Dispatching energy management system for assisting power peak regulation

Abstract

The invention relates to the technical field of electric power peak regulation of energy storage equipment, in particular to a dispatching energy management system for assisting electric power peak regulation, which comprises a load prediction unit, a dynamic adjustment unit and a model optimization unit. A dynamic adjustment unit monitors multiple parameters of an energy storage equipment power system through principal component analysis dimensionality reduction, a peak regulation threshold value of the energy storage equipment is dynamically adjusted according to working conditions and load prediction in combination with a fuzzy logic reasoning model, and a model optimization unit analyzes operation data, compares deviation, adjusts fuzzy rules and evaluates the peak regulation effect of the energy storage equipment. The method can flexibly adjust the threshold according to the system operation parameters and the load prediction, improves the peak regulation accuracy of the energy storage equipment, guarantees the stable operation of a power grid and the efficient utilization of energy, further optimizes the peak regulation effect of the energy storage equipment through the feedback improvement rule, and solves the problem of insufficient adaptability of the peak regulation strategy of the existing energy storage equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of power peak regulation of energy storage equipment, and in particular to a dispatching energy management system for assisting power peak regulation. Background Art

[0002] Grid analysis and control is an important technology. With the development of social economy and the growth of energy demand, energy storage technology is increasingly used in power systems. Energy storage equipment plays a key role in balancing electricity supply and demand and improving the stability and flexibility of power systems. However, energy storage equipment faces many challenges in participating in power peak regulation.

[0003] Electricity load is affected by many factors, such as economic activities, residents' living habits, meteorological conditions, etc. Its large fluctuations bring great challenges to the peak-shaving operation of energy storage equipment. At present, the peak-shaving strategy of energy storage equipment is often based on fixed rules or experience settings, lacking sufficient flexibility and adaptability. For example, in actual operation, the electricity load will change dramatically due to special circumstances such as large-scale events, extreme weather and natural disasters. At this time, the traditional fixed peak-shaving strategy cannot adjust the charging and discharging thresholds of the energy storage equipment in time, resulting in the ineffective utilization of energy storage resources, and may cause over-charging or under-charging, which in turn affects the service life and peak-shaving effect of the energy storage equipment, ultimately causing power waste or insufficient power supply, reducing the overall operation efficiency of the power system. In order to solve this technical problem, we provide a dispatching energy management system for assisting power peak-shaving. Summary of the invention

[0004] The object of the present invention is to provide a dispatching energy management system for assisting power peak load regulation to solve the problems raised in the above-mentioned background technology.

[0005] The threshold adjustment strategy model constructed based on the fuzzy logic reasoning model in the present invention dynamically adjusts the start threshold and stop threshold of the peak-shaving of the energy storage device, and optimizes the peak-shaving effect of the energy storage device by feeding back the improved new fuzzy rules in the fuzzy logic reasoning model to the threshold adjustment strategy model, so that the energy storage device can better participate in the power peak-shaving process.

[0006] To achieve the above-mentioned purpose, a dispatching energy management system for assisting power peak regulation is provided, comprising a load prediction unit, a dynamic adjustment unit and a model optimization unit;

[0007] The load forecasting unit uses multi-source data fusion and a long short-term memory network model improved based on an attention mechanism to forecast power load, and outputs the forecast results to the dynamic adjustment unit and the model optimization unit;

[0008] The dynamic adjustment unit monitors the frequency deviation, voltage fluctuation, spare capacity of power generation equipment, power grid flow distribution and load forecast results of the power system in real time, and dynamically adjusts the start threshold and stop threshold of peak load regulation according to different operating conditions and load forecast conditions in combination with the threshold adjustment strategy model constructed based on the fuzzy logic reasoning model;

[0009] The model optimization unit analyzes the operating data of the power system and compares it with the peak-shaving threshold adjustment suggestions output by the fuzzy logic reasoning model and the effects after actual application, obtains the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold, and sets the fuzzy rule adjustment method based on deviation analysis, special events and trend changes to obtain new fuzzy rules. Finally, the new fuzzy rules are evaluated through the power system simulation model to obtain evaluation results.

[0010] As a further improvement of the technical solution, when the load forecasting unit performs multi-source data fusion, the principal component analysis method is used to perform dimensionality reduction processing on the fused data, as follows:

[0011] The data set obtained after fusion of multi-source data is standardized, the covariance matrix of the standardized data set is calculated, and the eigenvalue decomposition of the covariance matrix is ​​performed to obtain the eigenvalue and the corresponding eigenvector;

[0012] Sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first k eigenvalues, and form a principal component transformation matrix, where k is the number of principal components selected, determined according to the cumulative contribution rate, which refers to the ratio of the sum of the first k eigenvalues ​​to the sum of all eigenvalues;

[0013] Finally, the reduced-dimensionality dataset is obtained by multiplying the original data matrix with the principal component transformation matrix.

[0014] As a further improvement of the technical solution, the dynamic adjustment unit constructs a threshold adjustment strategy model based on fuzzy logic reasoning, specifically as follows:

[0015] The power load change rate, the percentage of spare capacity of power generation equipment, and the grid frequency deviation are used as input variables. For each input variable, several fuzzy sets are defined to fuzzify its value range. The fuzzy sets are obtained by historical experience analysis, and each fuzzy set has a corresponding membership function.

[0016] A fuzzy rule base is constructed, and the fuzzy rule base is determined by formulating rule premises, determining rule conclusions and writing complete rules. The input variables are substituted into corresponding membership functions, and the membership degree of the input variables belonging to each fuzzy set is calculated. Inference calculation is performed based on the fuzzy rule base to determine the activation degree of each rule. All activated rule conclusion parts are synthesized to obtain fuzzy output about the peak-shaving threshold adjustment amount. Finally, the fuzzy output is converted into the peak-shaving threshold adjustment amount through the centroid method.

[0017] As a further improvement of the technical solution, the dynamic adjustment unit adopts a network topology analysis method based on graph theory when monitoring the power grid power distribution, which is as follows:

[0018] The power network is represented as a graph, each node in the power grid is abstracted as a vertex in the graph, and the transmission lines and transformers are abstracted as edges in the graph. A graph model corresponding to the actual power grid is constructed, and the power flow calculation algorithm is used to calculate the voltage amplitude and phase angle of each node in the power grid and the power flow of each branch according to the topological structure and component parameters of the power grid, as well as the output of the generator and the load requirements;

[0019] Calculate the degree of each node to obtain the degree centrality. The degree of a node refers to the number of edges connected to the node. Then determine the betweenness centrality of the node by calculating the ratio of the number of times all the shortest paths in the network pass through a node to the total number of shortest paths. Find the key lines and nodes connecting different areas to analyze the connectivity of the power grid diagram. Finally, calculate the ratio of the power flow of each branch to the line transmission capacity to obtain the line load rate. Define the line with a line load rate exceeding 85% as a weak link.

[0020] When it is monitored that the voltage amplitude and phase angle of the key nodes are beyond the normal range or the line load rate of the weak link reaches the limit threshold, the peak-shaving threshold adjustment mechanism is activated.

[0021] As a further improvement of the technical solution, the peak-shaving threshold adjustment mechanism in the dynamic adjustment unit includes lowering the peak-shaving start threshold and increasing the peak-shaving stop threshold, as follows:

[0022] Calculate the voltage deviation percentage of key nodes. When the voltage deviation exceeds ±5%, the operation status of the node is recorded as abnormal. At the same time, obtain the current load level and load change trend. If the load forecast shows that the load increase rate in a fixed time period in the future exceeds 50% of the increase rate in the same time period before and the load rate of the weak link line reaches 80%, reduce the peak load start threshold;

[0023] During the peak load period, the load trend is continuously monitored. When the load starts to decrease and the load decrease rate remains within the preset range for 15 to 30 minutes, prepare to evaluate whether to increase the peak-shaving stop threshold. If the voltage deviation of the key nodes returns to within ±3% and the line load factor drops below 70%, the threshold for stopping the peak-shaving equipment when the load drops to 60% of the peak value will be raised to 65%.

[0024] As a further improvement of the technical solution, the model optimization unit regularly updates and evaluates the threshold adjustment strategy model, as follows:

[0025] Collect the operation data of the power system over a period of time, including the actual load changes, the operating status of the power generation equipment, the frequency deviation of the power grid, the voltage fluctuation and the actual implementation of the peak-shaving operation. Analyze the operation data and compare it with the peak-shaving threshold adjustment suggestions output by the fuzzy logic reasoning model and the effects after actual application. Calculate the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold under different load levels and power generation reserve capacity combinations;

[0026] Based on deviation analysis, special events and trend changes, the fuzzy rules that need to be adjusted are identified, and the fuzzy rule adjustment method is set to adjust the fuzzy rules that need to be adjusted. After the adjustment is completed, the new fuzzy rules are simulated and tested, and the output results of the fuzzy logic reasoning model after applying the new rules are observed. The simulation test results are compared with the expected goals in the historical actual operation data to evaluate the new fuzzy rules.

[0027] As a further improvement of the technical solution, the fuzzy rule adjustment method includes respectively modifying the rule premise, optimizing the rule conclusion part, adding or deleting rules, and adjusting the priority of fuzzy rules, as follows:

[0028] Check the premise part of the fuzzy rule, that is, the conditional statement describing the operating conditions of the power system, and adjust the variable range or membership function in the conditional statement according to the analysis results of the actual operating data;

[0029] For the conclusion part of the fuzzy rule, i.e., the adjustment amount suggestion of the peak-shaving threshold, it is adjusted according to the deviation between the actual effect and the expected effect, i.e., the fuzzy set definition of the adjustment amount is changed or its corresponding membership function is adjusted;

[0030] Add new fuzzy rules according to operating conditions or special circumstances, such as large-scale events, extreme weather and natural disasters, evaluate the peak load regulation effect of fuzzy rules in actual operation, and determine whether to delete the fuzzy rules based on the evaluation results;

[0031] If there are multiple fuzzy rules applicable at the same time, set priorities for different fuzzy rules based on actual operating experience.

[0032] As a further improvement of the technical solution, the model optimization unit evaluates the new fuzzy rule as follows:

[0033] Use historical data to build a power system simulation model, input different operating conditions, observe the output of the fuzzy logic reasoning model after applying the new rules, compare the simulation test results with the expected goals in the historical actual operating data, and evaluate the new fuzzy rules;

[0034] The fuzzy rules that have undergone simulation tests are gradually applied to the dynamic adjustment unit, and online debugging is carried out to monitor the peak-shaving effect and actual operation status in real time. The fuzzy rules are optimized according to the actual operation status.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] In a dispatching energy management system for assisting power peak load regulation, a threshold adjustment strategy model constructed based on a fuzzy logic reasoning model can flexibly adjust the thresholds related to the peak load regulation of energy storage equipment according to various operating parameters of the power system and load forecast conditions, thereby improving the accuracy and adaptability of the peak load regulation of energy storage equipment, and ensuring the stable operation of the power grid and efficient use of energy with the participation of energy storage equipment. By feeding back the improved new fuzzy rules in the fuzzy logic reasoning model to the threshold adjustment strategy model, the model can better adapt to changes in the power system, improve the effectiveness of the threshold adjustment of energy storage equipment, further ensure the stable operation of the power grid and optimize the peak load regulation effect of energy storage equipment, avoid the problem of low power transmission efficiency due to power waste or insufficient power supply, and improve power transmission efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is an overall block diagram of the present invention.

[0038] The meaning of each number in the figure is:

[0039] 1. Load forecasting unit; 2. Dynamic adjustment unit; 3. Model optimization unit. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Existing power peak-shaving strategies have problems when facing complex operating environments and load fluctuations. The power load fluctuates greatly due to a variety of factors. Existing peak-shaving strategies based on fixed rules or experience are difficult to adapt to special situations such as large-scale events and extreme weather, resulting in power waste or insufficient supply, thus affecting power transmission efficiency.

[0042] The present invention provides a dispatching energy management system for assisting power peak regulation. Figure 1 As shown, it includes a load forecasting unit 1, a dynamic adjustment unit 2, and a model optimization unit 3;

[0043] The present invention is closely related to the peak-shaving application of energy storage equipment. The load prediction unit predicts the load through multi-source data fusion and an improved long short-term memory network model, and reduces the dimension through principal component analysis, so as to provide an accurate basis for the charging and discharging control of the energy storage equipment. The dynamic adjustment unit monitors multiple parameters of the power system, combines the fuzzy logic reasoning model, and dynamically adjusts the peak-shaving threshold according to the working condition and load prediction. It can accurately control the input and exit time of the energy storage equipment, so that it can better adapt to the changes in the power system and improve the peak-shaving efficiency of the energy storage equipment. The model optimization unit continuously optimizes the peak-shaving strategy by analyzing the operating data, comparing the deviation, adjusting the fuzzy rules and evaluating, so as to ensure the effectiveness of the energy storage equipment in the power peak-shaving, thereby ensuring the stable operation of the power grid and the efficient use of energy, solving the problem of insufficient adaptability of the existing peak-shaving strategy, and further enhancing the role of the energy storage equipment in the power peak-shaving.

[0044] Electricity load is affected by a variety of complex factors. A single data source cannot fully reflect the changing patterns of load. For example, relying solely on historical electricity load data cannot take into account the impact of changing meteorological conditions on residential and industrial electricity consumption, nor can it capture load fluctuations caused by social and economic activities. Integrating information from multiple data sources can provide a more comprehensive understanding of the various factors affecting electricity load, thereby excavating more complex load change patterns and providing a basis for more accurate predictions.

[0045] The load forecasting unit 1 uses multi-source data fusion and a long short-term memory network model improved based on the attention mechanism to forecast the power load, so that the model can adaptively use various data information according to the actual situation, better adapt to the complex and changeable operating environment of the power system, and output the forecast results to the dynamic adjustment unit 2 and the model optimization unit 3;

[0046] The data dimensionality after multi-source data fusion is often high and contains a lot of information, but it also brings problems of increased data complexity and increased computing costs;

[0047] When the load forecasting unit 1 performs multi-source data fusion, it uses the principal component analysis method to perform dimensionality reduction processing on the fused data, as follows:

[0048] Because data from different data sources have different dimensions and scales, the data set obtained after the fusion of multi-source data is standardized. Through standardization, all data are converted into standard normal distribution data with a mean of 0 and a standard deviation of 1. The covariance matrix of the standardized data set is calculated. The covariance matrix reflects the correlation between each variable, and the covariance matrix is ​​eigenvalue decomposed to obtain the eigenvalue and the corresponding eigenvector;

[0049] Sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first k eigenvalues, retain the feature information in the original data that contributes more to the data variance, and form a principal component transformation matrix, which reduces the data dimension and helps reduce the complexity and computational cost of data processing. K is the number of principal components selected, which is determined based on the cumulative contribution rate. The cumulative contribution rate refers to the ratio of the sum of the first k eigenvalues ​​to the sum of all eigenvalues.

[0050] Finally, by multiplying the original data matrix with the principal component transformation matrix, we get the reduced-dimensional data set. The data after dimensionality reduction is more concise and effective, and the interference factors are removed, so that the load forecasting model can better focus on key information, thereby improving the accuracy of load forecasting. After data dimensionality reduction, the amount of calculation is reduced and the real-time processing capability of the system is enhanced. In the actual application scenario of power peak regulation, the system can process and analyze real-time data more quickly and make peak regulation decisions in a timely manner.

[0051] The operation of the power system is affected by many factors, such as the user's electricity consumption behavior, the operating status of the power generation equipment, the topological structure of the power grid and the natural environment. These factors interact with each other, making the parameters of the power system such as frequency, voltage, and power flow distribution in a constantly changing state;

[0052] The dynamic adjustment unit 2 monitors the frequency deviation, voltage fluctuation, spare capacity of power generation equipment, power grid flow distribution and load forecast results of the power system in real time. Combined with the threshold adjustment strategy model constructed based on the fuzzy logic reasoning model, it can more accurately judge the operating status and peak-shaving demand of the power system. According to different operating conditions and load forecast conditions, it dynamically adjusts the start threshold and stop threshold of peak-shaving to balance power supply and demand and ensure the stable operation of the power grid.

[0053] The method of the dynamic adjustment unit 2 in constructing the threshold adjustment strategy model based on fuzzy logic reasoning is as follows:

[0054] The power load change rate, the percentage of spare capacity of power generation equipment, and the grid frequency deviation are used as input variables. These variables can reflect the current operating status of the power system and the degree of demand for peak load regulation. For each input variable, several fuzzy sets are defined to fuzzify its value range. The fuzzy sets are obtained by historical experience analysis, and each fuzzy set has a corresponding membership function.

[0055] Taking the power load change rate as an example, if the historical data shows that the load change rate usually fluctuates between -20% and 20%, we can define a "low load change rate" fuzzy set (-20% to -5%), a "medium load change rate" fuzzy set (-5% to 5%) and a "high load change rate" fuzzy set (5% to 20%). Each fuzzy set has a corresponding membership function to describe the degree to which the input value belongs to the fuzzy set. For the "low load change rate" fuzzy set, a triangular membership function can be used, with a membership of 1 at -20% and 0 at -5%, with a linear change in the middle.

[0056] Different fuzzy sets and membership functions can be adjusted and optimized according to actual conditions, so that the model can flexibly handle various input situations, instead of being restricted by fixed rules and parameters like traditional precise models. The definitions of fuzzy sets and membership functions are based on historical experience analysis, which enables the model to make full use of historical operating data and more accurately adjust the peak-shaving threshold according to the actual operating conditions of the power system.

[0057] When the power load change rate is in the "high load change rate" fuzzy set and the percentage of reserve capacity of power generation equipment is in the "low reserve capacity percentage" fuzzy set, the model can more accurately determine the adjustment direction and amplitude of the peak-shaving threshold according to the pre-set fuzzy rules, thereby improving the accuracy of peak-shaving and ensuring the stable operation of the power grid;

[0058] Construct a fuzzy rule base. The fuzzy rule base is determined by formulating rule premises, determining rule conclusions, and writing complete rules. The construction of the fuzzy rule base is based on expert experience and actual data. Expert experience can provide a qualitative understanding of the operation of the power system, while actual data can be used to determine the membership function and verify the effectiveness of the rules. By integrating the two, a more accurate and practical threshold adjustment strategy model can be constructed;

[0059] The rule premise is a logical expression composed of fuzzy sets of input variables, such as "if the power load change rate is a high load change rate and the power generation equipment reserve capacity percentage is a low reserve capacity percentage";

[0060] For each rule premise, determine the corresponding rule conclusion, that is, the adjustment amount of the peak-shaving threshold. The rule conclusion is also expressed in the form of a fuzzy set, such as "large peak-shaving threshold adjustment amount", "medium peak-shaving threshold adjustment amount" and "small peak-shaving threshold adjustment amount". For the above rule premise, the corresponding rule conclusion can be "large peak-shaving threshold adjustment amount";

[0061] Combine the rule premise and rule conclusion to form a complete fuzzy rule, and build a fuzzy rule base containing multiple rules to cover various possible combinations of power system operating conditions and load states;

[0062] Substitute the input variables into the corresponding membership function and calculate the membership degree of the input variables to each fuzzy set. By accurately calculating the membership degree and performing reasonable reasoning calculations, the accuracy of the model can be improved and it can better adapt to the complex changes of the power system.

[0063] For the power load change rate, if its actual value is x, the membership function of the "high load change rate" fuzzy set is μ g (x), then the membership degree of x to the “high load change rate” fuzzy set is calculated to be μ g (x);

[0064] Based on the fuzzy rule base, the inference calculation is carried out. For each rule, the activation degree of each rule is determined by fuzzy logic operation according to the membership degree of the input variable;

[0065] For the rule "If the power load change rate is a high load change rate and the reserve capacity percentage of the power generation equipment is a low reserve capacity percentage", if the power load change rate belongs to the "high load change rate" fuzzy set, the membership degree is μ g (x), the percentage of spare capacity of power generation equipment belongs to the fuzzy set of “low spare capacity percentage”, and the membership degree is μ d (y), then the activation degree of the rule can be obtained through fuzzy AND operation, that is, min(μ g (x), μ d (y));

[0066] The “maximum-minimum synthesis method” is used to synthesize the conclusions of all activated rules to obtain a fuzzy output about the peak-shaving threshold adjustment amount, namely:

[0067] If there are two activation rules, one rule has the conclusion of "large peak-shaving threshold adjustment", and its membership degree is μ big , the other rule conclusion is "mid-peak threshold adjustment", and its membership degree is μ mid , then the synthesized fuzzy output is max(min(μ big ,α),min(μ mid, β)), α is the large peak-shaving threshold adjustment amount, β is the medium peak-shaving threshold adjustment amount;

[0068] Assume that the membership function of the fuzzy output is μ(z), z is the value range of the peak-shaving threshold adjustment amount, and the calculation formula of the centroid method is: where z * The peak-shaving threshold adjustment value is finally obtained. Through the above series of operations, the peak-shaving threshold can be accurately adjusted according to the actual operating conditions of the power system, thereby improving the accuracy and effectiveness of peak-shaving and ensuring the stable operation of the power grid and the balance of power supply and demand.

[0069] Reasonable peak-shaving threshold adjustment can avoid problems such as excessive frequency deviation and voltage instability in the power grid, and improve the stability of power grid operation. At the same time, by optimizing the peak-shaving strategy, it can reduce the operating time of power generation equipment in the inefficient operating range, improve the utilization rate of power generation equipment, reduce power generation costs, and achieve economical power grid operation.

[0070] The electric power grid is a complex network system. Its flow distribution, node importance, weak links and other information are crucial to the stable operation of the power system. The network topology analysis method based on graph theory can comprehensively analyze the power grid from the perspective of network structure, reveal its inherent topological characteristics and operating laws, and provide an important basis for the operation and maintenance of the power system.

[0071] The dynamic adjustment unit 2 uses a network topology analysis method based on graph theory when monitoring the power grid flow distribution, as follows:

[0072] The power network is represented as a graph, each load node in the power grid is abstracted as a vertex in the graph, and the transmission lines and transformers are abstracted as edges in the graph. A graph model corresponding to the actual power grid is constructed, and a power flow calculation algorithm is used to calculate the voltage amplitude and phase angle of each node in the power grid and the power flow of each branch according to the topological structure and component parameters of the power grid, as well as the output of the generator and the load requirements. The specific steps are as follows:

[0073] Assume that the node voltage vector is V = [V1, V2, ..., V n ] T , where n is the total number of nodes, V i is the complex representation of the voltage amplitude and phase angle of the ith node, V i，mag is the voltage amplitude, θ iis the phase angle, and the node admittance matrix Y is constructed according to the topological structure and component parameters of the power grid, and the power flow equation I=YV is established; where I is the node injection current vector, for the generator node, the injection current is the known generator output current, and for the load node, the injection current is the negative value of the known load demand current. The above equation is solved iteratively until the iteration error is less than the set value. In each iteration, the estimated value of the node injection current is calculated according to the current voltage estimate, and then the voltage estimate is updated according to the node admittance matrix;

[0074] Calculate the degree of each node to obtain the degree centrality. The degree of a node refers to the number of edges connected to the node. Then, the betweenness centrality of the node is determined by calculating the ratio of the number of times all shortest paths in the network pass through a node to the total number of shortest paths. Let the degree of node i be d i ,but where a ij is an element in the adjacency matrix, and the betweenness centrality of the node is b i , for each pair of nodes (u, v), count the number of shortest paths passing through node i n uv (i), then the betweenness centrality of node i is in is the total number of shortest paths between all node pairs;

[0075] Finally, calculate the ratio of the power flow of each branch to the line transmission capacity to obtain the line load rate. Suppose the power flow of branch l is P l , the line transmission capacity is C l , then the line load factor Lines with a load rate of more than 85% are defined as weak links. When the voltage amplitude and phase angle of the key nodes are monitored to be beyond the normal range or the line load rate of the weak link reaches the limit threshold, the peak-shaving threshold adjustment mechanism is activated. By quickly identifying key nodes and weak links and giving them special attention and protection during the operation of the power grid, the probability of power grid failures can be reduced and the safety and reliability of the power grid can be improved.

[0076] Reasonable adjustment of the peak-shaving threshold according to the load change trend and the operating status of the power grid can enable the peak-shaving equipment to start and stop at a more appropriate time, improve the utilization rate of the peak-shaving equipment, reduce unnecessary start and stop times, thereby improving the peak-shaving efficiency, reducing the peak-shaving cost, and achieving the economy of power grid operation.

[0077] The peak-shaving threshold adjustment mechanism in the dynamic adjustment unit 2 includes lowering the peak-shaving start threshold and raising the peak-shaving stop threshold, as follows:

[0078] Calculate the voltage deviation percentage of the key node, that is, V sj Expressed as actual voltage, V edIndicates the rated voltage. When the voltage deviation percentage exceeds ±5%, the node operation status is recorded as abnormal. At the same time, the current load level L is obtained. fh and load change trend. If the load forecast shows that the load increase rate in the future fixed time period t exceeds 50% of the increase rate in the previous same time period and the load rate of the weak link line reaches 80%, the peak load start threshold is lowered;

[0079] During the peak load period, the load change trend is continuously monitored. When the load starts to decrease and the load decrease rate is maintained at [ΔL min , ΔL max ] range, prepare to evaluate whether to increase the peak-shaving stop threshold. If the voltage deviation of the key nodes returns to within ±3% and the line load rate drops below 70%, the threshold for stopping the peak-shaving equipment when the load drops to 60% of the peak value will be raised to 65%.

[0080] This peak-shaving threshold adjustment mechanism can make adaptive adjustments based on the real-time operating status of the power grid and load change trends, rather than using a fixed threshold. This makes the peak-shaving strategy more flexible and better able to adapt to the complex and changeable operating environment of the power system, avoiding grid frequency deviations and voltage fluctuations caused by excess or insufficient power, and ensuring the stable operation of the power grid.

[0081] The model optimization unit 3 analyzes the operating data of the power system and compares it with the peak-shaving threshold adjustment suggestions output by the fuzzy logic reasoning model and the effects after actual application, obtains the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold, and sets the fuzzy rule adjustment method based on deviation analysis, special events and trend changes to obtain new fuzzy rules. Finally, the new fuzzy rules are evaluated through the power system simulation model to obtain evaluation results.

[0082] The power system is a dynamically changing system. With the passage of time, the access of new energy sources, changes in load trends, etc., the original threshold adjustment strategy model may no longer be applicable. Regularly updating and evaluating the model can make the model adapt to these changes and improve the accuracy and effectiveness of the model.

[0083] The model optimization unit 3 regularly updates and evaluates the threshold adjustment strategy model, as follows:

[0084] Collect the operation data of the power system over a period of time, including the actual load changes, the operating status of the power generation equipment, the frequency deviation of the power grid, the voltage fluctuation and the actual implementation of the peak-shaving operation. Analyze the operation data and compare it with the peak-shaving threshold adjustment suggestions output by the fuzzy logic reasoning model and the effects after actual application. This regular update and evaluation mechanism can continuously optimize the performance of the threshold adjustment strategy model, calculate the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold under different load levels and power generation reserve capacity combinations. Through the analysis of the deviation and the adjustment of the fuzzy rules, the model can better adapt to various operating conditions and improve the prediction accuracy and decision-making ability.

[0085] Based on deviation analysis, special events and trend changes, the fuzzy rules that need to be adjusted are identified, and the fuzzy rule adjustment method is set to adjust the fuzzy rules that need to be adjusted, so that the model can flexibly respond to the complex changes of the power system and have stronger adaptability. After the adjustment is completed, the new fuzzy rules are simulated and tested, and the output results of the fuzzy logic reasoning model after applying the new rules are observed. The simulation test results are compared with the expected goals in the historical actual operation data, and the new fuzzy rules are evaluated. Accurate peak-shaving threshold adjustment helps to balance the supply and demand of electricity, avoid problems such as excessive frequency deviation and voltage instability in the power grid, and improve the stability of power grid operation. At the same time, a reasonable peak-shaving strategy can reduce the operating time of power generation equipment in the inefficient operating range, improve the utilization rate of power generation equipment, reduce power generation costs, and achieve the economy of power grid operation.

[0086] The operating conditions of the power system are complex and dynamically changing, and are affected by many factors. The original fuzzy rules may not accurately reflect the current operating conditions and need to be adjusted according to actual operating data and special circumstances to adapt to the changes in the power system.

[0087] The fuzzy rule adjustment methods include modifying the rule premise, optimizing the rule conclusion part, adding or deleting rules, and adjusting the priority of fuzzy rules, as follows:

[0088] Check the premise part of the fuzzy rule, that is, the conditional statement describing the operating conditions of the power system, and adjust the variable range or membership function in the conditional statement according to the analysis results of the actual operating data;

[0089] For example, in the original fuzzy rule premise, the value range of a variable h is [a, b]. If the actual data shows that the value of this variable is concentrated in [c, d] in most cases, and Then adjust the value range of variable h to [c, d];

[0090] For the conclusion part of the fuzzy rule, i.e., the adjustment amount suggestion of the peak load threshold, it is adjusted according to the deviation between the actual effect and the expected effect, i.e., the fuzzy set definition of the adjustment amount is changed;

[0091] If the deviation indicates that the actual adjustment amount is always greater than the expected adjustment amount and exceeds a certain range, the fuzzy set range of the original "small peak-shaving threshold adjustment amount" can be narrowed, and the fuzzy set range of the original "large peak-shaving threshold adjustment amount" can be expanded to more accurately reflect the actual situation, and the linear membership function can be adjusted to a nonlinear membership function to better adapt to the actual peak-shaving situation.

[0092] Add new fuzzy rules according to operating conditions or special circumstances. Special circumstances include large-scale events, extreme weather and natural disasters. Evaluate the peak-shaving effect of fuzzy rules in actual operation and decide whether to delete the fuzzy rules based on the evaluation results.

[0093] During large-scale events, the power load will suddenly increase with a certain regularity. A fuzzy rule can be added: "If it is during a large-scale event and the load increase rate exceeds a certain value, a large peak-shaving threshold adjustment will be used." If the number of times a rule is triggered during multiple peak-shaving processes is less than 5% of the total peak-shaving times, the rule can be considered for deletion.

[0094] If there are multiple fuzzy rules that are applicable at the same time, set priorities for different fuzzy rules based on actual operation experience, that is, set the priority of rule 1 to 1 and the priority of rule 2 to 2, which means that rule 1 takes precedence over rule 2.

[0095] By adjusting the premise and conclusion of the fuzzy rules, adding and deleting rules, and setting priorities, the fuzzy rules can more accurately reflect the actual operation of the power system, improve the accuracy and adaptability of the fuzzy rules, thereby improving the accuracy and effectiveness of the peak-shaving threshold adjustment, reducing the deviation of the peak-shaving threshold adjustment, and improving the accuracy and effectiveness of the peak-shaving.

[0096] The model optimization unit 3 evaluates the new fuzzy rules as follows:

[0097] The power system simulation model is established by using historical data. Through simulation model evaluation and online debugging, the performance of the new fuzzy rules can be fully tested. Different operating condition data can be input to find possible problems with the rules in advance. The output results of the fuzzy logic reasoning model after applying the new rules can be observed and the output results can be set as T new , T new The expected target in the simulation test results and historical actual operation data is compared to evaluate the new fuzzy rules. The expected target is T yq , the new fuzzy rules are evaluated by calculating the deviation index, namely:

[0098]

[0099] Where n is the number of test data points. The fuzzy rules that have been simulated and tested are gradually applied to the dynamic adjustment unit 2, and online debugging is performed to monitor the peak regulation effect and actual operation in real time. The fuzzy rules are optimized according to the actual operation. If it is found that the grid frequency deviation Δf is too large and is inconsistent with the peak regulation threshold adjustment amount T new If there is any problem, the reasons are analyzed and the relevant fuzzy rules are adjusted. The new fuzzy rules after evaluation and optimization can more accurately adjust the peak-shaving threshold according to the actual operation of the power system, improve the accuracy and adaptability of peak-shaving, and ensure the stable operation of the power grid.

[0100] In the present invention, the load is predicted by the load prediction unit using multi-source data fusion and an improved long short-term memory network model, and the dimension is reduced by principal component analysis. The dynamic adjustment unit monitors multiple parameters of the power system and combines the fuzzy logic reasoning model to dynamically adjust the peak-shaving threshold according to the operating conditions and load forecasts. The model optimization unit analyzes the operating data, compares the deviation, adjusts the fuzzy rules and evaluates them. The threshold can be flexibly adjusted according to the system operating parameters and load forecasts, thereby improving the accuracy and adaptability of peak-shaving, ensuring the stable operation of the power grid and the efficient use of energy. The peak-shaving effect is further optimized through feedback improvement rules to solve the problem of insufficient adaptability of existing peak-shaving strategies.

[0101] 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. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A dispatching energy management system for assisting power peak regulation, characterized in that: It includes a load forecasting unit (1), a dynamic adjustment unit (2) and a model optimization unit (3); The load prediction unit (1) uses multi-source data fusion and a long short-term memory network model improved based on an attention mechanism to predict power load, and outputs the prediction results to the dynamic adjustment unit (2) and the model optimization unit (3); The dynamic adjustment unit (2) monitors the frequency deviation, voltage fluctuation, spare capacity of power generation equipment, power grid flow distribution and load forecast results of the power system in real time, and dynamically adjusts the start threshold and stop threshold of peak load regulation according to different operating conditions and load forecast results in combination with a threshold adjustment strategy model constructed based on a fuzzy logic reasoning model; The model optimization unit (3) analyzes the operation data of the power system and compares the analysis with the peak-shaving threshold adjustment suggestion output by the fuzzy logic reasoning model and the effect after actual application, obtains the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold, and sets a fuzzy rule adjustment method based on the deviation analysis, special events and trend changes to obtain a new fuzzy rule, and finally evaluates the new fuzzy rule through the power system simulation model to obtain an evaluation result.

2. A dispatching energy management system for assisting power peak load regulation according to claim 1, characterized in that: When the load forecasting unit (1) fuses the multi-source data, it uses the principal component analysis method to perform dimensionality reduction processing on the fused data, as follows: The data set obtained after fusion of multi-source data is standardized, the covariance matrix of the standardized data set is calculated, and the eigenvalue decomposition of the covariance matrix is ​​performed to obtain the eigenvalue and the corresponding eigenvector; Sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first k eigenvalues, and form a principal component transformation matrix, where k is the number of principal components selected, determined according to the cumulative contribution rate, which refers to the ratio of the sum of the first k eigenvalues ​​to the sum of all eigenvalues; Finally, the reduced-dimensionality dataset is obtained by multiplying the original data matrix with the principal component transformation matrix.

3. The dispatching energy management system for assisting power peak load regulation according to claim 1, characterized in that: The dynamic adjustment unit (2) constructs a threshold adjustment strategy model based on fuzzy logic reasoning in the following manner: The power load change rate, the percentage of spare capacity of power generation equipment, and the grid frequency deviation are used as input variables. For each input variable, several fuzzy sets are defined to fuzzify its value range. The fuzzy sets are obtained by historical experience analysis, and each fuzzy set has a corresponding membership function. A fuzzy rule base is constructed, and the fuzzy rule base is determined by formulating rule premises, determining rule conclusions and writing complete rules. The input variables are substituted into corresponding membership functions, and the membership degree of the input variables belonging to each fuzzy set is calculated. Inference calculation is performed based on the fuzzy rule base to determine the activation degree of each rule. All activated rule conclusion parts are synthesized to obtain fuzzy output about the peak-shaving threshold adjustment amount. Finally, the fuzzy output is converted into the peak-shaving threshold adjustment amount through the centroid method.

4. The dispatching energy management system for assisting power peak load regulation according to claim 1, characterized in that: The dynamic adjustment unit (2) uses a network topology analysis method based on graph theory when monitoring the power grid power flow distribution, which is specifically as follows: The power network is represented as a graph, each node in the power grid is abstracted as a vertex in the graph, and the transmission lines and transformers are abstracted as edges in the graph. A graph model corresponding to the actual power grid is constructed, and the power flow calculation algorithm is used to calculate the voltage amplitude and phase angle of each node in the power grid and the power flow of each branch according to the topological structure and component parameters of the power grid, as well as the output of the generator and the load requirements; Calculate the degree of each node to obtain the degree centrality. The degree of a node refers to the number of edges connected to the node. Then determine the betweenness centrality of the node by calculating the ratio of the number of times all the shortest paths in the network pass through a node to the total number of shortest paths. Find the key lines and nodes connecting different areas to analyze the connectivity of the power grid diagram. Finally, calculate the ratio of the power flow of each branch to the line transmission capacity to obtain the line load rate. Define the line with a line load rate exceeding 85% as a weak link. When it is monitored that the voltage amplitude and phase angle of the key nodes are beyond the normal range or the line load rate of the weak link reaches the limit threshold, the peak-shaving threshold adjustment mechanism is activated.

5. The dispatching energy management system for assisting power peak load regulation according to claim 1, characterized in that: The peak-shaving threshold adjustment mechanism in the dynamic adjustment unit (2) includes lowering the peak-shaving start threshold and raising the peak-shaving stop threshold, as follows: Calculate the voltage deviation percentage of key nodes. When the voltage deviation exceeds ±5%, the operation status of the node is recorded as abnormal. At the same time, obtain the current load level and load change trend. If the load forecast shows that the load increase rate in a fixed time period in the future exceeds 50% of the increase rate in the same time period before and the load rate of the weak link line reaches 80%, reduce the peak load start threshold; During the peak load period, the load trend is continuously monitored. When the load starts to decrease and the load decrease rate remains within the preset range for 15 to 30 minutes, prepare to evaluate whether to increase the peak-shaving stop threshold. If the voltage deviation of the key nodes returns to within ±3% and the line load factor drops below 70%, the threshold for stopping the peak-shaving equipment when the load drops to 60% of the peak value will be raised to 65%.

6. The dispatching energy management system for assisting power peak load regulation according to claim 1, characterized in that: The model optimization unit (3) regularly updates and evaluates the threshold adjustment strategy model, specifically as follows: Collect the operation data of the power system over a period of time, including the actual load changes, the operating status of the power generation equipment, the frequency deviation of the power grid, the voltage fluctuation and the actual implementation of the peak-shaving operation. Analyze the operation data and compare it with the peak-shaving threshold adjustment suggestions output by the fuzzy logic reasoning model and the effects after actual application. Calculate the deviation between the peak-shaving threshold recommended by the fuzzy rule and the actual optimal peak-shaving threshold under different load levels and power generation reserve capacity combinations; Based on deviation analysis, special events and trend changes, the fuzzy rules that need to be adjusted are identified, and the fuzzy rule adjustment method is set to adjust the fuzzy rules that need to be adjusted. After the adjustment is completed, the new fuzzy rules are simulated and tested, and the output results of the fuzzy logic reasoning model after applying the new rules are observed. The simulation test results are compared with the expected goals in the historical actual operation data to evaluate the new fuzzy rules.

7. A dispatching energy management system for assisting power peak load regulation according to claim 6, characterized in that: The fuzzy rule adjustment method includes respectively modifying the rule premise, optimizing the rule conclusion part, adding or deleting the rule, and adjusting the priority of the fuzzy rule, as follows: Check the premise part of the fuzzy rule, that is, the conditional statement describing the operating conditions of the power system, and adjust the variable range or membership function in the conditional statement according to the analysis results of the actual operating data; For the conclusion part of the fuzzy rule, i.e., the adjustment amount suggestion of the peak-shaving threshold, it is adjusted according to the deviation between the actual effect and the expected effect, i.e., the fuzzy set definition of the adjustment amount is changed or its corresponding membership function is adjusted; Add new fuzzy rules according to operating conditions or special circumstances, such as large-scale events, extreme weather and natural disasters, evaluate the peak load regulation effect of fuzzy rules in actual operation, and determine whether to delete the fuzzy rules based on the evaluation results; If there are multiple fuzzy rules applicable at the same time, set priorities for different fuzzy rules based on actual operating experience.

8. The dispatching energy management system for assisting power peak load regulation according to claim 6, characterized in that: The model optimization unit (3) evaluates the new fuzzy rules as follows: Use historical data to build a power system simulation model, input different operating conditions, observe the output of the fuzzy logic reasoning model after applying the new rules, compare the simulation test results with the expected goals in the historical actual operating data, and evaluate the new fuzzy rules; The fuzzy rules that have undergone simulation tests are gradually applied to the dynamic adjustment unit (2), and online debugging is performed to monitor the peak regulation effect and actual operation status in real time. The fuzzy rules are optimized according to the actual operation status.

Images