Hydraulic power plant electromechanical equipment state monitoring informatization management system

Abstract

According to the hydraulic power plant electromechanical equipment state monitoring informatization management system provided by the embodiment of the invention, data monitoring and analysis are performed on pressure of a water inlet and a water outlet of a water turbine by adopting an artificial intelligence technology based on deep learning, and time sequence dynamic change modes of the pressure of the water inlet and the pressure of the water outlet in a local time domain are respectively extracted; and principal component feature joint coding is carried out on the water inlet pressure and the water outlet pressure in a local time domain to obtain pressure cooperative change information of the water inlet and the water outlet of the water turbine, and then global time domain aggregation is carried out on the pressure cooperative change information to realize intelligent identification of the leakage state of the water turbine. Therefore, the operation efficiency and the safety performance of the hydropower station can be effectively improved, and meanwhile the maintenance cost and the fault processing time are reduced.

Description

Technical Field

[0001] The embodiments of the present disclosure belong to the technical field of electromechanical equipment in hydropower plants, and specifically relate to an information management system for monitoring the status of electromechanical equipment in hydropower plants. Background Art

[0002] As one of the important sources of clean energy, hydropower stations occupy an important position in the global energy structure. By utilizing the potential energy and kinetic energy of water, hydropower stations convert the flow of water into electrical energy, which not only provides a large amount of renewable energy for modern society, but also effectively reduces dependence on fossil fuels, thus helping to mitigate climate change and environmental pollution. As the core electromechanical equipment of a hydropower station, the operating status of the turbine is directly related to the power generation efficiency and safety performance of the entire power station.

[0003] Traditionally, the monitoring of turbine operation status mainly relies on regular inspection and maintenance by experienced technicians. However, this method has many limitations, such as limited inspection frequency, greater influence of human factors, and inability to monitor in real time, which can easily lead to delayed detection and handling of faults.

[0004] At present, some existing equipment status monitoring systems use sensor technology for data collection. Although they can provide real-time data on the equipment's operating status, they often lack the ability to conduct in-depth analysis and intelligent processing of the data, and are unable to provide corresponding fault warnings and maintenance recommendations.

[0005] Therefore, an information management system for monitoring the status of electromechanical equipment in a hydropower plant is expected to achieve efficient monitoring and intelligent analysis of the operating status of turbine equipment. Summary of the invention

[0006] The embodiments of the present disclosure provide an information management system for monitoring the status of electromechanical equipment in a hydropower plant, aiming to solve at least one of the technical problems existing in the prior art, such as effectively improving the operating efficiency and safety performance of the hydropower station while reducing maintenance costs and fault handling time.

[0007] The embodiment of the present disclosure provides an information management system for monitoring the status of electromechanical equipment in a hydropower plant, including: a pressure monitoring module, which is used to obtain, on the edge side, a time queue of water inlet pressure and a time queue of water outlet pressure collected by pressure sensors deployed at the water inlet and water outlet of a water turbine; a data transmission module, which is used to transmit the time queue of the water inlet pressure and the time queue of the water outlet pressure to an information management center through a communication module on the edge side; and a pressure data local time series analysis module, which is used to perform time series analysis on the time queue of the water inlet pressure and the time queue of the water outlet pressure based on a local time scale at the information management center to obtain a time queue of the water inlet pressure. A module for jointly analyzing inlet and outlet pressure characteristics is used to perform local time scale-based characteristic principal component joint encoding on the sequence of implicit feature vectors of local time series association of inlet pressure and the sequence of implicit feature vectors of local time series association of outlet pressure in the information management center to obtain a sequence of inlet-outlet pressure local time series joint representation vectors; a leakage detection module is used to determine whether there is leakage in the turbine based on the global time series aggregation characteristics of the sequence of inlet-outlet pressure local time series joint representation vectors in the information management center.

[0008] In one possible implementation, the pressure data local time series analysis module is used to: split the time queue of the water inlet pressure and the time queue of the water outlet pressure according to a predetermined time scale and input them into a pressure sequence encoder based on a Bi-LSTM model respectively to obtain a sequence of implicit feature vectors associated with the local time series of the water inlet pressure and a sequence of implicit feature vectors associated with the local time series of the water outlet pressure.

[0009] In one possible implementation, the inlet and outlet pressure feature joint analysis module is used to: input each group of corresponding inlet pressure local time series associated implicit feature vectors and outlet pressure local time series associated implicit feature vectors in the sequence of inlet pressure local time series associated implicit feature vectors and the sequence of outlet pressure local time series associated implicit feature vectors into a significant fusion network based on feature principal component query matching to obtain a sequence of inlet-outlet pressure local time series joint representation vectors.

[0010] In a possible implementation, the inlet and outlet pressure feature joint analysis module includes: a standardization processing unit, which is used to standardize the inlet pressure local time series associated implicit feature vector and the corresponding outlet pressure local time series associated implicit feature vector to obtain a standardized inlet pressure local time series associated implicit feature vector and a standardized outlet pressure local time series associated implicit feature vector; a sample covariance matrix calculation unit, which is used to respectively calculate the sample covariance matrix of the standardized inlet pressure local time series associated implicit feature vector and the standardized outlet pressure local time series associated implicit feature vector to obtain an inlet pressure sample covariance matrix and an outlet pressure sample covariance matrix; a principal component feature extraction unit, which is used to perform feature vector extraction based on matrix decomposition on the inlet pressure sample covariance matrix and the outlet pressure sample covariance matrix to obtain the inlet pressure local time series principal component matrix. A set of feature vectors and a set of main component feature vectors of local time series of outlet pressure; a query matching unit, used to input the set of main component feature vectors of local time series of inlet pressure and the set of main component feature vectors of local time series of outlet pressure into a maximum approximation query matching network to obtain a set of best matching pairs of main component feature vectors of local time series of inlet pressure and main component feature vectors of local time series of outlet pressure; a matching union unit, used to input each best matching pair in the set of best matching pairs of main component feature vectors of local time series of inlet pressure and main component feature vectors of local time series of outlet pressure into a semantic fine-grained gated union module to obtain a set of inlet pressure-outlet pressure main component joint feature vectors; a principal component feature fusion unit, used to cascade the set of inlet pressure-outlet pressure main component joint feature vectors to obtain the inlet-outlet pressure local time series joint representation vector.

[0011] In a possible implementation, the query matching unit is used to: calculate the cosine similarity between the i-th water inlet pressure local time series principal component feature vector in the set of the water inlet pressure local time series principal component feature vectors and each water outlet pressure local time series principal component feature vector in the set of the water outlet pressure local time series principal component feature vectors to obtain a set of cosine similarities; and form the water outlet pressure local time series principal component feature vector corresponding to the maximum value in the set of cosine similarities with the i-th water inlet pressure local time series principal component feature vector into the best matching pair.

[0012] In a possible implementation, the matching joint unit is used to: calculate the point-by-point addition, point-by-point subtraction and point-by-point multiplication between the local time series principal component feature vector of the inlet pressure and the local time series principal component feature vector of the outlet pressure in the best matching pair to obtain a first matching fusion result, a second matching fusion result and a third matching fusion result; cascade the first matching fusion result, the second matching fusion result and the third matching fusion result into an inlet pressure-outlet pressure principal component multi-scale fusion feature vector, and then perform one-dimensional convolution processing and maximum pooling processing on it to obtain the inlet pressure-outlet pressure principal component joint feature vector.

[0013] In one possible implementation, the leakage detection module includes: a feature aggregation unit, which is used to input the sequence of the inlet-outlet pressure local time series joint representation vector into a feature aggregation network guided by node feature significance attenuation to obtain the inlet-outlet pressure time series aggregated joint representation vector; a leakage detection result generation unit, which is used to input the inlet-outlet pressure time series aggregated joint representation vector into a classifier-based state monitoring generator to obtain a state monitoring result, and the state monitoring result is used to indicate whether there is a leakage.

[0014] In a possible implementation, the feature aggregation unit includes: a feature significance calculation subunit, which is used to calculate the feature significance description factor of each inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector; a feature significance attenuation factor calculation subunit, which is used to use the last inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector as the current inlet-outlet pressure local time series joint representation vector, and based on the comparison between each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector and the current inlet-outlet pressure local time series joint representation vector. The distance span between them is used to construct the characteristic significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector; a significance modulation subunit is used to calculate the product between the characteristic significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector and its characteristic significance description factor to obtain a sequence of characteristic significance attenuation description factors; a weighting subunit is used to input the sequence of characteristic significance attenuation description factors into a gated mask module to obtain a sequence of characteristic significance attenuation weight factors; a weight application subunit is used to calculate the weighted sum of the sequence of inlet-outlet pressure local time series joint representation vectors based on the sequence of characteristic significance attenuation weight factors to obtain the inlet-outlet pressure time series aggregated joint representation vector.

[0015] In one possible implementation, the feature significance calculation subunit is used to: calculate the expected value of the fourth power of the difference between each eigenvalue in the inlet-outlet pressure local time series joint representation vector and its feature mean, and divide the expected value by the square of the feature variance of the inlet-outlet pressure local time series joint representation vector to obtain the feature significance description factor.

[0016] In one possible implementation, the feature significance attenuation factor calculation subunit is used to: calculate the difference between the maximum eigenvalue of the current inlet-outlet pressure local time series joint representation vector and the maximum eigenvalue of each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector, and then divide the difference by the number of eigenvectors between the two to obtain the feature significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector.

[0017] The embodiment of the present disclosure provides a hydropower plant electromechanical equipment status monitoring information management system, which uses artificial intelligence technology based on deep learning to monitor and analyze the pressure of the water inlet and outlet of the turbine, respectively extracts the time series dynamic change mode of the water inlet pressure and the water outlet pressure in the local time domain, and obtains the pressure coordinated change information of the water turbine inlet and outlet by jointly encoding the main component features of the water inlet pressure and the water outlet pressure in the local time domain, and then realizes the intelligent identification of the leakage state of the turbine by global time domain aggregation of this pressure coordinated change information. In this way, the operation efficiency and safety performance of the hydropower station can be effectively improved, while reducing maintenance costs and fault handling time. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic block diagram of an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to an embodiment of the present disclosure;

[0019] Figure 2 A data flow diagram of an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to another embodiment of the present disclosure;

[0020] Figure 3 It is a schematic block diagram of a joint analysis module for water inlet and outlet pressure characteristics in an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to another embodiment of the present disclosure;

[0021] Figure 4 A schematic block diagram of a leakage detection module in an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to another embodiment of the present disclosure;

[0022] Figure 5A schematic block diagram of a feature aggregation unit in an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to another embodiment of the present disclosure;

[0023] Figure 6 It is a schematic flowchart of an information management method for monitoring the status of electromechanical equipment in a hydropower plant according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solution of the present disclosure, the technical solution of the present disclosure is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0025] In view of the above technical problems, the technical concept of the embodiment of the present disclosure is to use artificial intelligence technology based on deep learning to monitor and analyze the pressure of the water inlet and outlet of the turbine, respectively extract the time series dynamic change mode of the water inlet pressure and the water outlet pressure in the local time domain, and jointly encode the main component features of the water inlet pressure and the water outlet pressure in the local time domain to obtain the pressure synergistic change information of the water turbine inlet and outlet, and then perform global time domain aggregation on this pressure synergistic change information, so as to realize the intelligent identification of the leakage state of the turbine. In this way, the operation efficiency and safety performance of the hydropower station can be effectively improved, while reducing maintenance costs and fault handling time.

[0026] Figure 1 It is a schematic block diagram of an information management system for monitoring the status of electromechanical equipment in a hydropower plant according to an embodiment of the present disclosure. Figure 2 Schematic diagram of data flow of the information management system for monitoring the status of electromechanical equipment in a hydropower plant according to an embodiment of the present disclosure. Figure 1 and Figure 2As shown, the information management system 100 for monitoring the status of electromechanical equipment in a hydropower plant includes: a pressure monitoring module 110, which is used to obtain the time queue of the water inlet pressure and the time queue of the water outlet pressure collected by the pressure sensors deployed at the water inlet and the water outlet of the turbine on the edge side. A data transmission module 120, which is used to transmit the time queue of the water inlet pressure and the time queue of the water outlet pressure to the information management center through the communication module on the edge side. A pressure data local time series analysis module 130, which is used to perform time series analysis based on the local time scale on the time queue of the water inlet pressure and the time queue of the water outlet pressure at the information management center to obtain a sequence of implicit feature vectors of the local time series association of the water inlet pressure and a sequence of implicit feature vectors of the local time series association of the water outlet pressure. The inlet and outlet pressure feature joint analysis module 140 is used to perform a local time scale-based feature principal component joint encoding on the sequence of the inlet pressure local time series associated implicit feature vectors and the sequence of the outlet pressure local time series associated implicit feature vectors in the information management center to obtain a sequence of inlet-outlet pressure local time series joint representation vectors. The leakage detection module 150 is used to determine whether there is a leak in the turbine based on the global time series aggregation characteristics of the sequence of the inlet-outlet pressure local time series joint representation vectors in the information management center.

[0027] In the above-mentioned hydropower plant electromechanical equipment status monitoring information management system, the pressure monitoring module 110 is used to obtain the time queue of the water inlet pressure and the time queue of the water outlet pressure collected by the pressure sensors deployed at the water inlet and outlet of the turbine on the edge side. It should be understood that when the turbine is operating normally, the pressure at its water inlet and outlet usually remains relatively stable. If the pressure at the water inlet and outlet fluctuates violently or changes abnormally, it may mean that the turbine may have abnormal problems such as leakage. Therefore, in the technical solution disclosed in the present invention, by real-time monitoring and analysis of the pressure at the water inlet and outlet, the abnormal state of the turbine can be effectively identified so that corresponding maintenance measures can be taken in time. In addition, by capturing pressure data in real time on the edge side, the timeliness and accuracy of data acquisition can be significantly improved.

[0028] In the above-mentioned hydropower plant electromechanical equipment status monitoring information management system, the data transmission module 120 is used to transmit the time queue of the water inlet pressure and the time queue of the water outlet pressure to the information management center through the communication module on the edge side. It should be understood that the time queue of the water inlet pressure and the time queue of the water outlet pressure are transmitted to the information management center through the communication module on the edge side, and the powerful computing resources equipped in the information management center are used to conduct in-depth mining and intelligent analysis of the pressure data of the water inlet and outlet of the turbine, so as to realize data monitoring and predictive maintenance of the operation status of the turbine.

[0029] In the above-mentioned hydropower plant electromechanical equipment status monitoring information management system, the pressure data local time series analysis module 130 is used to perform time series analysis based on the local time scale on the time queue of the water inlet pressure and the time queue of the water outlet pressure in the information management center to obtain a sequence of implicit feature vectors associated with the local time series of the water inlet pressure and a sequence of implicit feature vectors associated with the local time series of the water outlet pressure. It should be understood that since the pressure data of the water inlet and the water outlet are typical time series data, they have dynamic characteristics that change over time. Therefore, in order to be able to capture the change pattern of the pressure data in the time dimension in a fine-grained manner, in the information management center, the time queue of the water inlet pressure and the time queue of the water outlet pressure are first divided and processed separately at a predetermined time scale to reduce the complexity of the global analysis and enable the model to more finely capture the pressure change pattern in the local time domain. Then, a pressure sequence encoder based on the Bi-LSTM model is used to perform time series encoding on the pressure data in each local time domain after segmentation. The time series dependency of the pressure data is captured through its bidirectional structure, and the dynamic change characteristics of the inlet and outlet pressures in the local time domain are effectively extracted, thereby obtaining a sequence of implicit feature vectors of the local time series association of the inlet pressure and a sequence of implicit feature vectors of the local time series association of the outlet pressure, providing important information basis for subsequent turbine leakage detection.

[0030] Optionally, in one embodiment of the present disclosure, the local time series analysis module for pressure data is used to: split the time queue of the water inlet pressure and the time queue of the water outlet pressure according to a predetermined time scale and input them into a pressure sequence encoder based on a Bi-LSTM model respectively to obtain a sequence of implicit feature vectors associated with the local time series of the water inlet pressure and a sequence of implicit feature vectors associated with the local time series of the water outlet pressure.

[0031] In the above-mentioned hydropower plant electromechanical equipment status monitoring information management system, the inlet and outlet pressure feature joint analysis module 140 is used in the information management center to perform characteristic principal component joint encoding based on the local time scale on the sequence of implicit feature vectors associated with the local time series of the inlet pressure and the sequence of implicit feature vectors associated with the local time series of the outlet pressure to obtain a sequence of inlet-outlet pressure local time series joint representation vectors. It should be understood that when a turbine leaks, the pressure at the outlet usually drops due to the flow of water in an unexpected path. The pressure at the inlet is usually determined by the pressure provided upstream and may not change significantly after the leak occurs. Therefore, it is possible to identify whether a leak occurs by jointly analyzing the pressure characteristics of the inlet and outlet. Based on this, in the technical solution disclosed in the present invention, a significant fusion network based on feature principal component query matching is introduced to jointly analyze each group of corresponding implicit feature vectors of local temporal association of inlet pressure and implicit feature vectors of temporal association of outlet pressure in the sequence of implicit feature vectors of local temporal association of inlet pressure and the sequence of implicit feature vectors of temporal association of outlet pressure. Specifically, the network first standardizes the implicit feature vectors of local temporal association of inlet pressure and the implicit feature vectors of temporal association of outlet pressure in the same local time domain to ensure that the data is distributed on a scale of zero mean and has unit variance, eliminating the influence of different dimensions and magnitudes. Then, the principal component analysis algorithm is used to extract the principal component features of the implicit feature vectors of local temporal association of inlet pressure and the implicit feature vectors of temporal association of outlet pressure after standardization, and the main components are mined to generate a set of two principal component feature vectors. Furthermore, based on the maximum approximate query matching between feature vectors, the best matching pair is found by calculating and comparing the similarity between the two sets of principal component feature vectors. Then, a fine-grained semantic union is performed on each best matching pair, and the feature fusion results of each best matching pair are cascaded to aggregate the core time series information of the inlet pressure and the outlet pressure, and a local time series joint representation vector of the inlet-outlet pressure corresponding to each local time domain is generated.

[0032] Optionally, in one embodiment of the present disclosure, the inlet and outlet pressure feature joint analysis module is used to: input each group of corresponding inlet pressure local time series associated implicit feature vectors and outlet pressure local time series associated implicit feature vectors in the sequence of inlet pressure local time series associated implicit feature vectors and the sequence of outlet pressure local time series associated implicit feature vectors into a significant fusion network based on feature principal component query matching to obtain a sequence of inlet-outlet pressure local time series joint representation vectors.

[0033] Optionally, in one embodiment of the present disclosure, the inlet and outlet pressure feature joint analysis module 140 includes: a standardization processing unit 141, which is used to perform standardization processing on the inlet pressure local time series associated implicit feature vector and the corresponding outlet pressure local time series associated implicit feature vector to obtain a standardized inlet pressure local time series associated implicit feature vector and a standardized outlet pressure local time series associated implicit feature vector. A sample covariance matrix calculation unit 142, which is used to calculate the sample covariance matrix of the standardized inlet pressure local time series associated implicit feature vector and the standardized outlet pressure local time series associated implicit feature vector respectively to obtain an inlet pressure sample covariance matrix and an outlet pressure sample covariance matrix. A principal component feature extraction unit 143, which is used to perform feature vector extraction based on matrix decomposition on the inlet pressure sample covariance matrix and the outlet pressure sample covariance matrix to obtain a set of inlet pressure local time series principal component feature vectors and a set of outlet pressure local time series principal component feature vectors. The query matching unit 144 is used to input the set of the inlet pressure local time series principal component feature vectors and the set of the outlet pressure local time series principal component feature vectors into the maximum approximate query matching network to obtain the set of the best matching pairs of the inlet pressure local time series principal component feature vectors and the outlet pressure local time series principal component feature vectors. The matching union unit 145 is used to input each best matching pair in the set of the best matching pairs of the inlet pressure local time series principal component feature vectors and the outlet pressure local time series principal component feature vectors into the semantic fine-grained gated union module to obtain the set of inlet pressure-outlet pressure principal component joint feature vectors. The principal component feature fusion unit 146 is used to cascade the set of the inlet pressure-outlet pressure principal component joint feature vectors to obtain the inlet-outlet pressure local time series joint representation vector.

[0034] Optionally, in one embodiment of the present disclosure, the query matching unit is used to: calculate the cosine similarity between the i-th water inlet pressure local time series principal component feature vector in the set of the water inlet pressure local time series principal component feature vectors and each water outlet pressure local time series principal component feature vector in the set of the water outlet pressure local time series principal component feature vectors to obtain a set of cosine similarities. The water outlet pressure local time series principal component feature vector corresponding to the maximum value in the set of cosine similarities and the i-th water inlet pressure local time series principal component feature vector form the best matching pair.

[0035] Optionally, in one embodiment of the present disclosure, the matching joint unit is used to: calculate the position point addition, position point subtraction and position point multiplication between the local time series principal component feature vector of the water inlet pressure and the local time series principal component feature vector of the water outlet pressure in the best matching pair to obtain the first matching fusion result, the second matching fusion result and the third matching fusion result. After the first matching fusion result, the second matching fusion result and the third matching fusion result are cascaded into a multi-scale fusion feature vector of the water inlet pressure-water outlet pressure principal component, one-dimensional convolution processing and maximum pooling processing are performed on them to obtain the water inlet pressure-water outlet pressure principal component joint feature vector.

[0036] In summary, in the above embodiment, the inlet and outlet pressure feature joint analysis module 140 is used to: fuse the inlet pressure local time series associated implicit feature vector and the outlet pressure local time series associated implicit feature vector using the following principal component query matching formula to obtain the inlet-outlet pressure local time series joint representation vector, wherein the principal component query matching formula is:

[0037]

[0038] C1=U1Λ1U1 T

[0039] C2=U2Λ2U2 T

[0040] U1=[v 11 , v 12 , ..., v 1m ]

[0041] Λ1=diag(λ 11 ,λ 12 , ..., λ 1m )

[0042] U2=[v 21 , v 22 , ..., v 2m ]

[0043] Λ2=diag(λ 21 ,λ 22 , ..., λ 2m )

[0044]

[0045] x=[v p1 ;v p2 ;...v pm ]

[0046] Wherein, V1 represents the implicit characteristic vector of the local time series correlation of the water inlet pressure, μ1 and σ1 represent the mean and standard deviation of the implicit characteristic vector of the local time series correlation of the water inlet pressure, respectively. 1norm represents the local time series associated implicit eigenvector of the standardized inlet pressure, V2 represents the local time series associated implicit eigenvector of the outlet pressure, μ2 and σ2 represent the mean and standard deviation of the local time series associated implicit eigenvector of the outlet pressure, respectively. 2norm represents the standardized outlet pressure local time series associated implicit eigenvector, the value of n1 is the length of the inlet pressure local time series associated implicit eigenvector, the value of n2 is the length of the outlet pressure local time series associated implicit eigenvector, C1 represents the inlet pressure sample covariance matrix, C2 represents the outlet pressure sample covariance matrix, U1 represents the matrix composed of the set of inlet pressure local time series principal component eigenvectors obtained by performing eigenvalue decomposition on the inlet pressure sample covariance matrix, Λ1 represents the diagonal matrix composed of the set of inlet pressure local time series principal component eigenvalues ​​obtained by performing eigenvalue decomposition on the inlet pressure sample covariance matrix, v 11 、v 12 、v 1i and v 1m Respectively represent the first, second, i-th and m-th principal component eigenvectors of the local time series of the water inlet pressure in the set of the principal component eigenvectors of the local time series of the water inlet pressure, λ 11 , 12 and λ 1m represents the first, second and mth local time series principal component eigenvalues ​​of the inlet pressure in the set of the local time series principal component eigenvalues ​​of the inlet pressure, respectively, the value of m is the number of the local time series principal component eigenvalues ​​of the inlet pressure, U2 represents the matrix composed of the set of main component eigenvectors of the local time series of the outlet pressure obtained by performing eigenvalue decomposition on the covariance matrix of the outlet pressure samples, Λ2 represents the diagonal matrix composed of the set of main component eigenvalues ​​of the local time series of the outlet pressure obtained by performing eigenvalue decomposition on the covariance matrix of the outlet pressure samples, v 21 、v 22 、v 2j 、v 2k and v 2m represents the first, second, jth, kth and mth local time series principal component eigenvectors of the outlet pressure in the set of the local time series principal component eigenvectors of the outlet pressure, respectively, 21 , 22 and λ 2mdenote the first, second and mth local time series principal component eigenvalues ​​of the outlet pressure in the set of the local time series principal component eigenvalues ​​of the outlet pressure, respectively; diag is the diag function used to construct a diagonal matrix; <·> denotes the inner product; ‖·‖ denotes the modulus of the eigenvector; argmax(·) denotes the index corresponding to the maximum value; k denotes the index of the local time series principal component eigenvector of the outlet pressure that forms the best matching pair with the i-th local time series principal component eigenvector of the inlet pressure; [·,·] denotes a cascade operation; MaxPool(·) denotes a maximum pooling operation; conv1D(·) denotes a one-dimensional convolution operation; ⊙ denotes a dot product; Indicates point addition, Indicates point minus, v pi represents the i-th inlet pressure-outlet pressure principal component joint eigenvector, and x represents the inlet-outlet pressure local time series joint representation vector.

[0047] In the above-mentioned hydropower plant electromechanical equipment status monitoring information management system, the leakage detection module 150 is used to determine whether there is a leak in the turbine based on the global time series aggregation characteristics of the sequence of the local time series joint representation vector of the water inlet-outlet pressure in the information management center. It should be understood that the local time series joint representation vector of the water inlet-outlet pressure in each local time domain is further integrated to capture the long-term trend and periodic changes of the operating state of the turbine. At the same time, considering that the feature information of the early time period may contribute less to the overall state recognition when performing information aggregation, the embodiment of the present disclosure introduces a feature aggregation network guided by node feature significance attenuation to process the sequence of the local time series joint representation vector of the water inlet-outlet pressure, and dynamically adjusts its weight in the aggregation process based on the feature significance and time series attenuation characteristics of each local time domain information, thereby improving the accuracy and reliability of feature aggregation. Specifically, the network first calculates its feature significance description factor based on the mean and variance of each local time series joint representation vector of the water inlet-outlet pressure to quantify the importance of the representation feature. Next, based on the distance span between each inlet-outlet pressure local time series joint representation vector and the current inlet-outlet pressure local time series joint representation vector (time series end), a feature significance attenuation factor is constructed, and its feature significance description factor is modulated to make the local time domain information closer to the current moment in the feature space obtain a higher weight, thereby more accurately reflecting the relative importance of each local time domain feature in the overall sequence. Then, the modulated description factor is further screened using the activation function and gating mechanism, and the screened description factor is used as the weight to perform weighted summation on the inlet-outlet pressure local time series joint representation vector in each local time domain, so as to integrate each local time domain information and generate a global inlet-outlet pressure time series aggregated joint representation vector.

[0048] Finally, the inlet-outlet pressure time series aggregated joint representation vector is input into a classifier-based state monitoring generator to obtain a state monitoring result, which is used to indicate whether there is a leak. In the technical solution disclosed in the present invention, the classifier can effectively identify the pressure feature differences under normal operation and leakage state through training and learning, thereby accurately judging whether there is a leakage problem in the turbine.

[0049] Optionally, in one embodiment of the present disclosure, the leakage detection module 150 includes: a feature aggregation unit 151, which is used to input the sequence of the local time series joint representation vector of the water inlet-water outlet pressure into a feature aggregation network guided by node feature significance attenuation to obtain the water inlet-water outlet pressure time series aggregated joint representation vector. A leakage detection result generation unit 152, which is used to input the water inlet-water outlet pressure time series aggregated joint representation vector into a state monitoring generator based on a classifier to obtain a state monitoring result, and the state monitoring result is used to indicate whether there is a leak.

[0050] Optionally, in one embodiment of the present disclosure, the feature aggregation unit 151 includes: a feature significance calculation subunit 1511, which is used to calculate the feature significance description factor of each inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector. A feature significance attenuation factor calculation subunit 1512, which is used to take the last inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector as the current inlet-outlet pressure local time series joint representation vector, and construct the feature significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector based on the distance span between each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector and the current inlet-outlet pressure local time series joint representation vector. The significance modulation subunit 1513 is used to calculate the product between the characteristic significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector and its characteristic significance description factor to obtain a sequence of characteristic significance attenuation description factors. The weighting subunit 1514 is used to input the sequence of characteristic significance attenuation description factors into the gated mask module to obtain a sequence of characteristic significance attenuation weight factors. The weight application subunit 1515 is used to calculate the weighted sum of the sequence of the inlet-outlet pressure local time series joint representation vector based on the sequence of characteristic significance attenuation weight factors to obtain the inlet-outlet pressure time series aggregated joint representation vector.

[0051] Optionally, in one embodiment of the present disclosure, the feature significance calculation subunit is used to: calculate the expected value of the fourth power of the difference between each eigenvalue in the inlet-outlet pressure local time series joint representation vector and its feature mean, and divide the expected value by the square of the feature variance of the inlet-outlet pressure local time series joint representation vector to obtain the feature significance description factor.

[0052] Optionally, in one embodiment of the present disclosure, the feature significance attenuation factor calculation subunit is used to: calculate the difference between the maximum eigenvalue of the current inlet-outlet pressure local time series joint representation vector and the maximum eigenvalue of each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector, and then divide the difference by the number of eigenvectors separated by the two to obtain the feature significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector.

[0053] In summary, in the above embodiment, the feature aggregation unit 151 is used to process the sequence of the inlet-outlet pressure local time series joint representation vectors using the following feature significance attenuation guided fusion formula to obtain the inlet-outlet pressure time series aggregated joint representation vector, wherein the feature significance attenuation guided fusion formula is:

[0054] X={x1,x2,...,x k ,...,x n}

[0055]

[0056] ws i =mask(w i )

[0057]

[0058] Wherein, X represents the sequence of the joint representation vector of the local time series of the water inlet-water outlet pressure, the value of n is the number of characteristic vectors in the sequence of the joint representation vector of the local time series of the water inlet-water outlet pressure, and x1, x2, x i 、x k and x n denote the first, second, i-th, k-th and current joint representation vectors of the local time series of the inlet-outlet pressure in the sequence of the joint representation vectors of the local time series of the inlet-outlet pressure, max(·) denotes the maximum eigenvalue of the eigenvector, x i (j) represents the jth eigenvalue of the local time series joint representation vector of the i-th inlet-outlet pressure, μ i and σ i 4 They represent the characteristic mean and square of the characteristic variance of the local time series joint representation vector of the i-th inlet-outlet pressure, respectively, E{[x i (j)-μ i ] 4} represents the fourth-order central moment of the joint representation vector of the local time series of the i-th inlet-outlet pressure, that is, the expected value of the fourth power of the difference between each eigenvalue and its eigenmean in the joint representation vector of the local time series of the i-th inlet-outlet pressure, i∈[0,n-1], τ i represents the characteristic significance attenuation factor of the local time series joint representation vector of the i-th inlet-outlet pressure, k i represents the descriptive factor of the i-th feature significance attenuation, w i represents the descriptive factor of the ith normalized feature significance attenuation, ws i represents the i-th feature significance attenuation weight factor, θ is the preset threshold, mask(·) represents mask processing, X a Represents the inlet-outlet pressure time series aggregation joint representation vector.

[0059] In summary, according to the disclosed embodiment of the hydropower plant electromechanical equipment status monitoring information management system is explained, and the artificial intelligence technology based on deep learning is used to monitor and analyze the pressure of the water inlet and outlet of the turbine, and the time series dynamic change mode of the water inlet pressure and the water outlet pressure in the local time domain is extracted respectively, and the main component feature joint encoding of the water inlet pressure and the water outlet pressure in the local time domain is performed to obtain the pressure synergistic change information of the water turbine inlet and outlet, and then the global time domain aggregation of this pressure synergistic change information is performed to realize the intelligent identification of the leakage state of the turbine. In this way, the operation efficiency and safety performance of the hydropower station can be effectively improved, while reducing the maintenance cost and fault handling time.

[0060] Figure 6 FIG. 1 is a schematic flow chart of an information management method for monitoring the status of electromechanical equipment in a hydropower plant according to an embodiment of the present disclosure. Figure 6As shown, the information management method for the condition monitoring of electromechanical equipment in a hydropower plant includes: S1, obtaining the time queue of the water inlet pressure and the time queue of the water outlet pressure collected by the pressure sensors deployed at the water inlet and the water outlet of the turbine on the edge side. S2, transmitting the time queue of the water inlet pressure and the time queue of the water outlet pressure to the information management center through the communication module on the edge side. S3, in the information management center, performing a time series analysis based on the local time scale on the time queue of the water inlet pressure and the time queue of the water outlet pressure to obtain a sequence of implicit feature vectors of the local time series association of the water inlet pressure and a sequence of implicit feature vectors of the local time series association of the water outlet pressure. S4, in the information management center, performing a characteristic principal component joint encoding based on the local time scale on the sequence of implicit feature vectors of the local time series association of the water inlet pressure and the sequence of implicit feature vectors of the local time series association of the water outlet pressure to obtain a sequence of joint representation vectors of the local time series of the water inlet-water outlet pressure. S5, at the information management center, determining whether there is leakage in the turbine based on the global time series aggregation characteristics of the sequence of the local time series joint representation vector of the water inlet-water outlet pressure.

[0061] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned hydropower plant electromechanical equipment status monitoring information management method have been referred to above. Figures 1 to 5 The invention has been introduced in detail in the description of the information management system for monitoring the condition of electromechanical equipment in a hydropower plant, and therefore, its repeated description will be omitted.

[0062] The basic principles of the embodiments of the present disclosure are described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the embodiments of the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details of the above embodiments are only for the purpose of illustration and ease of understanding, rather than limitation, and the above details do not limit the embodiments of the present disclosure to the necessity of adopting the above specific details to be implemented.

[0063] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided in the embodiments of the present disclosure, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0064] In addition, each functional module in each embodiment of the present disclosure may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0065] It is obvious to those skilled in the art that the embodiments of the present disclosure are not limited to the details of the above exemplary embodiments, and that the embodiments of the present disclosure can be implemented in other specific forms without departing from the spirit or basic features of the embodiments of the present disclosure. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the embodiments of the present disclosure is defined by the attached claims rather than the above description, and it is intended that all changes that fall within the meaning and scope of the equivalent elements of the claims are included in the embodiments of the present disclosure. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0066] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the device claim can also be implemented by one unit through software or hardware.

[0067] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the embodiments of the present disclosure, but the embodiments of the present disclosure are not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the embodiments of the present disclosure, and these modifications and improvements are also considered to be within the protection scope of the embodiments of the present disclosure.

Claims

1. A hydropower plant electromechanical equipment status monitoring information management system, characterized in that: include: A pressure monitoring module, used to obtain, at the edge side, a time queue of water inlet pressure and a time queue of water outlet pressure collected by pressure sensors deployed at the water inlet and water outlet of the turbine; A data transmission module, used for transmitting the time queue of the water inlet pressure and the time queue of the water outlet pressure to an information management center through the communication module on the edge side; A local time series analysis module for pressure data is used to perform time series analysis based on the local time scale on the time queue of the water inlet pressure and the time queue of the water outlet pressure in the information management center to obtain a sequence of implicit feature vectors associated with the local time series of the water inlet pressure and a sequence of implicit feature vectors associated with the local time series of the water outlet pressure; The inlet and outlet pressure characteristic joint analysis module is used to perform characteristic principal component joint encoding based on the local time scale on the sequence of implicit characteristic vectors associated with the local time series of the inlet pressure and the sequence of implicit characteristic vectors associated with the local time series of the outlet pressure in the information management center to obtain a sequence of joint representation vectors of the local time series of the inlet and outlet pressures; The leakage detection module is used to determine whether there is leakage in the turbine based on the global time series aggregation characteristics of the sequence of the local time series joint representation vector of the water inlet-water outlet pressure in the information management center.

2. The hydropower plant electromechanical equipment status monitoring information management system according to claim 1 is characterized in that: The pressure data local time series analysis module is used for: The time queue of the water inlet pressure and the time queue of the water outlet pressure are divided into predetermined time scales and respectively input into the pressure sequence encoder based on the Bi-LSTM model to obtain a sequence of implicit feature vectors of local time series association of the water inlet pressure and a sequence of implicit feature vectors of local time series association of the water outlet pressure.

3. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 2 is characterized in that: The inlet and outlet pressure characteristics joint analysis module is used to: Each group of corresponding implicit feature vectors of local time series association of inlet pressure and implicit feature vectors of local time series association of outlet pressure in the sequence of implicit feature vectors of local time series association of inlet pressure and the sequence of implicit feature vectors of local time series association of outlet pressure is input into a significant fusion network based on feature principal component query matching to obtain a sequence of joint representation vectors of local time series association of inlet and outlet pressure.

4. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 3 is characterized in that: The inlet and outlet pressure characteristics joint analysis module includes: A standardization processing unit, used for standardizing the water inlet pressure local time series associated implicit feature vector and the corresponding water outlet pressure local time series associated implicit feature vector to obtain a standardized water inlet pressure local time series associated implicit feature vector and a standardized water outlet pressure local time series associated implicit feature vector; A sample covariance matrix calculation unit, used to calculate the sample covariance matrices of the standardized local time series associated implicit eigenvector of the water inlet pressure and the standardized local time series associated implicit eigenvector of the water outlet pressure respectively to obtain a water inlet pressure sample covariance matrix and a water outlet pressure sample covariance matrix; A principal component feature extraction unit, used for performing feature vector extraction based on matrix decomposition on the water inlet pressure sample covariance matrix and the water outlet pressure sample covariance matrix to obtain a set of principal component feature vectors of the local time series of the water inlet pressure and a set of principal component feature vectors of the local time series of the water outlet pressure; A query matching unit, used for inputting the set of the inlet pressure local time series principal component feature vectors and the set of the outlet pressure local time series principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of the inlet pressure local time series principal component feature vectors and the outlet pressure local time series principal component feature vectors; A matching and combining unit, used for inputting each best matching pair in the set of best matching pairs of the local time series principal component feature vector of the water inlet pressure and the local time series principal component feature vector of the water outlet pressure into a semantic fine-grained gated combining module to obtain a set of the water inlet pressure-water outlet pressure principal component combined feature vectors; The principal component feature fusion unit is used to cascade the set of the inlet pressure-outlet pressure principal component joint feature vectors to obtain the inlet-outlet pressure local time series joint representation vector.

5. The hydropower plant electromechanical equipment status monitoring information management system according to claim 4 is characterized in that: The query matching unit is used to: Calculating the cosine similarity between the i-th water inlet pressure local time series principal component eigenvector in the set of water inlet pressure local time series principal component eigenvectors and each water outlet pressure local time series principal component eigenvector in the set of water outlet pressure local time series principal component eigenvectors to obtain a set of cosine similarities; The outlet pressure local time series principal component feature vector corresponding to the maximum value in the set of cosine similarities and the i-th inlet pressure local time series principal component feature vector form the best matching pair.

6. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 5 is characterized in that: The matching unit is used for: Calculate the position point addition, position point subtraction and position point multiplication between the local time series principal component eigenvector of the water inlet pressure and the local time series principal component eigenvector of the water outlet pressure in the best matching pair to obtain a first matching fusion result, a second matching fusion result and a third matching fusion result; After the first matching fusion result, the second matching fusion result and the third matching fusion result are cascaded into a multi-scale fusion feature vector of the inlet pressure-outlet pressure principal component, one-dimensional convolution processing and maximum pooling processing are performed on them to obtain the inlet pressure-outlet pressure principal component joint feature vector.

7. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 6 is characterized in that: The leakage detection module comprises: A feature aggregation unit, used for inputting the sequence of the inlet-outlet pressure local time series joint representation vector into a feature aggregation network guided by node feature significance attenuation to obtain an inlet-outlet pressure time series aggregated joint representation vector; The leakage detection result generating unit is used to input the water inlet-water outlet pressure time series aggregation joint representation vector into the classifier-based state monitoring generator to obtain a state monitoring result, and the state monitoring result is used to indicate whether there is a leakage.

8. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 7 is characterized in that: The feature aggregation unit comprises: A feature significance calculation subunit, used to calculate a feature significance description factor of each water inlet-outlet pressure local time series joint representation vector in the sequence of the water inlet-outlet pressure local time series joint representation vector; A characteristic significance attenuation factor calculation subunit is used to use the last inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector as the current inlet-outlet pressure local time series joint representation vector, and construct characteristic significance attenuation factors of each other inlet-outlet pressure local time series joint representation vector based on the distance span between each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector and the current inlet-outlet pressure local time series joint representation vector; A significance modulation subunit, used for calculating the product between the characteristic significance attenuation factor of each other water inlet-water outlet pressure local time series joint representation vector and its characteristic significance description factor to obtain a sequence of characteristic significance attenuation description factors; A weighting subunit, configured to input the sequence of feature significance attenuation description factors into a gated mask module to obtain a sequence of feature significance attenuation weighting factors; The weight applying subunit is used to calculate the weighted sum of the sequence of the inlet-outlet pressure local time series joint representation vectors based on the sequence of the feature significance attenuation weight factors to obtain the inlet-outlet pressure time series aggregated joint representation vector.

9. The hydropower plant electromechanical equipment status monitoring information management system according to claim 8 is characterized in that: The feature significance calculation subunit is used to: Calculate the expected value of the fourth power of the difference between each eigenvalue in the inlet-outlet pressure local time series joint representation vector and its eigenvalue mean, and divide the expected value by the square of the eigenvalue variance of the inlet-outlet pressure local time series joint representation vector to obtain the feature significance description factor.

10. The information management system for monitoring the status of electromechanical equipment in a hydropower plant according to claim 9, characterized in that: The feature significance attenuation factor calculation subunit is used to: Calculate the difference between the maximum eigenvalue of the current inlet-outlet pressure local time series joint representation vector and the maximum eigenvalue of each other inlet-outlet pressure local time series joint representation vector in the sequence of the inlet-outlet pressure local time series joint representation vector, and then divide the difference by the number of eigenvectors between the two to obtain the characteristic significance attenuation factor of each other inlet-outlet pressure local time series joint representation vector.

Images