A method for formulating energy storage operation and maintenance plan

Abstract

The present invention discloses a method for formulating an energy storage operation and maintenance plan. The method obtains the equipment operation data from the sensor, performs data preprocessing on the equipment operation data, and obtains the initial equipment operation data. The Apriori association rule analysis algorithm is used to obtain the potential correlation between different operation parameters and between operation parameters and faults in the initial equipment operation data. The operation status of the equipment in the associated equipment operation data is classified based on the DBSCAN cluster analysis algorithm. The similar equipment operation data is input into the RNN-SVM equipment status evaluation model for training. The energy storage equipment status is evaluated according to the equipment status evaluation result. When the equipment status indicator approaches the preset threshold, the early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel. The equipment status can be evaluated more accurately, providing a scientific basis for formulating a reasonable operation and maintenance plan, and effectively improving the pertinence and accuracy of the operation and maintenance plan.

Description

Technical Field

[0001] The present invention relates to the field of energy storage technology, and in particular to a method for formulating an energy storage operation and maintenance plan. Background Art

[0002] With the continuous development of energy storage technology, energy storage systems are increasingly used in the fields of electricity, new energy, etc. However, the stable operation and efficient maintenance of energy storage systems are crucial to their performance and service life. At present, traditional energy storage operation and maintenance plan formulation methods often lack systematicity and scientificity, and it is difficult to fully consider factors such as the performance changes of energy storage equipment, the operating environment, and the cost-effectiveness of different maintenance strategies, resulting in inaccurate operation and maintenance plans, and there may be problems such as excessive maintenance increasing costs or insufficient maintenance affecting equipment reliability. Therefore, it is necessary to develop a more scientific and reasonable energy storage operation and maintenance plan formulation method to improve the operation and maintenance efficiency and economy of energy storage systems. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and to design a method for formulating energy storage operation and maintenance plan.

[0004] To achieve the above object, the technical solution of the present invention is that, further, in the above energy storage operation and maintenance plan formulation method, the crowd flow monitoring method includes the following steps:

[0005] Acquire device operation data from a sensor, perform data preprocessing on the device operation data, and obtain initial device operation data;

[0006] Apriori association rule analysis algorithm is used to obtain potential associations between different operating parameters in the initial equipment operating data and between operating parameters and faults, so as to obtain associated equipment operating data;

[0007] Classifying the operating status of the devices in the associated device operating data based on the DBSCAN clustering analysis algorithm to obtain similar device operating data;

[0008] Establish an RNN-SVM equipment status assessment model based on RNN recurrent neural network and SVM support vector machine;

[0009] Inputting the similar equipment operation data into the RNN-SVM equipment status assessment model for training to obtain equipment status assessment results;

[0010] The energy storage device status is evaluated according to the device status evaluation result. When the device status indicator approaches a preset threshold, an early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel.

[0011] Furthermore, in the above energy storage operation and maintenance plan formulation method, the acquiring of equipment operation data from the sensor and the preprocessing of the equipment operation data to obtain initial equipment operation data include:

[0012] Acquire device operation data from the sensor, wherein the device operation data at least includes voltage operation data, current operation data, temperature operation data, charge and discharge times operation data, energy conversion efficiency operation data, temperature data, humidity data, and air quality data of the device;

[0013] Integrate and clean the equipment operation data, remove abnormal values ​​and noise in the equipment operation data, and obtain complete equipment operation data;

[0014] The complete equipment operation data is normalized to obtain initial equipment operation data.

[0015] Furthermore, in the above-mentioned energy storage operation and maintenance plan formulation method, the Apriori association rule analysis algorithm is used to obtain the potential association relationship between different operating parameters and between operating parameters and faults in the initial equipment operation data to obtain the associated equipment operation data, including:

[0016] Scan the initial equipment operation data, count the number of occurrences of each single fault type, and determine a frequent 1-item set that meets a support threshold;

[0017] Based on the frequent 1-item set, a candidate 2-item set is generated through a connection operation, and the initial device operation data is scanned again, the number of occurrences of the candidate 2-item set is counted, and a frequent 2-item set that meets a support threshold is determined;

[0018] Repeat the above process, continuously generate higher-order candidate item sets through connection and pruning operations, and filter out frequent item sets that meet the support threshold until no new frequent item sets can be generated;

[0019] For each frequent item set, all possible non-empty subsets are generated, and the confidence between each subset and other items is calculated. If the confidence meets the set threshold, the association relationship between the subset and other items is output as a valid association rule to obtain the associated equipment operation data.

[0020] Furthermore, in the above-mentioned energy storage operation and maintenance plan formulation method, the operation status of the equipment in the associated equipment operation data is classified based on the DBSCAN clustering analysis algorithm to obtain similar equipment operation data, including:

[0021] Determine the neighborhood radius Eps and the minimum number of samples MinPts of the DBSCAN clustering analysis algorithm;

[0022] For each point in the associated device operation data, the number of points in its neighborhood is calculated according to the neighborhood radius Eps and the minimum number of samples MinPts;

[0023] Starting from any core point that has not been visited, all points in its neighborhood are assigned to the same cluster. For the points in these neighborhoods, their neighborhoods are continued to be expanded, and newly discovered points belonging to the cluster are added to the cluster until no new points can be added, and the initial cluster is obtained;

[0024] Repeat the above process, and continue to expand the clusters for other core points that are not assigned to any clusters until all core points are processed and the operation data of similar devices are obtained.

[0025] Furthermore, in the above-mentioned energy storage operation and maintenance plan formulation method, the similar equipment operation data is input into the RNN-SVM equipment status assessment model for training to obtain the equipment status assessment result, including:

[0026] Dividing the similar equipment operation data to obtain a training data set and a test data set;

[0027] Use the training set data to train the RNN, input the preprocessed equipment operation data into the RNN model according to the batch size and time step, and calculate the output of the model through forward propagation;

[0028] Calculate the loss function based on the output results and the actual device status label, use the back propagation algorithm to calculate the gradient of the loss function to the parameters of each layer of the RNN, and update the parameters through the optimization algorithm;

[0029] The feature vector obtained by RNN training is used as the input of SVM, and the label corresponding to the device status is used as the output target of SVM;

[0030] After using the trained RNN to extract features from the validation set data, the extracted features and corresponding labels are input into the SVM for training to obtain the equipment status assessment results.

[0031] Furthermore, in the above energy storage operation and maintenance plan formulation method, the energy storage device status is evaluated according to the device status evaluation result, and when the device status indicator approaches a preset threshold, an early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel, including:

[0032] Set corresponding thresholds according to each equipment status evaluation indicator, wherein the thresholds at least include a normal range threshold, a warning threshold and a danger threshold;

[0033] The normal range threshold defines the reasonable fluctuation range of the indicator under normal operation of the equipment. The warning threshold is the value set when the indicator is close to the boundary of a problem. The danger threshold is the extreme value of the indicator.

[0034] When the equipment status indicators approach the preset warning threshold and danger threshold, the warning mechanism is immediately triggered and warning information is sent to the operation and maintenance personnel.

[0035] Furthermore, a system for formulating a method for energy storage operation and maintenance plan is implemented, characterized in that the system includes the following modules:

[0036] A data preprocessing module is used to obtain the device operation data in the sensor, perform data preprocessing on the device operation data, and obtain initial device operation data;

[0037] An association analysis module is used to obtain potential associations between different operating parameters and between operating parameters and faults in the initial equipment operating data by using an Apriori association rule analysis algorithm to obtain associated equipment operating data;

[0038] A cluster analysis module, used for classifying the operation status of the equipment in the associated equipment operation data based on the DBSCAN cluster analysis algorithm to obtain similar equipment operation data;

[0039] A model building module is used to build an RNN-SVM equipment status assessment model based on the RNN recurrent neural network and the SVM support vector machine;

[0040] A state recognition module is used to input the similar equipment operation data into the RNN-SVM equipment state assessment model for training to obtain equipment state assessment results;

[0041] The energy storage early warning module is used to evaluate the state of the energy storage device according to the device state evaluation result. When the device state indicator approaches a preset threshold, the early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel.

[0042] Furthermore, a system for formulating a method for energy storage operation and maintenance plan is implemented, characterized in that the system includes the following modules:

[0043] An acquisition submodule is used to acquire device operation data in the sensor, wherein the device operation data at least includes voltage operation data, current operation data, temperature operation data, charge and discharge times operation data, energy conversion efficiency operation data, temperature data, humidity data, and air quality data of the device;

[0044] A cleaning submodule is used to integrate and clean the equipment operation data, remove abnormal values ​​and noise in the equipment operation data, and obtain complete equipment operation data;

[0045] The submodule is used to normalize the complete device operation data to obtain initial device operation data.

[0046] Furthermore, a system for formulating a method for energy storage operation and maintenance plan is implemented, characterized in that the system includes the following modules:

[0047] A scanning submodule, used to scan the initial equipment operation data, count the number of occurrences of each single fault type, and determine a frequent 1-item set that meets a support threshold;

[0048] A connection submodule, configured to generate candidate 2-item sets through a connection operation based on the frequent 1-item set, and scan the initial device operation data again, count the number of occurrences of the candidate 2-item sets, and determine the frequent 2-item sets that meet the support threshold;

[0049] The screening submodule is used to repeat the above process, continuously generate higher-order candidate item sets through connection and pruning operations, and screen out frequent item sets that meet the support threshold until no new frequent item sets can be generated;

[0050] The output submodule is used to generate all possible non-empty subsets of each frequent item set and calculate the confidence between each subset and other items. If the confidence meets the set threshold, the association relationship between the subset and other items is output as a valid association rule to obtain the associated device operation data.

[0051] Its beneficial effects are that, by acquiring the equipment operation data in the sensor, the equipment operation data is preprocessed to obtain the initial equipment operation data; the Apriori association rule analysis algorithm is used to obtain the potential association between different operation parameters and between operation parameters and faults in the initial equipment operation data to obtain the associated equipment operation data; the operation status of the equipment in the associated equipment operation data is classified based on the DBSCAN clustering analysis algorithm to obtain similar equipment operation data; an RNN-SVM equipment state evaluation model is established based on the RNN recurrent neural network and the SVM support vector machine; the similar equipment operation data is input into the RNN-SVM equipment state evaluation model for training to obtain the equipment state evaluation result; the energy storage equipment state is evaluated according to the equipment state evaluation result, and when the equipment state index approaches the preset threshold, the early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel. The equipment state can be evaluated more accurately, providing a scientific basis for formulating a reasonable operation and maintenance plan, and effectively improving the pertinence and accuracy of the operation and maintenance plan. It can monitor the operation status of the equipment in real time, predict faults in advance, realize preventive and predictive maintenance, reduce the probability of sudden equipment failures, and improve the reliability and service life of the equipment. It can improve operation and maintenance efficiency, reduce maintenance costs, and achieve economical and sustainable development of energy storage system operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0053] Figure 1 A schematic diagram of a first embodiment of a method for formulating an energy storage operation and maintenance plan in an embodiment of the present invention;

[0054] Figure 2 A schematic diagram of a second embodiment of a method for formulating an energy storage operation and maintenance plan in an embodiment of the present invention;

[0055] Figure 3 Schematic diagram of a third embodiment of a method for formulating an energy storage operation and maintenance plan in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0057] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0058] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1 As shown, a method for formulating an energy storage operation and maintenance plan includes the following steps:

[0059] Step 101: Acquire device operation data from a sensor, perform data preprocessing on the device operation data, and obtain initial device operation data;

[0060] Specifically, in this embodiment: the device operation data in the sensor is obtained, and the device operation data at least includes the device's voltage operation data, current operation data, temperature operation data, charge and discharge times operation data, energy conversion efficiency operation data, temperature data, humidity data, and air quality data; the device operation data is integrated and cleaned to remove abnormal values ​​and noise in the device operation data to obtain complete device operation data; the complete device operation data is normalized to obtain initial device operation data.

[0061] Step 102: Apriori association rule analysis algorithm is used to obtain potential associations between different operating parameters in the initial equipment operating data and between the operating parameters and the faults, and associated equipment operating data is obtained;

[0062] Specifically, in this embodiment: the initial equipment operation data is scanned, the number of occurrences of each single fault type is counted, and a frequent 1-item set that meets the support threshold is determined; based on the frequent 1-item set, a candidate 2-item set is generated through a connection operation, and the initial equipment operation data is scanned again, the number of occurrences of the candidate 2-item set is counted, and a frequent 2-item set that meets the support threshold is determined; the above process is repeated, and higher-order candidate item sets are continuously generated through connection and pruning operations, and frequent item sets that meet the support threshold are screened out until new frequent item sets cannot be generated; for each frequent item set, all possible non-empty subsets thereof are generated, and the confidence between each subset and other items is calculated. If the confidence meets the set threshold, the association relationship between the subset and the other items is output as a valid association rule to obtain the associated equipment operation data.

[0063] Step 103: classify the operating status of the devices in the associated device operating data based on the DBSCAN clustering analysis algorithm to obtain similar device operating data;

[0064] Specifically, in this embodiment: determine the neighborhood radius Eps and the minimum number of samples MinPts of the DBSCAN clustering analysis algorithm; for each point in the associated device operation data, calculate the number of points in its neighborhood according to the neighborhood radius Eps and the minimum number of samples MinPts; starting from any core point that has not been visited, assign all points in its neighborhood to the same cluster, and for the points in these neighborhoods, continue to expand their neighborhoods, and add newly discovered points belonging to the cluster to the cluster until no new points can be added, thereby obtaining an initial cluster; repeat the above process, and for other core points that have not been assigned to any cluster, continue to expand the cluster until all core points have been processed, thereby obtaining similar device operation data.

[0065] Step 104: Establish an RNN-SVM equipment status assessment model based on the RNN recurrent neural network and the SVM support vector machine;

[0066] Specifically, in this embodiment:

[0067] RNN part construction: Determine the structure type of RNN, such as the common long short-term memory network (LSTM) or gated recurrent unit (GRU), which can better capture long-term dependencies when processing time series data. Select the appropriate number of layers and the number of neurons in each layer according to the characteristics of the equipment operation data and the complexity of the problem. Define the input layer, hidden layer, and output layer of the RNN. The number of nodes in the input layer should match the number of features of the equipment operation data, the number of neurons in the hidden layer can be determined by empirical formulas or experimental adjustments, and the number of nodes in the output layer is set according to the number of classifications of the equipment status or the number of evaluation indicators. For example, if the equipment status is to be divided into three types: normal, minor fault, and major fault, the output layer can be set with 3 nodes, corresponding to different status categories. Select a suitable activation function, such as the commonly used tanh function or ReLU function in the hidden layer, and select a suitable activation function in the output layer according to the task requirements, such as the softmax function for classification tasks or the linear activation function for regression tasks. Set other parameters of the RNN, such as time step (if the data is time series data, determine the number of data points contained in each time step), batch size (the number of samples input into the model each time during training), etc.

[0068] SVM part construction: Select the kernel function of SVM, common ones include linear kernel function, polynomial kernel function, radial basis function (RBF) kernel function, etc. For equipment status assessment problems, RBF kernel function usually has good performance, but it is selected experimentally according to specific data. Determine the relevant parameters of SVM, such as penalty coefficient C and kernel function parameter γ. The penalty coefficient C is used to control the model's tolerance to training errors. A larger C value will make the model fit the training data more strictly, but may lead to overfitting; a smaller C value will make the model pay more attention to generalization ability, but may lead to underfitting. The kernel function parameter γ affects the distribution of data in high-dimensional space, and different γ values ​​will have a greater impact on the performance of the model. The optimal combination of C and γ values ​​can be found through grid search, random search, or cross-validation.

[0069] Model combination: Determine how to combine RNN and SVM. Use the output of RNN as the input feature of SVM. After RNN learns and extracts the time series features of the equipment operation data, it obtains a feature vector of fixed length, which is input into SVM to classify or evaluate the equipment status. This combination is achieved by defining appropriate data transfer and processing flows in the code. Establish the connection and interaction mechanism between models to ensure that data can flow and process correctly between RNN and SVM. During the training process, the training results of RNN are promptly passed to SVM, and the parameters of RNN are adjusted according to the feedback of SVM.

[0070] Step 105: input similar equipment operation data into the RNN-SVM equipment status assessment model for training to obtain equipment status assessment results;

[0071] Specifically, in this embodiment: similar equipment operation data are divided to obtain a training data set and a test data set; the training set data is used to train the RNN, the preprocessed equipment operation data is input into the RNN model according to the batch size and time step, and the output result of the model is calculated by forward propagation; the loss function is calculated according to the output result and the actual equipment status label, the gradient of the loss function to the parameters of each layer of the RNN is calculated using the back propagation algorithm, and the parameters are updated through the optimization algorithm; the feature vector obtained by the RNN training is used as the input of the SVM, and the label corresponding to the equipment status is used as the output target of the SVM; after the trained RNN is used to extract features of the validation set data, the extracted features and the corresponding labels are input into the SVM for training to obtain the equipment status evaluation result.

[0072] Step 106: Evaluate the state of the energy storage device according to the device state evaluation result. When the device state indicator approaches a preset threshold, immediately trigger the early warning mechanism and send early warning information to the operation and maintenance personnel.

[0073] Specifically, in this embodiment: corresponding thresholds are set according to each equipment status evaluation indicator, and the thresholds include at least a normal range threshold, a warning threshold and a danger threshold; the normal range threshold defines the reasonable fluctuation range of the indicator under normal operating conditions of the equipment, the warning threshold is the value set when the indicator is close to the boundary where problems occur, and the danger threshold is the limit value of the indicator; when the equipment status indicator approaches the preset warning threshold and danger threshold, the warning mechanism is immediately triggered to send warning information to the operation and maintenance personnel.

[0074] Its beneficial effects are that by obtaining the equipment operation data in the sensor, the equipment operation data is preprocessed to obtain the initial equipment operation data; the Apriori association rule analysis algorithm is used to obtain the potential correlation between different operation parameters and between operation parameters and faults in the initial equipment operation data to obtain the associated equipment operation data; the operation status of the equipment in the associated equipment operation data is classified based on the DBSCAN clustering analysis algorithm to obtain similar equipment operation data; an RNN-SVM equipment state evaluation model is established based on the RNN recurrent neural network and the SVM support vector machine; the similar equipment operation data is input into the RNN-SVM equipment state evaluation model for training to obtain the equipment state evaluation result; the energy storage equipment state is evaluated according to the equipment state evaluation result, and when the equipment state indicator approaches the preset threshold, the early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel. The equipment state can be evaluated more accurately, providing a scientific basis for formulating a reasonable operation and maintenance plan, and effectively improving the pertinence and accuracy of the operation and maintenance plan. It can monitor the operation status of the equipment in real time, predict faults in advance, realize preventive and predictive maintenance, reduce the probability of sudden equipment failures, and improve the reliability and service life of the equipment. It can improve operation and maintenance efficiency, reduce maintenance costs, and achieve economical and sustainable development of energy storage system operation and maintenance.

[0075] See also Figure 2 In a method for formulating an energy storage operation and maintenance plan, the operation status of the equipment in the associated equipment operation data is classified based on the DBSCAN clustering analysis algorithm, and obtaining the similar equipment operation data includes the following steps:

[0076] Step 201, determine the neighborhood radius Eps and the minimum number of samples MinPts of the DBSCAN cluster analysis algorithm;

[0077] Step 201: for each point in the associated device operation data, the number of points in its neighborhood is calculated according to the neighborhood radius Eps and the minimum number of samples MinPts;

[0078] Step 201: Starting from any core point that has not been visited, all points in its neighborhood are assigned to the same cluster. For the points in these neighborhoods, their neighborhoods are continuously expanded, and newly discovered points belonging to the cluster are added to the cluster until no new points can be added, thereby obtaining an initial cluster.

[0079] Step 201 , repeat the above process, and continue to expand the clusters for other core points that are not assigned to any clusters, until all core points are processed and similar device operation data are obtained.

[0080] See also Figure 3 In a method for formulating an energy storage operation and maintenance plan, similar equipment operation data is input into the RNN-SVM equipment status assessment model for training, and obtaining the equipment status assessment result includes the following steps:

[0081] Step 301: Divide the operation data of similar devices to obtain a training data set and a test data set;

[0082] Step 301: Use the training set data to train the RNN, input the pre-processed equipment operation data into the RNN model according to the batch size and time step, and calculate the output result of the model through forward propagation;

[0083] Step 301: Calculate the loss function based on the output result and the actual device status label, use the back propagation algorithm to calculate the gradient of the loss function to the parameters of each layer of the RNN, and update the parameters through the optimization algorithm;

[0084] Step 301: Use the feature vector obtained by RNN training as the input of SVM, and the label corresponding to the device status as the output target of SVM;

[0085] Step 301: After using the trained RNN to extract features from the validation set data, the extracted features and corresponding labels are input into the SVM for training to obtain the equipment status assessment results.

[0086] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for formulating an energy storage operation and maintenance plan, characterized in that: The energy storage operation and maintenance plan formulation method comprises the following steps: Acquire device operation data from a sensor, perform data preprocessing on the device operation data, and obtain initial device operation data; Apriori association rule analysis algorithm is used to obtain potential associations between different operating parameters in the initial equipment operating data and between operating parameters and faults, so as to obtain associated equipment operating data; Classifying the operating status of the devices in the associated device operating data based on the DBSCAN clustering analysis algorithm to obtain similar device operating data; Establish an RNN-SVM equipment status assessment model based on RNN recurrent neural network and SVM support vector machine; Inputting the similar equipment operation data into the RNN-SVM equipment status assessment model for training to obtain equipment status assessment results; The energy storage device status is evaluated according to the device status evaluation result. When the device status indicator approaches a preset threshold, an early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel.

2. A method for formulating an energy storage operation and maintenance plan according to claim 1, characterized in that: The step of acquiring the device operation data from the sensor and performing data preprocessing on the device operation data to obtain initial device operation data includes: Acquire device operation data from the sensor, wherein the device operation data at least includes voltage operation data, current operation data, temperature operation data, charge and discharge times operation data, energy conversion efficiency operation data, temperature data, humidity data, and air quality data of the device; Integrate and clean the equipment operation data, remove abnormal values ​​and noise in the equipment operation data, and obtain complete equipment operation data; The complete equipment operation data is normalized to obtain initial equipment operation data.

3. A method for formulating an energy storage operation and maintenance plan according to claim 1, characterized in that: The Apriori association rule analysis algorithm is used to obtain potential associations between different operating parameters and between operating parameters and faults in the initial equipment operation data to obtain associated equipment operation data, including: Scan the initial equipment operation data, count the number of occurrences of each single fault type, and determine a frequent 1-item set that meets a support threshold; Based on the frequent 1-item set, a candidate 2-item set is generated through a connection operation, and the initial device operation data is scanned again, the number of occurrences of the candidate 2-item set is counted, and a frequent 2-item set that meets a support threshold is determined; Repeat the above process, continuously generate higher-order candidate item sets through connection and pruning operations, and filter out frequent item sets that meet the support threshold until no new frequent item sets can be generated; For each frequent item set, all possible non-empty subsets are generated, and the confidence between each subset and other items is calculated. If the confidence meets the set threshold, the association relationship between the subset and other items is output as a valid association rule to obtain the associated equipment operation data.

4. A method for formulating an energy storage operation and maintenance plan according to claim 1, characterized in that: The DBSCAN clustering analysis algorithm is used to classify the operation status of the devices in the associated device operation data to obtain similar device operation data, including: Determine the neighborhood radius Eps and the minimum number of samples MinPts of the DBSCAN clustering analysis algorithm; For each point in the associated device operation data, the number of points in its neighborhood is calculated according to the neighborhood radius Eps and the minimum number of samples MinPts; Starting from any core point that has not been visited, all points in its neighborhood are assigned to the same cluster. For the points in these neighborhoods, their neighborhoods are continued to be expanded, and newly discovered points belonging to the cluster are added to the cluster until no new points can be added, and the initial cluster is obtained; Repeat the above process, and continue to expand the clusters for other core points that are not assigned to any clusters until all core points are processed and the operation data of similar devices are obtained.

5. A method for formulating an energy storage operation and maintenance plan according to claim 1, characterized in that: The inputting the similar equipment operation data into the RNN-SVM equipment status assessment model for training to obtain equipment status assessment results includes: Dividing the similar equipment operation data to obtain a training data set and a test data set; Use the training set data to train the RNN, input the preprocessed equipment operation data into the RNN model according to the batch size and time step, and calculate the output of the model through forward propagation; Calculate the loss function based on the output results and the actual device status label, use the back propagation algorithm to calculate the gradient of the loss function to the parameters of each layer of the RNN, and update the parameters through the optimization algorithm; The feature vector obtained by RNN training is used as the input of SVM, and the label corresponding to the device status is used as the output target of SVM; After using the trained RNN to extract features from the validation set data, the extracted features and corresponding labels are input into the SVM for training to obtain the equipment status assessment results.

6. A method for formulating an energy storage operation and maintenance plan according to claim 1, characterized in that: The energy storage device status is evaluated according to the device status evaluation result. When the device status indicator approaches a preset threshold, an early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel, including: Set corresponding thresholds according to each equipment status evaluation indicator, wherein the thresholds at least include a normal range threshold, a warning threshold and a danger threshold; The normal range threshold defines the reasonable fluctuation range of the indicator under normal operation of the equipment. The warning threshold is the value set when the indicator is close to the boundary of a problem. The danger threshold is the extreme value of the indicator. When the equipment status indicators approach the preset warning threshold and danger threshold, the warning mechanism is immediately triggered and warning information is sent to the operation and maintenance personnel.

7. A system for implementing a method for formulating an energy storage operation and maintenance plan as claimed in claim 1, characterized in that: The system includes the following modules: A data preprocessing module is used to obtain the device operation data in the sensor, perform data preprocessing on the device operation data, and obtain initial device operation data; An association analysis module is used to obtain potential associations between different operating parameters and between operating parameters and faults in the initial equipment operating data by using an Apriori association rule analysis algorithm to obtain associated equipment operating data; A cluster analysis module, used for classifying the operation status of the equipment in the associated equipment operation data based on the DBSCAN cluster analysis algorithm to obtain similar equipment operation data; A model building module is used to build an RNN-SVM equipment status assessment model based on the RNN recurrent neural network and the SVM support vector machine; A state recognition module is used to input the similar equipment operation data into the RNN-SVM equipment state assessment model for training to obtain equipment state assessment results; The energy storage early warning module is used to evaluate the state of the energy storage device according to the device state evaluation result. When the device state indicator approaches a preset threshold, the early warning mechanism is immediately triggered to send early warning information to the operation and maintenance personnel.

8. A system for implementing a method for formulating an energy storage operation and maintenance plan as claimed in claim 1, characterized in that: The system includes the following submodules: An acquisition submodule is used to acquire device operation data in the sensor, wherein the device operation data at least includes voltage operation data, current operation data, temperature operation data, charge and discharge times operation data, energy conversion efficiency operation data, temperature data, humidity data, and air quality data of the device; A cleaning submodule is used to integrate and clean the equipment operation data, remove abnormal values ​​and noise in the equipment operation data, and obtain complete equipment operation data; The submodule is used to normalize the complete device operation data to obtain initial device operation data.

9. A system for implementing a method for formulating an energy storage operation and maintenance plan as claimed in claim 1, characterized in that: The system includes the following submodules: A scanning submodule, used to scan the initial equipment operation data, count the number of occurrences of each single fault type, and determine a frequent 1-item set that meets a support threshold; A connection submodule, configured to generate candidate 2-item sets through a connection operation based on the frequent 1-item set, and scan the initial device operation data again, count the number of occurrences of the candidate 2-item sets, and determine the frequent 2-item sets that meet the support threshold; The screening submodule is used to repeat the above process, continuously generate higher-order candidate item sets through connection and pruning operations, and screen out frequent item sets that meet the support threshold until no new frequent item sets can be generated; The output submodule is used to generate all possible non-empty subsets of each frequent item set and calculate the confidence between each subset and other items. If the confidence meets the set threshold, the association relationship between the subset and other items is output as a valid association rule to obtain the associated device operation data.

Images