Green electric power intelligent operation and maintenance system and method for low-carbon power station

Abstract

The invention discloses a green electric power intelligent operation and maintenance system and method for a low-carbon power station, and belongs to the technical field of low-carbon operation and maintenance of a transformer substation, and the method specifically comprises the steps: constructing a transformer substation capacitor region digital twin model through three-dimensional laser scanning, deploying a sensor to collect operation data in real time, and transmitting the operation data to a centralized monitoring center after edge calculation preprocessing; extracting data features by using a deep learning algorithm, analyzing and predicting a fault probability and residual life by combining with a time sequence, linking a prediction result with a transformer substation capacitor area digital twinborn model, and marking a fault risk area on a three-dimensional interface; a panoramic visual interface is developed based on the three-dimensional model, a dynamic three-level threshold value is generated in real time, a reinforcement learning algorithm is started to generate an optimal capacitor switching strategy when early warning is triggered, and switching operation is executed in combination with security constraints; the operation safety of the power system is improved, and the operation and maintenance cost and carbon emission are reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-carbon operation and maintenance of substations, and specifically relates to a low-carbon power station green power intelligent operation and maintenance system and method. Background Art

[0002] There are problems in the traditional capacitor management method, such as incomplete information acquisition, untimely processing, and complex operations, which are difficult to meet the requirements of modern power grids for capacitor monitoring. Therefore, an efficient and intelligent digital analysis method is needed to achieve panoramic visualization monitoring of substation capacitors. This method is based on modern information technologies such as the Internet of Things, big data, and cloud computing. By collecting the operation data of capacitors, real-time processing and analysis are carried out to generate a panoramic visualization display effect, helping managers comprehensively and accurately understand the operation status of capacitors, timely discover potential problems, and improve management efficiency.

[0003] For example, Chinese Patent with the publication number CN117638862A discloses a low-carbon scheduling method for an integrated energy system based on an oxy-fuel combustion-thermal power station, including: introducing an oxy-fuel combustion unit and a thermal power station into the integrated energy system, respectively analyzing the operation characteristics of the oxy-fuel combustion unit and the thermal power station and modeling them; analyzing the operation characteristics of two-stage power-to-gas and mathematically modeling the two-stage power-to-gas model; establishing a collaborative operation framework for the oxy-fuel combustion unit - thermal power station, studying the mechanism of their combined operation and the internal operation of the integrated energy system; constructing balance equations for four energy flows of electricity, heat, gas, and hydrogen and a stepped carbon trading model, establishing a low-carbon economic scheduling model with the minimum sum of carbon trading costs, unit operation and gas purchase costs, carbon sequestration costs, operation and maintenance and abandoned wind costs as the objective function, and calling the Gurobi solver to solve and analyze different set scenarios; this technical solution can effectively improve the operation efficiency of the energy system, reduce carbon emissions, and have good operation economy.

[0004] The above existing technologies all have the following problems: lack of refined ability for equipment-level operation and maintenance; rely on offline modeling and optimization solving, the scheduling strategy lags behind the actual operation condition changes, and lack of real-time data-driven closed-loop control; do not integrate digital twin and artificial intelligence technologies, and have weak adaptability to complex nonlinear systems. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention proposes a low-carbon power station green power intelligent operation and maintenance method. A digital twin model of the substation capacitor area is constructed through 3D laser scanning, sensors are deployed to collect operation data in real time, and after preprocessing by edge computing, the data is transmitted to the centralized monitoring center. Deep learning algorithms are used to extract data features, combined with time series analysis to predict the failure probability and remaining life. The prediction results are linked with the digital twin model of the substation capacitor area, and the fault risk area is marked on the 3D interface. A panoramic visualization interface is developed based on the 3D model, a dynamic three-level threshold is generated in real time, and when an alarm is triggered, a reinforcement learning algorithm is started to generate the optimal capacitor switching strategy, and the switching operation is executed in combination with safety constraints. The present invention improves the operation safety of the power system and reduces the operation and maintenance costs and carbon emissions.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A low-carbon power station green power intelligent operation and maintenance method, including:

[0008] Step S1: Establish a digital twin model of the substation capacitor area, collect the operation data of the substation capacitor in real time, and transmit the processed operation data of the substation capacitor to the centralized monitoring center;

[0009] Step S2: In the centralized monitoring center, use deep learning strategies to extract features from the received operation data of the substation capacitor, combine time series analysis strategies to predict the failure probability and remaining life, link the prediction results with the digital twin model of the substation capacitor area, and mark the fault risk area on the 3D interface;

[0010] Step S3: Build a panoramic visualization interface of the substation capacitor based on the 3D model of the substation capacitor area, generate equipment operation and maintenance decision support information through real-time analysis, and establish a data sharing mechanism; the equipment operation and maintenance decision support information includes equipment status assessment reports, fault warning prompts, and maintenance suggestions;

[0011] Step S4: Set a three-level threshold. When any parameter triggers an alarm, start a reinforcement learning algorithm to generate a switching strategy, calculate the optimal switching combination, and generate and execute a switching instruction according to the optimal switching combination.

[0012] Specifically, in the step S1 of establishing a digital twin model of the substation capacitor area and collecting the operation data of the substation capacitor in real time, it includes:

[0013] S1.1: Use a 3D laser scanner to collect point cloud data of the substation capacitor area. Combine the imported electrical CAD drawings, integrate the geometric parameters of the equipment, electrical connection relationships, and sensor position information to generate a digital twin model of the substation capacitor area; the equipment geometric parameters include dimensions and installation positions; the electrical connection relationships include capacitance values and rated voltages.

[0014] S1.2: Bind the digital twin model of the substation capacitor area to the unique identifier of the physical device through the Internet of Things platform to establish a virtual-entity two-way mapping relationship.

[0015] S1.3: Based on the digital twin model of the substation capacitor area, plan the sensor positions, collect the operation data of the substation capacitors in real time through the deployed sensor network, and use edge computing devices to preprocess the operation data of the substation capacitors; the operation data of the substation capacitors includes capacitor current, voltage, temperature, humidity, and partial discharge amount parameters.

[0016] S1.4: Perform hash certification on the preprocessed operation data of the substation capacitors through blockchain nodes and encrypt and transmit it to the centralized monitoring center.

[0017] Specifically, the process of using edge computing devices to preprocess the operation data of the substation capacitors in S1.3 includes:

[0018] S1.31: Set the 3D image data obtained by the 3D laser scanner as , convert the 3D image data to the 3D frequency domain to obtain 3D frequency domain coefficients , where represents the number of pixel values in each dimension of the 3D image data, and represent the first-dimension frequency domain, second-dimension frequency domain, and third-dimension frequency domain indices in sequence.

[0019] S1.32: Through a 3D quantization table, perform quantization analysis on the obtained 3D frequency domain coefficients to obtain the quantized 3D frequency domain coefficients ;

[0020] S1.33: Perform zero coefficient processing on the quantized 3D frequency domain coefficients , perform entropy coding on the non-zero coefficients to obtain the final compressed data, and perform an inverse coding operation at the decoding end to restore the original image data to obtain the preprocessed operation data of the substation capacitors , where i represents the index value of the preprocessed operation data of the substation capacitors.

[0021] Specifically, the specific steps of step S2 include:

[0022] S2.1: Based on the processed operation data of substation capacitors , construct a multi-dimensional feature vector , and use a long short-term memory network to perform double-branch prediction on the multi-dimensional feature vector to obtain the fault probability prediction result and the remaining life prediction result , where s represents the index value of the multi-dimensional feature vector, and t represents the current moment;

[0023] S2.2: Calculate the risk value according to the fault probability prediction result and the remaining life prediction result , and divide the risk level according to the risk value ;

[0024] S2.3: Based on the fault probability prediction result , the remaining life prediction result and the divided risk level , use the formula to map the risk level of the feature space to the three-dimensional physical space to obtain the mapping result of the risk level of the feature space , where represents the physical space coordinate index, represents the interpolation function based on inverse distance weighting;

[0025] S2.4: According to the mapping result of the risk level of the feature space, mark the fault risk area on the digital twin model of the substation capacitor area, and perform color mapping on different risk levels to generate a digital twin model file of the substation capacitor area with risk markings.

[0026] Specifically, the specific steps of step S4 include:

[0027] S4.1: Set three-level thresholds, and dynamically adjust the three-level thresholds in combination with the real-time load rate; the three-level thresholds include voltage threshold, temperature threshold, and life threshold;

[0028] S4.2: Obtain the operation data of substation capacitors, and set the state space and action space according to the operation data of substation capacitors;

[0029] The state space is composed of the operation data of substation capacitors;

[0030] The action space is the switching state, including cut-off or input;

[0031] S4.3: Input the operation data of substation capacitors into the state space after normalization processing;

[0032] S4.4: If any parameter in the substation capacitor operation data is greater than the dynamically adjusted three-level threshold, the switching probability of each capacitor group is calculated and the value of the state-action pair is evaluated;

[0033] S4.5: Based on the value evaluation result of the state-action pair and the switching state, the optimal switching combination is obtained through mixed integer programming, and the timing switching instruction sequence is generated in combination with the safety constraints, and the timing switching instruction sequence is issued through the system.

[0034] Specifically, the specific steps of obtaining the optimal switching combination in S4.5 include:

[0035] S4.5.1: Set the reactive power of the system to , the reactive power provided by the current capacitor bank is , calculate the power factor of the system ,in, Indicates the current voltage phase, Indicates the current phase;

[0036] S4.5.2: Calculate the reactive power of the system The reactive power provided by the capacitor bank The difference between ;

[0037] S4.5.3: Based on the reactive power compensation demand, combined with the performance parameters of the capacitors and the voltage conditions of the system, screen the capacitors, optimize the switching combination, and obtain the optimal switching combination.

[0038] A low-carbon power station green power intelligent operation and maintenance system, including: a data acquisition module, a fault prediction module, an operation and maintenance decision module, and an intelligent switching control module;

[0039] The data acquisition module is used to build a digital twin model of the substation capacitor area and collect and process the substation capacitor operation data in real time;

[0040] The fault prediction module is used to extract features and predict faults from the substation capacitor operation data using deep learning strategies and time series analysis strategies, and to link the substation capacitor regional digital twin model to achieve fault risk regional early warning;

[0041] The operation and maintenance decision module is used to build a panoramic visualization interface, generate equipment operation and maintenance decision support information, and establish a data sharing mechanism;

[0042] The intelligent switching control module is used to perform intelligent switching control on the capacitor bank by setting a threshold and a reinforcement learning algorithm.

[0043] Specifically, the fault prediction module includes: a feature extraction unit, a time series analysis unit, and a digital twin linkage unit;

[0044] The feature extraction unit is used to extract features from the operation data of substation capacitors using deep learning algorithms;

[0045] The time series analysis unit is used to analyze the temporal correlation of data based on time series analysis strategies, and predict the fault probability and remaining life of capacitors;

[0046] The digital twin linkage unit is used to link the prediction results with the digital twin model of the substation capacitor area, and real-time mark the fault risk area in the 3D interface.

[0047] Specifically, the intelligent switching control module includes: a threshold setting unit, a reinforcement learning algorithm unit, and a switching instruction execution unit;

[0048] The threshold setting unit is used to set three-level thresholds and automatically adjust the threshold range according to the real-time load rate of the power grid;

[0049] The reinforcement learning algorithm unit is used to automatically generate switching strategies using reinforcement learning algorithms when parameters trigger warnings, calculate the switching combinations of capacitor banks, and determine the number and sequence of capacitors to be put into or cut out;

[0050] The switching instruction execution unit is used to generate switching instructions according to the optimal switching combination and control the on-site equipment to execute.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] 1. The present invention proposes a low-carbon power station green power intelligent operation and maintenance system, and has optimized and improved the architecture, operation steps and processes. The system has the advantages of simple processes, low investment and operation costs, and low production work costs, and realizes efficient and intelligent capacitor management.

[0053] 2. The present invention proposes a low-carbon power station green power intelligent operation and maintenance method. Through the integration of digital twin technology and multi-source data, a virtual mapping model of the substation capacitor area is constructed, realizing the real-time collection and edge computing preprocessing of operation data, ensuring the timeliness and accuracy of data; using deep learning and time series analysis algorithms, it can accurately extract equipment features and predict fault probability and remaining life, and dynamically mark the fault risk area in the 3D interface in combination with the digital twin model of the substation capacitor area, providing an intuitive view of the equipment health status for operation and maintenance personnel, early warning of potential faults, changing passive maintenance to active prevention, and improving the timeliness of fault handling and operation and maintenance efficiency.

[0054] 3. The present invention proposes a method for intelligent operation and maintenance of green power in a low-carbon power station. By establishing a panoramic visualization interface and a data sharing mechanism, this method realizes the real-time generation and cross-departmental sharing of equipment status evaluation reports, fault early warning prompts, and maintenance suggestions, and optimizes the operation and maintenance decision-making process. By setting three-level dynamic thresholds and combining with a reinforcement learning algorithm, it can automatically trigger the optimization of switching strategies according to real-time operation data, dynamically generate the optimal switching combination, ensure the voltage stability of the power system and the compliance of the power factor, balance the equipment life, reduce energy consumption, and improve the safety and stability of system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a flowchart of the steps of a method for intelligent operation and maintenance of green power in a low-carbon power station according to the present invention;

[0056] Figure 2 is a flowchart of the algorithm of a method for intelligent operation and maintenance of green power in a low-carbon power station according to the present invention;

[0057] Figure 3 is a flowchart of the preprocessing of the operation data of the substation capacitor in a method for intelligent operation and maintenance of green power in a low-carbon power station according to the present invention;

[0058] Figure 4 is a flowchart of the switching strategy of a method for intelligent operation and maintenance of green power in a low-carbon power station according to the present invention;

[0059] Figure 5 is an architecture diagram of a system for intelligent operation and maintenance of green power in a low-carbon power station according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] Example 1:

[0061] Please refer to Figures 1-4 , an embodiment provided by the present invention: A method for intelligent operation and maintenance of green power in a low-carbon power station includes the following steps:

[0062] Step S1: Establish a digital twin model for the substation capacitor area, collect the operation data of the substation capacitor in real time, and transmit the processed operation data of the substation capacitor to the centralized monitoring center;

[0063] Step S2: In the centralized monitoring center, use a deep learning strategy to extract features from the received operation data of the substation capacitor, combine with a time series analysis strategy to predict the fault probability and remaining life, link the prediction results with the digital twin model of the substation capacitor area, and mark the fault risk area in the three-dimensional interface;

[0064] Step S3: Based on the 3D model of the substation capacitor area, construct a panoramic visualization interface for the substation capacitors, generate decision support information for equipment operation and maintenance through real-time analysis, and establish a data sharing mechanism; the equipment operation and maintenance decision support information includes equipment status assessment reports, fault warning prompts, and maintenance suggestions.

[0065] The present invention uses a panoramic visualization interface for substation capacitors developed based on a 3D model, making the display of data more intuitive and easy to understand. Operation and maintenance personnel can view information such as the operation status and fault warnings of capacitors in real time through the interface, improving work efficiency.

[0066] Step S4: Set three-level thresholds. When any parameter triggers an alarm, start the reinforcement learning algorithm to generate switching strategies, calculate the optimal switching combinations, and generate and execute switching instructions according to the optimal switching combinations.

[0067] The switching strategy refers to automatically or manually determining when to connect or disconnect capacitors according to the real-time operating status and requirements of the power system, in order to optimize system operation, improve the power factor, and reduce power losses. Connecting capacitors can increase the reactive power of the system and improve the voltage level, while disconnecting capacitors can avoid overvoltage or reactive power surplus situations.

[0068] In the above-mentioned step S1, a digital twin model of the substation capacitor area is established, and the operation data of the substation capacitors is collected in real time, including:

[0069] S1.1: Use a 3D laser scanner to collect point cloud data of the substation capacitor area, combine with the imported electrical CAD drawings, integrate equipment geometric parameters, electrical connection relationships, and sensor position information to generate a digital twin model of the substation capacitor area; the equipment geometric parameters include dimensions and installation positions; the electrical connection relationships include capacitance values and rated voltages.

[0070] S1.2: Bind the digital twin model of the substation capacitor area to the unique identifier of the physical device through the Internet of Things platform to establish a virtual-entity two-way mapping relationship.

[0071] S1.3: Plan the sensor positions based on the digital twin model of the substation capacitor area, collect the operation data of the substation capacitors in real time through the deployed sensor network, and preprocess the operation data of the substation capacitors using edge computing devices; the operation data of the substation capacitors includes capacitor current, voltage, temperature, humidity, and partial discharge amount parameters.

[0072] S1.4: Perform hash certification on the preprocessed operation data of the substation capacitors through blockchain nodes and encrypt and transmit it to the centralized monitoring center.

[0073] The process of preprocessing the operation data of substation capacitors by using edge computing devices in S1.3 includes:

[0074] S1.31: Set the 3D image data obtained by the 3D laser scanner as , and convert the 3D image data into the 3D frequency domain to obtain the 3D frequency domain coefficients . The formula is:

[0075] ;

[0076] Among them, represents the number of pixel values in each dimension of the 3D image data, represent the first - dimension frequency domain, second - dimension frequency domain, and third - dimension frequency domain indices in sequence, represent the first - dimension frequency, second - dimension frequency, and third - dimension frequency weighting coefficients in sequence, represent the first - dimension frequency, second - dimension frequency, and third - dimension frequency index coefficients in sequence, represents the pixel point position in each dimension of the 3D image data, represents in the 3D image data the pixel value at the position, represents the window function, represents the product;

[0077] S1.32: Through the 3D quantization table, perform quantization analysis on the obtained 3D frequency domain coefficients to obtain the quantized 3D frequency domain coefficients . Among them, represents the quantization step corresponding to the frequency domain index, represents rounding;

[0078] Further, the selection of the quantization step: The quantization step is a key factor affecting the compression ratio and the quality of the reconstructed image. A larger quantization step will result in a higher compression ratio but will also introduce more distortion; a smaller quantization step can retain more image details but the compression ratio will be lower.

[0079] S1.33: Perform zero - coefficient processing on the quantized 3D frequency domain coefficients , and perform entropy coding on the non - zero coefficients to obtain the final compressed data. At the same time, perform an inverse coding operation at the decoding end to restore the original image data to obtain the processed operation data of substation capacitors , where i represents the index value of the processed operation data of substation capacitors.

[0080] Furthermore, the reasons for zero coefficient processing: The quantization step usually causes many high-frequency coefficients to become zero or values close to zero. These zero coefficients or coefficients close to zero can be specially processed during encoding to reduce the data size.

[0081] Entropy coding is a lossless compression method that assigns different lengths of codewords according to the statistical characteristics of the data. After entropy coding, the final compressed data is obtained, and these data are output to a file or memory in the form of a binary stream.

[0082] The specific steps of step S2 include:

[0083] S2.1: Based on the processed operation data of the substation capacitor , construct a multi-dimensional feature vector , and use a long short-term memory network to perform double-branch prediction on the multi-dimensional feature vector to obtain the fault probability prediction result and the remaining life prediction result , where s represents the index value of the multi-dimensional feature vector, and t represents the current moment;

[0084] It should be noted that the relationship between the processed operation data of the substation capacitor and the time series analysis strategy: The operation data processed through S1.3 is the input of the time series analysis strategy. These input data contain the key information of the capacitor operation and are the basis for the model to make predictions; based on the input data, the time series model is trained to learn the operation rules and potential fault modes of the capacitor. Once the model training is completed, the model is used to predict future data, including the life state and potential fault points of the capacitor.

[0085] Furthermore, the architecture of the double-branch long short-term memory network is:

[0086] Health status branch: Predict the remaining life , where represents the hidden state sequence from the historical moment T to the current moment t, represents the long short-term memory network model for remaining life prediction, which belongs to the time series prediction model in deep learning;

[0087] Among them, the historical moment is any starting moment in the historical sequence and satisfies: .

[0088] Fault probability branch: Predict the probabilities of multiple types of faults , where represents the normalized exponential function, It represents a long short-term memory network model for fault classification, sharing part of the underlying network with the health status branch, such as the input layer and the initial long short-term memory network layer, but the top layer network is independent.

[0089] S2.2: According to the fault probability prediction result and the remaining life prediction result Calculate the risk value and divide the risk level according to the risk value ;

[0090] Furthermore, according to the fault probability prediction result and the remaining life prediction result Calculate the risk value, including:

[0091] (1) Set the risk prediction model , where represents the initial risk value at time t, represents the weight coefficient. In the present invention, , represents the designed life of the device. The capacitor in the present invention is 10 years, represents the maximum value function;

[0092] (2) Adjust the initial risk value according to the importance of the device in the power grid to obtain the preliminarily adjusted risk value ;

[0093] (3) Perform a secondary correction on the preliminarily adjusted risk value according to the environmental stress to obtain the final risk value , where represents the environmental stress normalization value, represents the environmental sensitivity coefficient. In the present invention, if it is a high-temperature environment, take , if it is a humid environment, take .

[0094] S2.3: Based on the fault probability prediction result , the remaining life prediction result and the divided risk level , use the formula to map the risk level of the feature space to the three-dimensional physical space to obtain the mapping result of the risk level of the feature space , where represents the physical space coordinate index, represents the interpolation function based on inverse distance weighting;

[0095] S2.4: According to the mapping result of the feature space risk level, mark the fault risk area on the digital twin model of the substation capacitor area, and perform color mapping for different risk levels to generate a digital twin model file of the substation capacitor area with risk markings.

[0096] Further, the process of dividing and performing color mapping for different risk levels includes:

[0097] If and , then the risk level is the safe level, that is , and it is mapped to green, where RUL represents the remaining useful life, which can also be called the Remaining Useful Life (RUL), and the unit is days;

[0098] If or , then the risk level is the attention level, that is , and it is mapped to yellow;

[0099] If or , then the risk level is the warning level, that is , and it is mapped to orange;

[0100] If and , then the risk level is the emergency level, that is , and it is mapped to red.

[0101] The specific steps of step S4 include:

[0102] S4.1: Set three-level thresholds, and dynamically adjust the three-level thresholds in combination with the real-time load rate; the three-level thresholds include voltage threshold, temperature threshold, and life threshold;

[0103] Further, the three-level thresholds include:

[0104] Voltage threshold (taking the rated voltage as the reference): Normal range: [0.95 , 1.05 ;

[0105] First-level warning: ;

[0106] Second-level warning: ;

[0107] Emergency threshold: less than 0.85 or greater than 1.15 ;

[0108] Temperature threshold (based on the maximum allowable temperature of the device ):

[0109] Normal range: less than or equal to 0.7 ;

[0110] First-level warning: ;

[0111] Second-level warning: ;

[0112] Emergency threshold: greater than 0.95 ;

[0113] Lifetime threshold (remaining useful life RUL):

[0114] Normal range: ;

[0115] First-level warning: ;

[0116] Second-level warning: ;

[0117] Emergency threshold: .

[0118] Furthermore, by calculating the ratio of the current active power to the maximum active power, the real-time load rate is obtained; the dynamic adjustment of the three-level threshold is achieved by multiplying the base threshold by the real-time load rate. Exemplarily, the voltage threshold adjustment formula is: , where represents the adjusted voltage threshold, represents the voltage base threshold, and satisfies , represents the adjustment coefficient, and , represents the real-time load rate.

[0119] S4.2: Obtain the operation data of the substation capacitor, and set the state space and action space according to the operation data of the substation capacitor;

[0120] The state space consists of the operation data of the substation capacitor;

[0121] The action space is the switching state, including cutting or putting in;

[0122] S4.3: Input the operation data of the substation capacitor into the state space after normalization;

[0123] S4.4: If any parameter in the operation data of the substation capacitor is greater than the dynamically adjusted three-level threshold, calculate the switching probability of each capacitor bank and evaluate the value of the state-action pair;

[0124] S4.5: Based on the value evaluation result of the state-action pair, combined with the switching state, obtain the optimal switching combination through mixed integer programming, considering the safety constraint conditions , generate a time-series switching instruction sequence, and issue the time-series switching instruction sequence through the system, where represents the maximized power factor, P is the power factor, m represents the minimized number of switchings, represents the switching state of capacitor r, 0 means not put in, 1 means put in, n represents the number of capacitor banks, and the mixed integer programming is the prior art content in the field and is not the creative solution of this application, so it will not be elaborated here.

[0125] Specifically, the specific steps for obtaining the optimal switching combination in S4.5 include:

[0126] S4.5.1: Set the reactive power of the system as , the reactive power provided by the current capacitor bank as , and calculate the power factor of the system , where represents the current voltage phase, represents the current current phase;

[0127] S4.5.2: By calculating the difference between the reactive power of the system and the reactive power provided by the current capacitor bank, obtain the current reactive power compensation demand ;

[0128] S4.5.3: Based on the reactive power compensation demand, combined with the performance parameters of the capacitors and the voltage conditions of the system, screen the capacitors and optimize the switching combination to obtain the optimal switching combination.

[0129] The purpose of optimizing the switching combination is to consider the number of switchings of the capacitor bank, the life balance of the capacitor bank, and the voltage stability factor of the system while meeting the reactive power compensation demand.

[0130] Judging whether it is necessary to adjust the switching state of the capacitor is usually based on the following considerations:

[0131] (1) Comparison of real-time monitoring data with thresholds: Compare the real-time monitored operation data of the substation capacitors with the preset thresholds, involving multiple aspects of data, such as voltage, current, power factor, and reactive power. If these data exceed the preset threshold range, it means that the capacitors need to adjust the switching state to maintain the stable operation of the system;

[0132] (2) System stability analysis: Analyze the stability of the current system, including the overall operating status of the system, the load conditions, and the harmonic content factors. If the system shows instability or abnormal fluctuations, it is necessary to adjust the switching state of the capacitors to improve the system stability;

[0133] (3) Power factor consideration: By analyzing the real-time monitored power factor data, determine whether it is necessary to switch the capacitors to increase or decrease the power factor, so as to achieve the purpose of energy conservation and improving the power quality;

[0134] (4) Historical data and trend analysis: Combine the historical operation data and the trend analysis of the current data to predict the future load changes and possible system requirements. If it is predicted that there will be load peaks or troughs, it is necessary to adjust the switching state of the capacitors in advance to cope with these changes;

[0135] (5) Fault warning and diagnosis: If the system detects a fault warning of the capacitors or related equipment, or discovers potential problems through fault diagnosis, it is necessary to adjust the switching state of the capacitors to avoid the occurrence of faults or reduce the impact of faults;

[0136] (6) Economic consideration: To save the electric energy cost and reduce the equipment loss.

[0137] Embodiment 2:

[0138] Please refer to Figure 5 , another embodiment provided by the present invention: A low-carbon power station green power intelligent operation and maintenance system, including:

[0139] A data acquisition module, a fault prediction module, an operation and maintenance decision-making module, and an intelligent switching control module;

[0140] The data acquisition module is used to build a digital twin model of the substation capacitor area and collect and process the operation data of the substation capacitors in real time;

[0141] The fault prediction module is used to extract features and predict faults for the operation data of the substation capacitors by using deep learning strategies and time series analysis strategies, and link the digital twin model of the substation capacitor area to realize the early warning of the fault risk area;

[0142] The operation and maintenance decision-making module is used to build a panoramic visualization interface, generate equipment operation and maintenance decision support information, establish a data sharing mechanism, and improve the operation and maintenance efficiency and collaboration ability;

[0143] The intelligent switching control module is used to perform intelligent switching control on the capacitor bank by setting thresholds and reinforcement learning algorithms, optimize the operation efficiency of the power system, and reduce energy consumption.

[0144] The data acquisition module includes: a digital twin modeling unit, a data acquisition unit, and a data processing unit;

[0145] A digital twin modeling unit is used to establish a digital twin model of the substation capacitor area, realize the virtual mapping of physical devices, and dynamically display the device status and spatial distribution;

[0146] A data acquisition unit is used to collect the operation data of substation capacitors in real time through devices such as sensors;

[0147] A data processing unit is used to perform preprocessing such as cleaning, noise reduction, and format conversion on the collected operation data of substation capacitors, and transmit the processed operation data of substation capacitors to the centralized monitoring center.

[0148] The fault prediction module includes: a feature extraction unit, a time series analysis unit, and a digital twin linkage unit;

[0149] The feature extraction unit is used to extract features from the operation data of substation capacitors using deep learning algorithms;

[0150] The time series analysis unit is used to analyze the temporal correlation of data based on the time series analysis strategy, and predict the fault probability and remaining life of the capacitor;

[0151] The digital twin linkage unit is used to link the prediction result with the digital twin model of the substation capacitor area, and real-time mark the fault risk area in the 3D interface to assist the operation and maintenance personnel to quickly locate the problem.

[0152] The operation and maintenance decision-making module includes: a visualization unit and a decision information generation unit;

[0153] The visualization unit is used to construct a panoramic 3D visualization interface of the substation capacitor based on the digital twin model of the substation capacitor area, and intuitively display the device layout, operation status, and warning information;

[0154] The decision information generation unit is used to analyze and generate device operation and maintenance decision support information in real time.

[0155] The intelligent switching control module includes: a threshold setting unit, a reinforcement learning algorithm unit, and a switching instruction execution unit;

[0156] The threshold setting unit is used to set three-level thresholds and automatically adjust the threshold range according to the real-time grid load rate;

[0157] The reinforcement learning algorithm unit is used to automatically generate a switching strategy using the reinforcement learning algorithm when the parameters trigger an alarm, calculate the switching combination of the capacitor bank, and determine the number and sequence of capacitors to be put in or cut out;

[0158] The switching instruction execution unit is used to generate switching instructions according to the optimal switching combination and control on-site equipment to execute, such as closing or disconnecting switches, so as to achieve dynamic reactive power regulation and improve the energy efficiency of the power grid.

[0159] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative rather than restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make changes, modifications, substitutions, and variations to the above embodiments without departing from the spirit and scope protected by the present invention and claims. All of these fall within the protection scope of the present invention.

Claims

1. A method for intelligent operation and maintenance of green power in a low-carbon power station, characterized in that, Including: Step S1: Establish a digital twin model for the substation capacitor area, collect the operation data of the substation capacitors in real time, process the operation data of the substation capacitors and transmit it to the centralized monitoring center; Step S2: At the centralized monitoring center, use the deep learning strategy to extract features from the received operation data of the substation capacitors, combine the time series analysis strategy to predict the failure probability and remaining life, link the prediction results with the digital twin model of the substation capacitor area, and mark the failure risk area on the 3D interface; Step S3: Build a panoramic visualization interface for the substation capacitors based on the 3D model of the substation capacitor area, generate equipment operation and maintenance decision support information through real-time analysis, and establish a data sharing mechanism; the equipment operation and maintenance decision support information includes equipment status assessment reports, failure warning prompts and maintenance suggestions; Step S4: Set three-level thresholds. When any parameter triggers an alarm, start the reinforcement learning algorithm to generate switching strategies, calculate the optimal switching combination, and generate and execute switching instructions according to the optimal switching combination.

2. The green power intelligent operation and maintenance method for a low-carbon power station according to claim 1, wherein In step S1, establishing a digital twin model for the substation capacitor area and collecting the operation data of the substation capacitors in real time includes: S1.1: Use a 3D laser scanner to collect point cloud data of the substation capacitor area, combine the imported electrical CAD drawings, integrate the equipment geometric parameters, electrical connection relationships and sensor position information to generate a digital twin model of the substation capacitor area; the equipment geometric parameters include dimensions and installation positions; the electrical connection relationships include capacitance values and rated voltages; S1.2: Bind the digital twin model of the substation capacitor area to the unique identifier of the physical device through the Internet of Things platform to establish a virtual-entity bidirectional mapping relationship; S1.3: Plan the sensor positions based on the digital twin model of the substation capacitor area, collect the operation data of the substation capacitors in real time through the deployed sensor network, and preprocess the operation data of the substation capacitors using edge computing devices; the operation data of the substation capacitors includes capacitor current, voltage, temperature, humidity, partial discharge amount parameters; S1.4: Perform hash certification on the preprocessed operation data of the substation capacitors through blockchain nodes and encrypt and transmit it to the centralized monitoring center.

3. The low-carbon power station green power intelligent operation and maintenance method according to claim 2, characterized in that, The process of preprocessing the operation data of the substation capacitors using edge computing devices in S1.3 includes: S1.31: Set the 3D image data obtained by the 3D laser scanner as , and convert the 3D image data into the 3D frequency domain to obtain 3D frequency domain coefficients , where represents the number of pixel values in each dimension of the 3D image data, successively represent the first dimension frequency domain, the second dimension frequency domain, and the third dimension frequency domain indices; S1.32: Through a three-dimensional quantization table, perform quantization analysis on the obtained three-dimensional frequency domain coefficients to obtain the quantized three-dimensional frequency domain coefficients ; S1.33: Quantize the three-dimensional frequency domain coefficients Perform zero coefficient processing on them, and perform entropy coding on non-zero coefficients to obtain the final compressed data. Then perform an inverse coding operation at the decoding end to restore the original image data, and obtain the processed operation data of the substation capacitor , where i represents the index value of the processed operation data of the substation capacitor 4. A low-carbon power station green power intelligent operation and maintenance method according to claim 3, characterized in that The specific steps of step S2 include: S2.1: Based on the processed operation data of substation capacitors , construct a multi-dimensional feature vector , and use a long short-term memory network to perform double-branch prediction on the multi-dimensional feature vector to obtain the fault probability prediction result and the remaining life prediction result , where s represents the index value of the multi-dimensional feature vector, and t represents the current moment; S2.2: According to the predicted results of the failure probability and the predicted results of the remaining life calculate the risk value, and divide the risk level according to the risk value ; S2.3: Based on the fault probability prediction result , the remaining life prediction result and the divided risk level , use the formula to map the risk level of the feature space to the three-dimensional physical space and obtain the mapping result of the risk level of the feature space , where represents the physical space coordinate index represents the interpolation function based on the inverse distance weight; S2.4: According to the feature space risk level mapping result, mark the failure risk area on the digital twin model of the substation capacitor area, and perform color mapping on different risk levels to generate a digital twin model file of the substation capacitor area with risk markings.

5. The method for intelligent operation and maintenance of green electricity in a low-carbon power station according to claim 4, characterized in that, The specific steps of step S4 include: S4.1: Set three-level thresholds and dynamically adjust the three-level thresholds in combination with the real-time load rate; the three-level thresholds include voltage thresholds, temperature thresholds, and life thresholds; S4.2: Obtain the operation data of the substation capacitors and set the state space and action space according to the operation data of the substation capacitors; The state space is composed of the operation data of the substation capacitors; The action space is the switching state, including disconnection or connection; S4.3: Input the normalized operation data of substation capacitors into the state space; S4.4: If any parameter in the operation data of substation capacitors is greater than the three-level threshold after dynamic adjustment, calculate the switching probability of each capacitor bank and evaluate the value of the state-action pair; S4.5: According to the value evaluation result of the state-action pair, combined with the switching state, obtain the optimal switching combination through mixed integer programming, combine the security constraint conditions, generate a time-series switching instruction sequence, and issue the time-series switching instruction sequence through the system.

6. The green power intelligent operation and maintenance method for a low-carbon power station according to claim 5, characterized in that, The specific steps for obtaining the optimal switching combination in S4.5 include: S4.5.1: Set the reactive power of the system to , the reactive power provided by the current capacitor bank is , calculate the power factor of the system , where represents the current voltage phase, represents the current current phase; S4.5.2: Calculate the reactive power of the system and the reactive power provided by the current capacitor bank to obtain the current reactive power compensation requirement ; S4.5.3: Based on the reactive power compensation demand, combined with the performance parameters of the capacitors and the voltage conditions of the system, screen the capacitors, optimize the switching combination, and obtain the optimal switching combination.

7. A low-carbon power station green electricity intelligent operation and maintenance system, which is used to implement the low-carbon power station green electricity intelligent operation and maintenance method described in any one of claims 1-6, and is characterized in that, Including: Data acquisition module, fault prediction module, operation and maintenance decision-making module, intelligent switching control module; The data acquisition module is used to construct a digital twin model of the substation capacitor area and collect and process the operation data of substation capacitors in real time; The fault prediction module is used to extract features and predict faults from the operation data of substation capacitors using deep learning strategies and time series analysis strategies, and link the digital twin model of the substation capacitor area to realize early warning of fault risk areas; The operation and maintenance decision-making module is used to construct a panoramic visualization interface, generate equipment operation and maintenance decision support information, and establish a data sharing mechanism; The intelligent switching control module is used to perform intelligent switching control on the capacitor bank by setting thresholds and reinforcement learning algorithms.

8. A low-carbon power station green power intelligent operation and maintenance system according to claim 7, characterized in that, The fault prediction module includes: feature extraction unit, time series analysis unit, digital twin linkage unit; The feature extraction unit is used to extract features from the operation data of substation capacitors using deep learning algorithms; The time series analysis unit is used to analyze the time series correlation of data based on time series analysis strategies, and predict the fault probability and remaining life of the capacitors; The digital twin linkage unit is used to link the prediction results with the digital twin model of the substation capacitor area and mark the fault risk area in real time on the three-dimensional interface.

9. A low-carbon power station green power intelligent operation and maintenance system according to claim 8, characterized in that, The intelligent switching control module includes: threshold setting unit, reinforcement learning algorithm unit, switching instruction execution unit; The threshold setting unit is used to set the three-level threshold and automatically adjust the threshold range according to the real-time load rate of the power grid; The reinforcement learning algorithm unit is used to automatically generate a switching strategy using the reinforcement learning algorithm when the parameter triggers an early warning, calculate the switching combination of the capacitor bank, and determine the number and order of the capacitors to be connected or disconnected; The switching instruction execution unit is used to generate a switching instruction according to the optimal switching combination and control the on-site equipment to execute.

Images