristic hydropower equipment. Also obtain the total number B of historical monitoring records in the historical monitoring record set of the operating conditions. sum Calculate the characteristic distribution trend β=B / B of the monitoring parameters of the characteristic hydropower equipment under the operating conditions with respect to the threshold setting data. sum ; Obtain the maximum value β of the characteristic distribution pattern of various monitoring parameters of characteristic hydroelectric equipment under operating conditions with respect to the threshold setting data. max and minimum value β min Set the characteristic distribution range, when β max and β min When all values ​​are within the range of the characteristic distribution pattern, the accuracy of the threshold setting data in diagnosing various monitoring parameters of characteristic hydropower equipment under different working conditions is determined. When β max or β min If the data is outside the characteristic distribution range, then calculate the average value μ of the characteristic distribution of the monitoring parameters of the characteristic hydropower equipment under various operating conditions relative to the threshold setting data. β Calculate the overall distribution difference P = |μ| of the monitoring parameters of the characteristic hydropower equipment under operating conditions with respect to the threshold setting data. β -β| / max(β,μ β ), calculate the maximum value P of the overall distribution difference of various monitoring parameters of characteristic hydroelectric equipment with respect to threshold setting data under operating conditions. max ; Set an overall distribution difference threshold P', when P' max When the threshold setting data is >P´, if the determination of the threshold setting data is inaccurate for the equipment status diagnosis of the characteristic hydropower equipment under the working condition, the working condition will be recorded as the target working condition of the characteristic hydropower equipment; otherwise, the working condition will not be processed.

[0008] Further, step S1 includes: Monitor the equipment status of the hydropower equipment under a certain working condition, obtain each historical monitoring record of the hydropower equipment under a certain working condition, and collect them to obtain the historical monitoring record set of a certain working condition; Obtain the historical monitoring record sets of various working conditions of the hydropower equipment, accumulate the total number of historical monitoring records in the historical monitoring record sets of various working conditions of the hydropower equipment to obtain the total number of working condition changes M of the hydropower equipment, obtain the duration T between the time point when the hydropower equipment is monitored and the current cycle, and calculate the working condition change frequency A = M / T; Set the change frequency threshold a. When A < a, it is determined that the working condition change of the hydropower equipment is not frequent. Otherwise, obtain the maximum value α max and the minimum value α min of the mean values of the monitoring parameters in the hydropower equipment in the historical monitoring record sets of various working conditions, and calculate the sensitivity value C of the monitoring parameters in the hydropower equipment to the working condition change = (α max -α min ) / α min . When there are monitoring parameters with sensitivity values to working condition changes greater than the preset sensitivity threshold among the various monitoring parameters in the hydropower equipment, determine the sensitivity degree of the monitoring parameters in the hydropower equipment to the working condition change, and record the hydropower equipment as a characteristic hydropower equipment. Otherwise, do not process the hydropower equipment.

[0009] Further, step S4 includes: Obtain the influence scale data of the characteristic hydropower equipment under the target working condition, and obtain the influence monitoring parameters of the marked hydropower equipment under the marked working condition that have an impact on the marked monitoring parameters in the characteristic hydropower equipment under the target working condition from the influence degree data; Respectively obtain the target historical records of the characteristic hydropower equipment and the marked hydropower equipment for each target historical period under the target working condition, and use the data set of the influence monitoring parameters in the target historical records of the marked hydropower equipment under the marked working condition that have an impact on the marked monitoring parameters in the characteristic hydropower equipment under the target working condition, and the data set of the marked monitoring parameters in the target historical records of the characteristic hydropower equipment under the target working condition as the input of the model within the target historical period; Set the characteristic duration and the abnormal variable W. When the characteristic hydropower equipment does not have an abnormality within the characteristic duration after the target historical period, the abnormal variable W of the characteristic hydropower equipment within the target historical period = 0. Otherwise, the abnormal variable W of the characteristic hydropower equipment within the target historical period = 1; Use the fault variable W of the characteristic hydropower equipment within the target historical period as the output, divide the target historical period into a training set and a test set according to a ratio, and use the large model technology to construct an abnormal monitoring model for the marked monitoring parameters in the characteristic hydropower equipment; During the current cycle, an anomaly monitoring model is used to monitor the marked monitoring parameters of characteristic hydropower equipment for anomalies. Threshold setting data is used to monitor the hydropower equipment in the hydropower station for anomalies, and the hydropower equipment in the hydropower station that is determined to be abnormal is obtained and reported through the platform.

[0010] To better implement the above methods, a smart monitoring system for hydropower equipment based on a large model is also proposed. The system includes a sensitivity analysis module for changes in operating conditions, a diagnostic accuracy analysis module, an impact scale analysis module, and an equipment monitoring module. The operating condition change sensitivity analysis module is used to analyze the sensitivity of monitoring parameters in hydropower equipment to changes in operating conditions, and to obtain characteristic hydropower equipment. The diagnostic accuracy analysis module is used to analyze the accuracy of threshold setting data in diagnosing the equipment status of characteristic hydropower equipment under operating conditions, and to obtain the target operating conditions. The impact scale analysis module is used to determine the marked hydropower equipment and marked operating conditions, analyze the impact scale of the marked hydropower equipment on the monitoring parameters of the characteristic hydropower equipment under the marked operating conditions, and generate impact scale data. The equipment monitoring module is used to construct anomaly monitoring models for characteristic hydropower equipment based on impact scale data, and to monitor the characteristic hydropower equipment.

[0011] Furthermore, the operating condition change sensitivity analysis module includes a record set acquisition unit and an operating condition change sensitivity analysis unit; The record set acquisition unit is used to monitor the status of hydropower equipment under operating conditions, acquire and collect various historical monitoring records of hydropower equipment under operating conditions, and obtain a historical monitoring record set of operating conditions. The operating condition change sensitivity analysis unit is used to calculate the sensitivity value of the monitoring parameters in the hydropower equipment to changes in operating conditions based on the historical monitoring record set of operating conditions, determine the sensitivity of the monitoring parameters in the hydropower equipment to changes in operating conditions, and obtain the characteristic hydropower equipment.

[0012] Furthermore, the diagnostic accuracy analysis module includes a distribution difference calculation unit and a diagnostic accuracy analysis unit; The distribution difference calculation unit is used to calculate the overall distribution difference of various monitoring parameters in characteristic hydropower equipment under operating conditions with respect to threshold setting data; The diagnostic accuracy analysis unit is used to analyze the accuracy of the diagnostic of the equipment status of characteristic hydropower equipment under the operating conditions based on the overall distribution differences and threshold setting data, so as to obtain the target operating conditions.

[0013] Furthermore, the influence scale analysis module includes a state correlation analysis unit and an influence scale analysis unit; The state correlation analysis unit is used to analyze the state correlation between other hydropower equipment under different operating conditions and the characteristic hydropower equipment under the target operating condition, and to determine the marked hydropower equipment and marked operating conditions. The influence scale analysis unit is used to analyze the influence scale of the marked hydropower equipment on the monitoring parameters of the characteristic hydropower equipment under the marked operating condition, and to generate the influence scale data of the characteristic hydropower equipment under the target operating condition.

[0014] Furthermore, the equipment monitoring module includes a model building unit and an equipment monitoring unit; The model building unit is used to construct anomaly monitoring models for the marked monitoring parameters of characteristic hydropower equipment under the target operating conditions using large model technology. The equipment monitoring unit uses an anomaly monitoring model and threshold setting data to monitor the hydropower equipment in the hydropower station for anomalies, identifies the hydropower equipment in the hydropower station that is determined to be abnormal, and reports it through the platform.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention realizes intelligent monitoring of hydropower equipment. Because hydropower equipment in hydropower stations needs to frequently change operating conditions, it first analyzes the sensitivity of monitoring parameters in hydropower equipment to changes in operating conditions, and combines threshold setting data to identify that there are preset thresholds that cannot accurately identify the operating conditions of hydropower equipment. Furthermore, by analyzing the scale of influence of monitoring parameters of other hydropower equipment under different operating conditions on the data of characteristic hydropower equipment under the target operating condition, it obtains the implicit influence relationship on the equipment status of hydropower equipment under the target operating condition. The anomaly monitoring model constructed through large model technology not only makes the diagnosis of hydropower equipment status more accurate, but also improves the operating efficiency of hydropower equipment, avoids accidental damage to hydropower equipment, and ensures the safe operation of hydropower stations. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the impact scale of an intelligent monitoring method for hydropower equipment based on a large model, as described in this invention. Figure 2 This is a schematic diagram of a module of an intelligent monitoring system for hydropower equipment based on a large model according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example: Figures 1-2 As shown, the present invention provides a technical solution, an intelligent monitoring method for hydropower equipment based on a large model, and the method includes: Step S1: Obtain the historical monitoring records of hydropower equipment in a hydropower station, analyze the sensitivity of the monitoring parameters in the hydropower equipment to the change of working conditions, and obtain the characteristic hydropower equipment; Among them, step S1 includes: Monitor the equipment state of the hydropower equipment under a certain working condition, obtain each historical monitoring record of the hydropower equipment under a certain working condition, and collect them to obtain the historical monitoring record set of a certain working condition; For example, the working conditions of hydropower equipment are different according to different hydropower equipment. When the hydropower equipment is a water turbine, the working conditions of the water turbine are stable power generation condition, vortex band condition area, pumping condition, etc.; Obtain the historical monitoring record sets of various working conditions of the hydropower equipment, accumulate the total number of historical monitoring records in the historical monitoring record sets of various working conditions of the hydropower equipment to obtain the total number of working condition changes M of the hydropower equipment, obtain the duration T between the time point when the hydropower equipment is monitored and the current cycle, and calculate the working condition change frequency A = M / T; Set a change frequency threshold a. When A < a, it is determined that the working condition change of the hydropower equipment is not frequent. Otherwise, obtain the maximum value α of the mean value of the monitoring parameters in the hydropower equipment in the historical monitoring record sets of various working conditions max and the minimum value α min , calculate the sensitivity value C of the monitoring parameters in the hydropower equipment to the working condition change = (α max - α min ) / α min . When there are monitoring parameters in the monitoring parameters of the hydropower equipment whose sensitivity values to the working condition change are greater than the preset sensitivity threshold, determine the sensitivity of the monitoring parameters in the hydropower equipment to the working condition change, and record the hydropower equipment as the characteristic hydropower equipment. Otherwise, do not process the hydropower equipment; For example, when the hydropower equipment is a power generation equipment, the monitoring parameters include current, voltage, stator temperature, etc.

[0019] Step S2: Obtain the threshold setting data of the characteristic hydropower equipment from the platform, analyze the diagnostic accuracy of the equipment state of the characteristic hydropower equipment under the working condition, and obtain the target working condition; Among them, step S2 includes: Obtain the threshold ranges of various monitoring parameters preset by the platform for the characteristic hydropower equipment from the threshold setting data; Obtain the average value μ of the maximum and minimum values ​​of the monitored parameters in the characteristic hydropower equipment within the threshold range. Extract the total number B of historical monitoring records where the monitored parameters are greater than the average value μ from the historical monitoring record set of the characteristic hydropower equipment. Also obtain the total number B of historical monitoring records in the historical monitoring record set of the operating conditions. sum Calculate the characteristic distribution trend β=B / B of the monitoring parameters of the characteristic hydropower equipment under the operating conditions with respect to the threshold setting data. sum ; Obtain the maximum value β of the characteristic distribution pattern of various monitoring parameters of characteristic hydroelectric equipment under operating conditions with respect to the threshold setting data. max and minimum value β min Set the characteristic distribution range, when β max and β min When all values ​​are within the range of the characteristic distribution pattern, the accuracy of the threshold setting data in diagnosing various monitoring parameters of characteristic hydropower equipment under different working conditions is determined. When β max or β min If the data is outside the characteristic distribution range, then calculate the average value μ of the characteristic distribution of the monitoring parameters of the characteristic hydropower equipment under various operating conditions relative to the threshold setting data. β Calculate the overall distribution difference P = |μ| of the monitoring parameters of the characteristic hydropower equipment under operating conditions with respect to the threshold setting data. β -β| / max(β,μ β ), calculate the maximum value P of the overall distribution difference of various monitoring parameters of characteristic hydroelectric equipment with respect to threshold setting data under operating conditions. max ; Set an overall distribution difference threshold P', when P' max When the threshold setting data is >P´, if the determination of the threshold setting data is inaccurate for the equipment status diagnosis of the characteristic hydropower equipment under the working condition, the working condition will be recorded as the target working condition of the characteristic hydropower equipment; otherwise, the working condition will not be processed.

[0020] Step S3: Analyze the state correlation between other hydropower equipment under different operating conditions and the characteristic hydropower equipment under the target operating condition, determine the marked hydropower equipment and marked operating conditions, analyze the data influence scale of the marked hydropower equipment under the marked operating conditions on the monitoring parameters of the characteristic hydropower equipment under the target operating conditions, and generate the influence scale data of the characteristic hydropower equipment under the target operating conditions. Step S3 includes: Obtain the target historical time period of the target operating condition of the characteristic hydropower equipment, and obtain the marked historical monitoring records of other hydropower equipment for the target operating condition; For example, the specific process for obtaining the target historical time period of the target operating conditions of characteristic hydropower equipment is as follows: Obtain historical monitoring records of other hydropower equipment under various operating conditions from the platform, obtain the historical time periods of the characteristic hydropower equipment under the target operating conditions, and record them as the target historical time periods of the target operating conditions. For example, obtaining historical monitoring records of other hydropower equipment for the target operating condition involves the following process: Collect historical monitoring records related to the target historical period from the historical monitoring records of other hydropower equipment under various operating conditions, and record them as the marked historical monitoring records of other hydropower equipment for the target operating conditions; For example, the specific method for determining historical monitoring records related to the target historical period is as follows; When the historical monitoring records of other hydropower equipment under various operating conditions are within the target historical period, it is determined that the historical monitoring records are related to the target historical period. When the historical monitoring records of other hydropower equipment in the historical monitoring record set of various operating conditions contain the target historical period, it is determined that the historical monitoring record is related to the target historical period. Obtain several operating conditions corresponding to the various markers of historical monitoring records of other hydropower equipment for the target operating condition, and obtain a set of markers of historical monitoring records for several operating conditions; For example, to obtain a set of marked historical monitoring records for several operating conditions, the specific acquisition process is as follows: Obtain historical monitoring records of various markers for the target operating condition from other hydropower equipment, obtain several operating conditions corresponding to each historical monitoring record of markers, and collect the historical monitoring records of several operating conditions according to the operating conditions of other hydropower equipment corresponding to the historical monitoring records of markers, to obtain a set of historical monitoring records of several operating conditions. Obtain the historical monitoring record set γ of the characteristic hydropower equipment under the target operating condition, and calculate the span value of the monitoring parameter in a certain historical monitoring record; For example, to calculate the span value of a monitoring parameter in a historical monitoring record, the specific calculation process is as follows: The difference between the maximum and minimum values ​​of a monitoring parameter in a specific historical monitoring record within the historical monitoring record set γ is used to obtain the span value of the monitoring parameter in that specific historical monitoring record. Calculate the data status ratio H of the monitoring parameters in the characteristic hydroelectric equipment under the target operating condition; For example, the specific calculation process for the data status ratio value H is as follows: Obtain the historical monitoring record set γ of the characteristic hydropower equipment under the target operating condition, obtain the difference between the maximum and minimum values ​​of the monitoring parameter in a certain historical monitoring record within the historical monitoring record set γ, and obtain the span value of the monitoring parameter in a certain historical monitoring record. Set state variable S, obtain the mean value F of the span value of the monitoring parameter in the historical monitoring record set γ, obtain the mean value F´ of the span value of the monitoring parameter of the characteristic hydropower equipment in the historical monitoring record set under various working conditions. When F>F´, the state variable S=1 of the monitoring parameter of the characteristic hydropower equipment in the historical monitoring record set γ; otherwise, the state variable S=0. Obtain the state variables of the monitoring parameters of the characteristic hydropower equipment in the historical monitoring record set γ and accumulate them to obtain the state characteristic value S of the monitoring parameters of the characteristic hydropower equipment under the target operating condition. sum Calculate the data state ratio H=S of the monitoring parameters in the characteristic hydropower equipment under the target operating condition. sum / γ sum , where γ sum The total number of historical monitoring records within the historical monitoring record set γ; For example, calculating the data status ratio H of various monitoring parameters of other hydropower equipment under the ε-th operating condition in several operating conditions. ε The specific calculation process is as follows: The state variables of the monitoring parameters of other hydropower equipment in the historical monitoring records of the marked historical monitoring record set for the ε-th operating condition are summed to obtain the state characteristic value S´ of the monitoring parameters of other hydropower equipment under the ε-th operating condition. The data state ratio value H of the monitoring parameters of the characteristic hydropower equipment under the target operating condition is then calculated. ε =S´ / ε sum , where ε sum The total number of historical monitoring records in the historical monitoring record set for other hydropower equipment under the ε-th operating condition; Obtain the marked operating conditions of the marked hydroelectric equipment; For example, the specific process for obtaining the marked operating conditions of marked hydroelectric equipment is as follows: Obtain the data state ratio values ​​of various monitoring parameters in the characteristic hydropower equipment under the target operating condition. Calculate the data state ratio values ​​of various monitoring parameters in other hydropower equipment under several operating conditions. Calculate the minimum value G of the absolute difference in data state between various monitoring parameters in the characteristic hydropower equipment under the target operating condition and various monitoring parameters in other hydropower equipment under several operating conditions. min ; For example, the absolute difference in data status between various monitoring parameters of a characteristic hydropower equipment under a target operating condition and various monitoring parameters of other hydropower equipment under several operating conditions is calculated as follows: Calculate the absolute value of the difference between the data status ratio of each monitoring parameter in the characteristic hydropower equipment under the target operating condition and the data status ratio of each monitoring parameter in other hydropower equipment under several operating conditions, and obtain the absolute difference of the data status between each monitoring parameter in the characteristic hydropower equipment under the target operating condition and the monitoring parameters of each monitoring parameter in other hydropower equipment under several operating conditions; Set an absolute threshold G', when G... min When >G´, it is determined that there is no state association between other hydropower equipment under several working conditions and the characteristic hydropower equipment under the target working condition; When G min When ≤G´, it is determined that there is a state association between other hydropower equipment under several working conditions and the characteristic hydropower equipment under the target working condition. Other hydropower equipment is recorded as the marked hydropower equipment of the characteristic hydropower equipment. The working conditions in which the absolute difference of the data state is less than the absolute threshold G´ are obtained from several working conditions and recorded as the marked working conditions of the marked hydropower equipment. Acquire the characteristic hydropower equipment and mark the target historical records ρ´ and σ´ of the hydropower equipment for a certain target historical period; For example, the specific process of obtaining the target historical records ρ´ and σ´ is as follows: Obtain the marked historical monitoring record set φ of the marked operating conditions of the marked hydropower equipment, obtain the historical monitoring record ρ of the d-th monitoring parameter of the characteristic hydropower equipment under the target operating condition within a certain target historical period, obtain the marked historical monitoring record σ that is related to a certain target historical period from the marked historical monitoring record set φ, extract the records of historical monitoring record ρ and marked historical monitoring record σ that are in the same historical period, and obtain the target historical records ρ´ and σ´ of the characteristic hydropower equipment and the marked hydropower equipment for a certain target historical period, respectively. Calculate the data influence values ​​of the d-th monitoring parameter in the characteristic hydropower equipment under the target operating condition and the k-th monitoring parameter in the marked hydropower equipment under the marked operating condition in each target historical period, and aggregate them to obtain the data influence value set Q; For example, calculating the data influence value S of the d-th monitoring parameter in the characteristic hydropower equipment under the target operating condition and the k-th monitoring parameter in the marked hydropower equipment under the marked operating condition within a certain target historical period. (d,k) The specific calculation process is as follows: Set the unit duration and calculate the data impact value S. (d,k) : , Where, x i Let d be the average value of the d-th monitoring parameter within the i-th unit of time in the target historical record ρ'; x' is the average value of the d-th monitoring parameter within each unit of time in the target historical record ρ'; y iy' is the average value of the k-th monitoring parameter within the i-th unit of time in the target historical record σ'; y' is the average value of the k-th monitoring parameter within each unit of time in the target historical record σ'; Obtain the variance δ of the data impact value set Q, and set the variance threshold δ. ◇ When δ>δ ◇ If δ ≤ δ, then it is determined that the k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition has no effect on the data change of the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition. ◇ At that time, obtain the average value Q of the data influence value of each element in the data influence value set Q. μ ; Set the data influence threshold q´, when Q μ If Q > q´, then it is determined that the k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition has an impact on the data change of the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition. The k-th monitoring parameter in the marked hydroelectric equipment under the marked operating condition is denoted as the monitoring parameter affecting the d-th monitoring parameter in the characteristic hydroelectric equipment under the target operating condition, and the d-th monitoring parameter is marked. μ When ≤q´, the kth monitoring parameter in the marked hydroelectric equipment under the marked working condition will not be processed; Several influencing monitoring parameters that affect the monitoring parameters of the characteristic hydropower equipment under the marked operating condition are obtained and collected to generate the influence scale data of the characteristic hydropower equipment under the target operating condition.

[0021] Step S4: Based on the impact scale data, construct an anomaly monitoring model for the characteristic hydropower equipment and monitor the characteristic hydropower equipment.

[0022] Step S4 includes: Obtain the impact scale data of characteristic hydropower equipment under the target operating condition, and extract the impact monitoring parameters of the marked hydropower equipment under the marked operating condition that have an impact on the marked monitoring parameters of the characteristic hydropower equipment under the target operating condition from the impact degree data; The target historical records of the feature hydropower equipment and the labeled hydropower equipment for each target historical period under the target operating condition are obtained respectively. The dataset of the influence monitoring parameters that affect the labeled monitoring parameters in the feature hydropower equipment under the target operating condition from the target historical records of the labeled hydropower equipment under the labeled operating condition, and the dataset of the labeled monitoring parameters in the target historical records of the feature hydropower equipment under the target operating condition are used as the input of the model within the target historical period. Set the characteristic duration and the abnormal variable W. When the characteristic hydropower equipment does not experience any abnormality within the characteristic duration after the target historical period, the abnormal variable W of the characteristic hydropower equipment in the target historical period is 0. Otherwise, the abnormal variable W of the characteristic hydropower equipment in the target historical period is 1. The fault variable W of the characteristic hydropower equipment in the target historical period is used as the output. The target historical period is divided into training set and test set according to the proportion and degree. The large model technology is used to construct the anomaly monitoring model of the marked monitoring parameters in the characteristic hydropower equipment. For example, using large model techniques, anomaly monitoring models for labeled monitoring parameters in characteristic hydropower equipment are constructed, specifically as follows: Using large model techniques, such as Timer-XL, the anomaly detection model is trained using the inputs and outputs of the target historical time period in the training set, and then trained using the test set. The ratio of correctly predicted samples to the total number of samples is calculated to obtain the accuracy of the anomaly detection model. When the accuracy is greater than a preset threshold, the anomaly detection model is considered to have been successfully built. Otherwise, the model parameters in the anomaly detection model need to be adjusted until the accuracy is greater than the preset threshold. During the current cycle, an anomaly monitoring model is used to monitor the marked monitoring parameters of the characteristic hydropower equipment, and threshold setting data is used to monitor the hydropower equipment in the hydropower station for anomalies. The hydropower equipment in the hydropower station that is judged to be abnormal is obtained and reported through the platform. For example, the specific process for obtaining information on hydroelectric equipment identified as abnormal in a hydropower station is as follows: For example, when the anomaly monitoring model performs anomaly monitoring on the marked monitoring parameters of the characteristic hydropower equipment and outputs anomaly variable W=1, it is determined that the characteristic hydropower equipment is abnormal in the current period. If the value of an unmarked monitoring parameter in a characteristic hydroelectric device or a monitoring parameter in a hydroelectric device is outside the threshold range set in the threshold data, the characteristic hydroelectric device or the hydroelectric device is determined to be abnormal.

[0023] To better implement the above methods, a smart monitoring system for hydropower equipment based on a large model is also proposed. The system includes a sensitivity analysis module for changes in operating conditions, a diagnostic accuracy analysis module, an impact scale analysis module, and an equipment monitoring module. The operating condition change sensitivity analysis module is used to analyze the sensitivity of monitoring parameters in hydropower equipment to changes in operating conditions, and to obtain characteristic hydropower equipment. The diagnostic accuracy analysis module is used to analyze the accuracy of threshold setting data in diagnosing the equipment status of characteristic hydropower equipment under operating conditions, and to obtain the target operating conditions. The impact scale analysis module is used to determine the marked hydropower equipment and marked operating conditions, analyze the impact scale of the marked hydropower equipment on the monitoring parameters of the characteristic hydropower equipment under the marked operating conditions, and generate impact scale data. The equipment monitoring module is used to construct anomaly monitoring models for characteristic hydropower equipment based on impact scale data, and to monitor the characteristic hydropower equipment.

[0024] The operating condition change sensitivity analysis module includes a record set acquisition unit and an operating condition change sensitivity analysis unit. The record set acquisition unit is used to monitor the status of hydropower equipment under operating conditions, acquire and collect various historical monitoring records of hydropower equipment under operating conditions, and obtain a historical monitoring record set of operating conditions. The operating condition change sensitivity analysis unit is used to calculate the sensitivity value of the monitoring parameters in the hydropower equipment to changes in operating conditions based on the historical monitoring record set of operating conditions, determine the sensitivity of the monitoring parameters in the hydropower equipment to changes in operating conditions, and obtain the characteristic hydropower equipment.

[0025] The diagnostic accuracy analysis module includes a distribution difference calculation unit and a diagnostic accuracy analysis unit. The distribution difference calculation unit is used to calculate the overall distribution difference of various monitoring parameters in characteristic hydropower equipment under operating conditions with respect to threshold setting data; The diagnostic accuracy analysis unit is used to analyze the accuracy of the diagnostic of the equipment status of characteristic hydropower equipment under the operating conditions based on the overall distribution differences and threshold setting data, so as to obtain the target operating conditions.

[0026] The influence scale analysis module includes a state correlation analysis unit and an influence scale analysis unit. The state correlation analysis unit is used to analyze the state correlation between other hydropower equipment under different operating conditions and the characteristic hydropower equipment under the target operating condition, and to determine the marked hydropower equipment and marked operating conditions. The influence scale analysis unit is used to analyze the influence scale of the marked hydropower equipment on the monitoring parameters of the characteristic hydropower equipment under the marked operating condition, and to generate the influence scale data of the characteristic hydropower equipment under the target operating condition.

[0027] The equipment monitoring module includes a model building unit and an equipment monitoring unit. The model building unit is used to construct anomaly monitoring models for the marked monitoring parameters of characteristic hydropower equipment under the target operating conditions using large model technology. The equipment monitoring unit uses an anomaly monitoring model and threshold setting data to monitor the hydropower equipment in the hydropower station for anomalies, identifies the hydropower equipment in the hydropower station that is determined to be abnormal, and reports it through the platform.

[0028] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A large model-based intelligent monitoring method for hydroelectric equipment, characterized in that, The method comprises: Step S1: obtaining historical monitoring records of hydropower equipment in a hydropower station, analyzing the sensitivity of monitoring parameters in the hydropower equipment to working condition changes, and obtaining characteristic hydropower equipment; Step S2: obtaining threshold setting data of the characteristic hydropower equipment from a platform, analyzing the accuracy of device state diagnosis of the characteristic hydropower equipment under the threshold setting data, and obtaining a target working condition; Step S3: analyzing the state correlation between other hydropower equipment under different working conditions and the characteristic hydropower equipment under the target working condition, determining a marker hydropower equipment and a marker working condition, analyzing the data influence scale of the marker hydropower equipment on the monitoring parameters in the characteristic hydropower equipment under the target working condition, and generating influence scale data of the characteristic hydropower equipment under the target working condition; Step S4: constructing an abnormal monitoring model of the characteristic hydropower equipment according to the influence scale data, and monitoring the characteristic hydropower equipment.

2. The large model-based intelligent monitoring method for hydroelectric equipment according to claim 1, characterized in that, The step S3 comprises: obtaining a target historical period of the target working condition of the characteristic hydropower equipment, and obtaining marker historical monitoring records of other hydropower equipment for the target working condition; obtaining a plurality of working conditions corresponding to each marker historical monitoring record of other hydropower equipment for the target working condition, and obtaining a marker historical monitoring record set of the plurality of working conditions; obtaining a historical monitoring record set γ of the characteristic hydropower equipment under the target working condition, and calculating a span value of a monitoring parameter in a certain historical monitoring record; calculating a data state proportion value H of the monitoring parameter in the characteristic hydropower equipment under the target working condition; obtaining a marker working condition of the marker hydropower equipment; obtaining target historical records ρ´ and σ´ of the characteristic hydropower equipment and the marker hydropower equipment for a certain target historical period; calculating data influence values of the dth monitoring parameter in the characteristic hydropower equipment under the target working condition and the kth monitoring parameter in the marker hydropower equipment under the marker working condition in each target historical period, and collecting the data influence values to obtain a data influence value set Q; Obtain variance δ in data influence value set Q, set variance threshold δ ◇ When δ> δ ◇ , then determine that the data change of the kth monitoring parameter in the marked hydroelectric equipment to the dth monitoring parameter in the characteristic hydroelectric equipment under the target working condition does not have influence under the marked working condition, when δ≤ δ ◇ , obtain the average value Q μ of the data influence value of each element in the data influence value set Q. A data influence threshold q' is set, when Q μ > q', it is determined that the kth monitoring parameter in the marked hydroelectric equipment under the marked working condition has influence on the data change of the dth monitoring parameter in the characteristic hydroelectric equipment under the target working condition, the kth monitoring parameter in the marked hydroelectric equipment under the marked working condition is marked as the influence monitoring parameter of the dth monitoring parameter in the characteristic hydroelectric equipment under the target working condition, and the dth monitoring parameter is marked, when Q μ ≤ q', the kth monitoring parameter in the marked hydroelectric equipment under the marked working condition is not processed. obtaining a plurality of influence monitoring parameters of the marker hydropower equipment under the marker working condition that have an influence on the monitoring parameters in the characteristic hydropower equipment under the target working condition, and collecting the influence monitoring parameters to generate the influence scale data of the characteristic hydropower equipment under the target working condition.

3. The large model-based intelligent monitoring method for hydroelectric equipment according to claim 1, characterized in that, The step S2 comprises: obtaining a threshold range of each monitoring parameter preset by the platform for the characteristic hydropower equipment from the threshold setting data; Obtaining the average value μ of the maximum value and the minimum value of the monitoring parameter in the threshold range in the characteristic hydropower equipment, obtaining the total number B of historical monitoring records of the monitoring parameter greater than the average value μ from the historical monitoring record set of the working condition in the characteristic hydropower equipment, and obtaining the total number B of historical monitoring records in the historical monitoring record set of the working condition sum , calculating the characteristic distribution situation β=B / B sum of the monitoring parameter in the characteristic hydropower equipment to the threshold setting data under the working condition. obtaining a maximum value β of a characteristic distribution trend of threshold setting data of each monitoring parameter in the characteristic hydroelectric equipment under the working condition max and a minimum value β min , setting a characteristic distribution trend range, when β max and β min are both in the characteristic distribution trend range, then determining the diagnostic accuracy of the threshold setting data to each monitoring parameter in the characteristic hydroelectric equipment under different working conditions; When β max or β min If the data is outside the characteristic distribution range, then calculate the average value μ of the characteristic distribution of the monitoring parameters of the characteristic hydropower equipment under various operating conditions relative to the threshold setting data. β Calculate the overall distribution difference P = |μ| of the monitoring parameters of the characteristic hydroelectric equipment under the stated operating condition with respect to the threshold setting data. β -β| / max(β,μ β ), calculate the maximum value P of the overall distribution difference of each monitoring parameter of the characteristic hydroelectric equipment with respect to the threshold setting data under the said operating condition. max ; A whole distribution difference threshold P' is set, and when P max If P' > P, it is determined that the threshold setting data is inaccurate for diagnosing the equipment state of the characteristic water and power equipment under the working condition, and the working condition is recorded as a target working condition of the characteristic water and power equipment, otherwise, the working condition is not processed.

4. The large model-based intelligent monitoring method for hydroelectric equipment according to claim 1, characterized in that, The step S1 comprises: monitoring the device state of the hydropower equipment under a certain working condition, obtaining each historical monitoring record of the hydropower equipment under the certain working condition, and collecting the historical monitoring records to obtain a historical monitoring record set of the certain working condition; obtaining historical monitoring record sets of each working condition of the hydropower equipment, accumulating the total number of historical monitoring records in the historical monitoring record sets of each working condition of the hydropower equipment, obtaining a total number M of working condition changes of the hydropower equipment, obtaining a time length T between a time point when the hydropower equipment is monitored and a current period, and calculating a working condition change frequency A of the hydropower equipment M / T; A change frequency threshold value a is set, when A < a, it is determined that the working condition of the hydroelectric equipment is not changed frequently, otherwise, the maximum value α of the average of the monitoring parameters in the hydroelectric equipment in the historical monitoring record set of each working condition is obtained max and the minimum value α min , the sensitivity value C of the monitoring parameters in the hydroelectric equipment to the working condition change is calculated=(α max -α min ) / α min When there is a monitoring parameter in each monitoring parameter in the hydroelectric equipment whose sensitivity value to the working condition change is greater than a preset sensitivity threshold value, it is determined that the sensitivity degree of the monitoring parameters in the hydroelectric equipment to the working condition change, and the hydroelectric equipment is recorded as a characteristic hydroelectric equipment, otherwise, the hydroelectric equipment is not processed.

5. The large model-based intelligent monitoring method for hydroelectric equipment according to claim 1, characterized in that, The step S4 comprises: obtaining the influence scale data of the characteristic hydropower equipment under the target working condition, and obtaining influence monitoring parameters of the marker hydropower equipment under the marker working condition that have an influence on the marked monitoring parameters in the characteristic hydropower equipment under the target working condition from the influence degree data; Respectively acquire the target historical records of the feature hydropower equipment and the labeled hydropower equipment for each target historical period under the target working condition, acquire the data set of the influence monitoring parameter in the target historical record of the labeled hydropower equipment under the labeled working condition, which has an influence on the labeled monitoring parameter in the feature hydropower equipment under the target working condition, and the data set of the labeled monitoring parameter in the target historical record of the feature hydropower equipment under the target working condition, as the input of the model within the target historical period; Set the feature duration and the abnormal variable W, when the feature hydropower equipment does not have an abnormality within the feature duration after the target historical period, the abnormal variable W of the feature hydropower equipment within the target historical period is 0, otherwise, the abnormal variable W of the feature hydropower equipment within the target historical period is 1; Take the failure variable W of the feature hydropower equipment within the target historical period as the output, divide the target historical period into a training set and a test set according to the proportion, and use the large model technology to construct the abnormal monitoring model of the labeled monitoring parameter in the feature hydropower equipment; Use the abnormal monitoring model to perform abnormal monitoring on the labeled monitoring parameter in the feature hydropower equipment within the current period, use the threshold setting data to perform abnormal monitoring on the hydropower equipment in the hydropower station, acquire the hydropower equipment judged as abnormal in the hydropower station, and report through the platform.

6. A large model-based intelligent monitoring system for hydroelectric equipment, configured to perform the large model-based intelligent monitoring method for hydroelectric equipment according to any one of claims 1-5. The system comprises a working condition change sensitivity analysis module, a diagnosis accuracy analysis module, an influence scale analysis module and an equipment monitoring module; The working condition change sensitivity analysis module is configured to analyze the sensitivity of the monitoring parameters in the hydropower equipment to the working condition change, and obtain the feature hydropower equipment. The diagnosis accuracy analysis module is configured to analyze the diagnosis accuracy of the threshold setting data to the equipment state of the feature hydropower equipment under the working condition, and obtain the target working condition. The influence scale analysis module is configured to determine the labeled hydropower equipment and the labeled working condition, analyze the data influence scale of the labeled hydropower equipment under the labeled working condition to the monitoring parameters in the feature hydropower equipment under the target working condition, and generate the influence scale data. The equipment monitoring module is configured to construct the abnormal monitoring model of the feature hydropower equipment according to the influence scale data, and monitor the feature hydropower equipment.

7. The large model-based hydroelectric equipment intelligent monitoring system according to claim 6, wherein, The working condition change sensitivity analysis module comprises a record set acquisition unit and a working condition change sensitivity analysis unit. The record set acquisition unit is configured to monitor the equipment state of the hydropower equipment under the working condition, acquire and collect the historical monitoring records of the hydropower equipment under the working condition, and obtain the historical monitoring record set of the working condition. The working condition change sensitivity analysis unit is configured to calculate the sensitive value of the monitoring parameters in the hydropower equipment to the working condition change based on the historical monitoring record set of the working condition, and determine the sensitivity of the monitoring parameters in the hydropower equipment to the working condition change. The diagnosis accuracy analysis module comprises a distribution difference calculation unit and a diagnosis accuracy analysis unit.

8. The large model-based hydroelectric equipment intelligent monitoring system according to claim 6, wherein, The distribution difference calculation unit is configured to calculate the overall distribution difference of each monitoring parameter in the feature hydropower equipment to the threshold setting data under the working condition. The diagnosis accuracy analysis unit is configured to analyze the diagnosis accuracy of the threshold setting data to the equipment state of the feature hydropower equipment under the working condition according to the overall distribution difference, and obtain the target working condition. ​ 9. The large model-based hydroelectric equipment intelligent monitoring system according to claim 6, wherein, The influence scale analysis module comprises a state correlation analysis unit and an influence scale analysis unit; The state correlation analysis unit is configured to analyze the state correlation conditions between the other hydropower equipment in different working conditions and the characteristic hydropower equipment in the target working condition, and determine the marked hydropower equipment and the marked working condition; The influence scale analysis unit is configured to analyze the data influence scale of the marked hydropower equipment in the marked working condition on the monitoring parameters of the characteristic hydropower equipment in the target working condition, and generate the influence scale data of the characteristic hydropower equipment in the target working condition.

10. The large model-based hydroelectric equipment intelligent monitoring system according to claim 6, wherein, The equipment monitoring module comprises a model construction unit and an equipment monitoring unit; The model construction unit is configured to use a large model technology to construct an abnormal monitoring model of the marked monitoring parameters of the characteristic hydropower equipment in the target working condition based on the influence scale data of the characteristic hydropower equipment in the target working condition; The equipment monitoring unit is configured to use the abnormal monitoring model and combine threshold setting data to perform abnormal monitoring on the hydropower equipment in the hydropower station, obtain the hydropower equipment determined to be abnormal in the hydropower station, and report the hydropower equipment through the platform.

Images