A situation awareness method for integrated energy systems

Abstract

The present invention relates to a situation awareness method for an integrated energy system, which belongs to the field of situation awareness of power systems and solves the problems of static analysis, strong subjectivity and lack of dynamic perception in existing perception. It includes determining situation safety indicators, collecting electrical and system parameter data of each energy site in the system within a historical time period, and calculating the situation safety index value at each time point; using fuzzy analytic hierarchy process to obtain the weight decision vector of each expert on the indicator and perform cluster analysis to obtain the weight of each indicator; based on the weight, weighted summation of the indicator value at each time point is performed to obtain the situation awareness result at each time point, and the indicator value at each time point is labeled according to the situation awareness result to form a training sample set to train the convolutional neural network and obtain a trained perception model; based on the electrical and system parameter data of each energy site collected in real time, the indicator value is calculated and input into the perception model to obtain the situation awareness result of the system. Real-time perception of system situation is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power system situation awareness, and in particular to a situation awareness method for an integrated energy system. Background Art

[0002] In recent years, integrated energy systems have developed rapidly to meet the growing energy demand and promote the application of clean energy. However, as the scale and complexity of the system increase, system security has become an increasingly important concern. Existing technologies mainly focus on the security assessment of independent subsystems and lack a comprehensive perception and assessment of the overall system situation. This makes it difficult to accurately identify potential safety hazards in integrated energy systems and make targeted safety decisions in a timely manner.

[0003] At present, most of the existing integrated energy system safety assessment methods still rely on traditional mathematical models and static data analysis, which cannot effectively capture the dynamic changes and uncertainties of the system. Traditional methods ignore the synergy of multiple sites, the correlation between equipment and the dynamic characteristics of the system, resulting in inaccurate and incomplete safety assessment results of the system. The existing technology is highly subjective in terms of expert decision-making weights and lacks an objective and scientific method to determine the weights of various indicators, which affects the objectivity and reliability of the assessment results.

[0004] In summary, there is an urgent need for a situation awareness method for an integrated energy system to improve the accuracy and intelligence level of system situation awareness based on the characteristics of the integrated energy system. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a situation awareness method for an integrated energy system, so as to solve the problems of static analysis, strong subjectivity of experts and lack of dynamic system perception in the existing situation awareness of integrated energy systems.

[0006] The embodiment of the present invention provides a situation awareness method for an integrated energy system, comprising the following steps:

[0007] Determine the situation safety index, collect the electrical parameter data and system parameter data of each energy site in the integrated energy system within the historical time period, and calculate the situation safety index value at each time point;

[0008] The fuzzy hierarchical analysis method is used to obtain the weight decision vector of each expert on the situation safety index, and the weight decision vector is clustered to obtain the weight of each situation safety index;

[0009] Based on the weights, weighted summation of the situation safety index values ​​at each time point is performed to obtain the situation awareness results at each time point, and the situation safety index values ​​at each time point are marked according to the situation awareness results to form a training sample set, and the convolutional neural network is trained to obtain a trained comprehensive energy system situation awareness model;

[0010] Based on the electrical parameter data and system parameter data of each energy site collected dynamically in real time, the situational safety index value is calculated and input into the comprehensive energy system situation awareness model to obtain the safety situation awareness result of the comprehensive energy system.

[0011] Furthermore, the method of obtaining the weighted decision vector of each expert on the situation safety index by using the fuzzy analytic hierarchy process includes:

[0012] Each expert compares the situation safety indicators in pairs to obtain fuzzy judgment results, and constructs a fuzzy complementary matrix based on the fuzzy judgment results, where each element value (i, j) of the matrix is ​​the fuzzy judgment result between situation safety indicators i and j;

[0013] Based on the element values ​​of the fuzzy complementary matrix, the weighted decision results of each expert on each situational safety index are calculated. , and obtain the weighted decision vector of each expert on the situational safety index ; Where Z=1…N, N represents the number of experts.

[0014] Furthermore, the fuzzy complementary matrix is ​​as follows:

[0015] ;

[0016] Among them, n is the number of situational safety indicators, , the element values ​​on the matrix diagonal are 1.0, and the element values ​​at other positions are 0.1~0.9.

[0017] Furthermore, based on the element values ​​of the fuzzy complementary matrix, the weighted decision results of each expert on each situational safety index are calculated as follows:

[0018] ;

[0019] in, is the weight decision result of the expert group on indicator i, and any two indicators m and j among the n indicators satisfy , The discernment of decision makers.

[0020] Furthermore, the weights of each situation safety index obtained by clustering analysis of the weight decision vector include:

[0021] Performing cluster analysis on the weight decision vector using a k-means algorithm;

[0022] Select the number of clusters k, and randomly select k experts' weighted decision vectors for situational safety indicators as the initial clustering center of each cluster ;

[0023] in, is the weight decision vector of the i-th cluster center;

[0024] Calculate the Euclidean distance from each weighted decision vector to the cluster center of each cluster, and assign the weighted decision vector to the cluster represented by the nearest cluster center;

[0025] For each cluster, the mean of all weight decision vectors in the cluster is calculated, and the mean is used as the new cluster center. The clustering operation is terminated until the convergence distance of each cluster center is less than the predetermined convergence parameter or the predetermined number of iterations is reached.

[0026] Based on the number of weighted decision vectors contained in each cluster, the weight of the weighted decision vector of each cluster is calculated;

[0027] Based on the weight of the weight decision vector of each cluster and the cluster center of each cluster, the weight of each situation safety index is calculated.

[0028] Furthermore, the weight of the weight decision vector in each cluster is calculated as follows:

[0029] ;

[0030] in, is the weight of the lth cluster, is the number of weight decision vectors in the lth cluster, k is the number of clusters, is the number of all weight decision vectors in k clusters.

[0031] Furthermore, the weight of the situational safety index is calculated as follows:

[0032] ;

[0033] in, is the weight of the i-th situational safety indicator, is the weight decision result of the i-th situational safety indicator in the weight decision vector corresponding to the cluster center of the l-th cluster.

[0034] Furthermore, the situation awareness results at each time point are calculated as follows:

[0035] ;

[0036] Among them, S, are the situation awareness result at each time point, the weight of the i-th situation safety index, and the situation safety index value.

[0037] Furthermore, the situational safety index values ​​at each time point are marked according to the situational awareness results, including:

[0038] The security situation awareness results are judged according to the preset threshold to obtain the situation security level;

[0039] The situational safety level includes four safety levels: normal, recovery, critical and emergency;

[0040] The situation security index value is marked according to the situation security level corresponding to the situation security index value at each time point.

[0041] Furthermore, the situational safety index includes a line flow distribution index , Functional quality qualification rate , Transient absolute safety margin , Transient comprehensive safety margin , New energy point grid-connected power quality grid-connected node voltage qualification rate , New energy point grid-connected power quality grid-connected node current qualification rate , New energy point grid-connected power quality grid-connected node frequency qualification rate , System load rate , System overload degree , Steady-state total adequacy , Steady-state component margin and transformer load factor ;

[0042] Furthermore, the integrated energy system includes energy storage stations, charging and swapping stations, substations, new energy stations, data center stations, 5G base stations and Beidou navigation stations.

[0043] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0044] 1. The present invention provides a situation awareness method for an integrated energy system, which fully considers the characteristics of the operation of a multi-station system in an integrated energy system, obtains the weight decision vector of each expert on the situation safety index by introducing the fuzzy hierarchical analysis method, and combines the expert decision results to perform cluster analysis on the weight decision vector of the situation safety index using the k-means algorithm to accurately calculate the weight of each situation safety index;

[0045] 2. The convolutional neural network is used to conduct a comprehensive analysis of the electrical parameter data and system parameter data of the integrated energy system collected in real time and dynamically, highlighting the real-time comprehensive perception and scientific situation perception of the system security situation; the present invention provides a new, more accurate and reliable real-time situation perception method for the safety management of the integrated energy system;

[0046] 3. Determine the situational safety indicators of the integrated energy system and provide a comprehensive situational safety indicator framework that can cover all aspects of the integrated energy system, including static and dynamic safety, real-time monitoring of system operation status, and timely perception of system problems and risks, which will help formulate safety strategies and measures, support decision-making processes, and prevent potential accidents.

[0047] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0049] Figure 1 A flow chart of a situational awareness method for an integrated energy system;

[0050] Figure 2 Flowchart for calculating the weights of various situational safety indicators using the k-means algorithm for cluster analysis. DETAILED DESCRIPTION

[0051] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0052] like Figure 1 As shown, a specific embodiment of the present invention discloses a situation awareness method for an integrated energy system, comprising the following steps:

[0053] Step S1, determine the situation safety index, collect the electrical parameter data and system parameter data of each energy site in the comprehensive energy system in the historical time period, and calculate the situation safety index value at each time point;

[0054] Step S2, using fuzzy analytic hierarchy process to obtain the weight decision vector of each expert for the situation safety index, and performing cluster analysis on the weight decision vector to obtain the weight of each situation safety index;

[0055] Step S3, based on the weights, weighted summing up the situation security index values ​​at each time point obtains the situation awareness results at each time point, annotating the situation security index values ​​at each time point according to the situation awareness results, forming a training sample set, training the convolutional neural network, and obtaining a trained comprehensive energy system situation awareness model;

[0056] Step S4: Calculate the situational safety index value based on the electrical parameter data and system parameter data of each energy site collected dynamically in real time, and input it into the comprehensive energy system situation awareness model to obtain the safety situation awareness result of the comprehensive energy system.

[0057] Step S1, includes steps S11-S12, specifically.

[0058] Step S11, determining the situational safety index.

[0059] The situational safety index includes a line flow distribution index , Functional quality qualification rate , Transient absolute safety margin , Transient comprehensive safety margin , New energy point grid-connected power quality grid-connected node voltage qualification rate , New energy point grid-connected power quality grid-connected node current qualification rate , New energy point grid-connected power quality grid-connected node frequency qualification rate , System load rate , System overload degree , Steady-state total adequacy , Steady-state component margin and transformer load factor .

[0060] The situational security indicators are as follows:

[0061] (1) Power grid flow distribution index

[0062] Refers to the average value of the difference between the maximum transmission capacity allowed by the line in the system and the active power flow of the line. This indicator reflects the stability of the system. The larger the indicator value, the farther the system is from the maximum transmission capacity allowed, and the more stable the system is.

[0063] 1) Use the power flow calculation method to calculate the active power flow of each line;

[0064] 2) For each line, determine its allowable limit transmission capacity based on equipment parameters and operating conditions, that is, the limit transmission capacity of each line;

[0065] 3) Obtain the total number of system lines n from the system topology;

[0066] 4) For each line, calculate the difference between its limit transmission capacity and the actual active power flow;

[0067] 5) Average the differences of all lines to obtain the power grid power flow distribution index value, as shown in formula (1).

[0068] ;

[0069] in, is the limit transmission capacity of line j in the system; n is the total number of lines in the system; is the active power flow of system line j. The active power flow is the actual power transmitted in the power system, and the unit is watt.

[0070] The numerical value of the power grid flow distribution index reflects the stability of the system. A larger index value means that the system is far from the limit transmission capacity under the current operating conditions and has better stability. This index can help system operation and maintenance personnel monitor the system load status and promptly discover potential overload risks, so as to take corresponding measures to ensure the safe operation of the power system.

[0071] (2) Functional quality pass rate

[0072] Functional quality qualification rate refers to whether the energy quality of the system is qualified. It is defined as the ratio of the amount of energy with qualified quality in the system to the amount of output energy types. The larger the value, the better the energy supply quality of the system. Functional quality qualification rate , as shown in formula (2):

[0073] ;

[0074] in, is a 0-1 variable of energy m; is the number of output energy types; e, h, g, w, l represent electricity, heat, gas, wind, and light energy flows respectively; is a 0-1 variable for energy m mass item i; is the actual value of energy quality item i; is the set of main mass items of energy m.

[0075] for is a 0-1 variable for energy w, i.e., if the mass item of energy w exists, the variable is 1, otherwise it is 0;

[0076] Output energy type refers to the types and categories of various energy generated by the system. Common energy types include but are not limited to: electrical energy, thermal energy, gas energy, mechanical energy, light energy, wind energy, etc.

[0077] Functional quality pass rate Quantitative indicators are provided for evaluating the performance and quality aspects of integrated energy systems.

[0078] (3) Transient absolute safety margin

[0079] The system transient absolute safety margin is the minimum value of the transient absolute safety margins of all nodes and branches in the system, and its physical meaning is the local vulnerability of the system.

[0080] Transient refers to the instantaneous response and change of the system after a transient disturbance, fault or emergency occurs, which may include sudden voltage fluctuations, short circuit faults and equipment disconnection.

[0081] Under transient conditions, the system may experience instantaneous voltage, current and power changes. Safety margin analysis and evaluation under transient conditions are very important to ensure the safe and stable operation of the system.

[0082] Absolute transient safety margin of the system , as shown in formula (3):

[0083] ;

[0084] in, is the transient absolute safety margin of the node or branch. The system’s transient absolute safety margin is the minimum value of the single-index transient safety margin in node or branch i. It is the transient safety margin of a single indicator, which is the ratio of the integral of the absolute value of the difference between the actual value and the standard value of a single indicator j and the integral of the absolute value of the difference between the standard value and the critical value within the maximum allowable fault time of the node or branch i at time t.

[0085] The transient absolute safety margin is used to measure the vulnerability of the system under transient conditions and help evaluate the overall transient safety of the system.

[0086] (4) Transient comprehensive safety margin

[0087] Transient comprehensive safety margin It is an evaluation index of the overall vulnerability of the system, which is the average value of the transient absolute safety margins of all nodes and branches. This index evaluates the vulnerability of the entire system when facing transient conditions by considering the transient absolute safety margins of all nodes and branches in the system. Its physical meaning is the overall vulnerability of the system.

[0088] Transient comprehensive safety margin , as shown in formula (4):

[0089] ;

[0090] in, It is the transient absolute safety margin of the node or branch.

[0091] The transient comprehensive safety margin is obtained by adding up the transient absolute safety margins of all nodes and branches in the system and dividing it by the total number of nodes and branches to obtain the average value.

[0092] The physical meaning of this indicator is to provide a comprehensive assessment of the transient safety of the system. By considering the vulnerability of each part, the overall performance of the system when facing transient conditions can be more comprehensively evaluated, which is conducive to the formulation of corresponding safety strategies and measures.

[0093] (5) Grid-connected power quality of new energy points

[0094] 1) In terms of grid-connected node voltage qualification rate, since the new energy station is connected to the grid in the integrated energy system, the new energy station injects a large amount of active power and reactive power into the grid, causing the voltage at the grid-connected node to increase or decrease significantly, making the voltage value in the active distribution network exceed the normal value range for safe and stable operation. As shown in formula (5):

[0095] ;

[0096] Among them, V is the total number of nodes connected to the grid of new energy points.

[0097] 2) In terms of the qualified rate of grid-connected node current, when the new energy station is under maintenance or fails, it continues to supply the power outage line and surrounding loads, thus forming an island effect, which is very likely to damage the power-consuming equipment. When the connection is restored, the phase of the distributed energy and the power grid is not synchronized, which will form a huge impact current that harms the system; as shown in formula (6):

[0098] ;

[0099] Among them, A is the total number of new energy point grid-connected nodes.

[0100] 3) In terms of the grid-connected node frequency qualification rate, as new energy stations are connected to the grid, the randomness and volatility of the output of new energy stations can easily affect the frequency stability in the active distribution network. As shown in formula (7):

[0101] ;

[0102] Among them, H is the total number of new energy point grid-connected nodes.

[0103] These indicators highlight the potential impact of the grid connection of new energy stations on the power quality of the power grid, and corresponding measures need to be taken in system design and operation to ensure the stability and reliability of power quality.

[0104] (6) System load rate

[0105] System load factor Refers to the ratio of the system line transmission power and the line's allowed maximum transmission capacity.

[0106] System load factor , as shown in formula (8):

[0107] ;

[0108] in, is the maximum transmission capacity allowed by the system line and, Transmit power and for system lines.

[0109] This indicator reflects the probability of a major power outage in the system. The larger the value of this indicator, the more it indicates that the system line transmission power is close to or exceeds the allowable limit transmission capacity of the system line, and the greater the probability of a major power outage in the system.

[0110] (7) System overload degree

[0111] When a component in the system fails, it may cause the load of some lines to exceed their normal load level, and the system overload degree It is the ratio of the number of overloaded lines in the system to the total number of remaining lines.

[0112] This indicator represents the degree of overloaded lines caused by system component failure. The larger the indicator is, the more lines in the system deviate from the normal state, the greater the degree of system overload, and the more dangerous the system state is.

[0113] System overload level , as shown in formula (9):

[0114] ;

[0115] Where m is the number of overloaded lines in the system, and n is the total number of lines in the system.

[0116] For example, the line overload determination standard is that the line load rate exceeds 1.1 times the normal line load rate. The threshold value (1.1 times) is preset according to the design or operation standard of the system.

[0117] The degree of system overload is measured by calculating the ratio of the number of overloaded lines m to the total number of remaining lines n. The larger the ratio, the greater the degree of system overload, and the system may be in a more dangerous state.

[0118] System overload level Helps to assess the extent to which the system may be affected by overloads when a system component fails.

[0119] (8) Steady-state total adequacy

[0120] Steady state means that the system is in a state of relative equilibrium, the changes in various physical quantities tend to be stable, and there is no obvious fluctuation over time, which is in contrast to the transient and transient states of the system.

[0121] The steady-state total amount margin refers to whether the total energy supply of the system meets user demand under steady-state conditions. It is the ratio of the total output energy of the system to the total load demand (i.e. user demand). When the total load demand is met, the steady-state total amount margin is 1.

[0122] Steady-state total adequacy , as shown in formula (10):

[0123] ;

[0124] in, is the total output energy of the energy quality coefficient; is the total load of the energy quality coefficient.

[0125] Under steady-state conditions, the total system output energy needs to meet the needs of the total system load. This balance is to ensure that the system can operate stably.

[0126] The steady-state total adequacy is an indicator that comprehensively considers energy quality and supply-demand balance, and helps to evaluate the energy supply status of the system in a steady-state system.

[0127] (9) Steady-state component margin

[0128] The steady-state component margin is used to evaluate whether the energy supply of various types of the system meets the corresponding load demand under steady-state conditions.

[0129] For each output energy i, the matching degree indicates the matching degree between its actual supply and demand. The matching degree is an indicator to measure the energy matching of the system at a specific moment;

[0130] The expected value represents the expected supply of output energy i by the system under given conditions. The expected value is based on previous experience.

[0131] The standard deviation is a measure of the degree of dispersion between a set of numerical values ​​of matching degree and expected value. It is used to measure the degree of dispersion of the matching degree of various types of energy in the system. The larger the standard deviation, the greater the overall variation in the matching degree of various types of energy.

[0132] Steady-state component margin For each energy matching The standard deviation of its expected value. As shown in formula (11):

[0133] ;

[0134] in, is the matching degree of output energy i, for The expected value of , are the supply and demand of output energy i respectively.

[0135] By calculating the steady-state component margin, we can understand the supply and demand of various energies in the system under steady-state conditions, as well as the overall stability of these energy matching degrees, which helps to evaluate the stability and reliability of the system.

[0136] (10) Transformer load factor

[0137] Transformer load factor Refers to the average value of the ratio of the actual output power of the system transformer to the rated capacity. This indicator reflects the load bearing capacity of the system transformer. The larger the indicator, the greater the load on the system transformer and the more likely the system is to have an accident.

[0138] Transformer load factor , as shown in formula (12):

[0139] ;

[0140] in, is the active power output by transformer b in the system, is the rated capacity of transformer b, is the number of transformers in the system.

[0141] Each transformer in the system has its own independent load condition. The output active power of transformer b is the active power actually output by transformer b, and the rated capacity of transformer b is the maximum power that the transformer can withstand.

[0142] By calculating the transformer load rate, we can understand the load conditions of each transformer in the system relative to its rated capacity. A higher transformer load rate indicates that the system is overloaded, increasing the risk of accidents in the system.

[0143] The 10 situation indicators are used to establish a comprehensive energy system situation security indicator system, which mainly characterizes the system situation security from the perspective of static security and dynamic security. Among them, (1)-(5) are static security indicators, and (6)-(10) are dynamic security indicators.

[0144] Step S12, collecting electrical parameter data and system parameter data of each energy site in the integrated energy system within a historical time period, and calculating the situation safety index value at each time point.

[0145] The comprehensive energy system includes energy storage stations, charging and swapping stations, substations, new energy stations, data center stations, 5G base stations and Beidou navigation stations.

[0146] Based on the situation safety index of step S11, the electrical data and system parameter data of each energy site in a historical time period are collected, and the situation safety index value at each time point in the historical time period is calculated.

[0147] Step S2, includes steps S21-S22, specifically.

[0148] Step S21, using fuzzy analytic hierarchy process to obtain the weight decision vector of each expert on the situation safety index.

[0149] The perception of the operating status of the integrated energy system requires a comprehensive evaluation. The relative importance of each indicator is different. How to set the weights between indicators is the key to the comprehensive evaluation.

[0150] Fuzzy analytic hierarchy process is a multi-index weight allocation method that combines fuzzy mathematics with analytic hierarchy process. It overcomes the problem that the consistency of the judgment matrix of analytic hierarchy process is difficult to test, and can eliminate the uncertainty problem in the allocation of index weights.

[0151] The expert group is determined, and the fuzzy analytic hierarchy process is used to obtain the weighted decision vector of each expert on the situational safety index.

[0152] The method of using fuzzy analytic hierarchy process to obtain the weight decision vector of each expert on the situation safety index includes:

[0153] Each expert compares the situation safety indicators in pairs to obtain fuzzy judgment results, and constructs a fuzzy complementary matrix based on the fuzzy judgment results, where each element value (i, j) of the matrix is ​​the fuzzy judgment result between situation safety indicators i and j;

[0154] Based on the element values ​​of the fuzzy complementary matrix, the weighted decision results of each expert on each situational safety index are calculated. , and obtain the weighted decision vector of each expert on the situational safety index ; Where Z=1…N, N represents the number of experts.

[0155] An expert group is determined, and the expert group includes N experts. The expert group is composed of professionals with professional knowledge and experience in this field. In the fuzzy hierarchical analysis method of the present invention, the expert group provides subjective judgments on the relative importance of situational security indicators.

[0156] For each two situational security indicators, the expert group made a fuzzy comparison of their relative importance and evaluated them using fuzzy numerical values ​​based on the subjective judgment of the expert group. The comparison results were expressed using values ​​between 0.1 and 0.9, with 0.1 indicating that the mutual influence between the two was weak and 0.9 indicating that the mutual influence between the two was strong.

[0157] For example, if the line power distribution index Relative to functional quality pass rate If the impact is greater, a higher comparison result is given, such as 0.7 or 0.8;

[0158] If the quality of the grid-connected power of the new energy point is relatively high compared to the transformer load rate If the effect is small, a lower comparison result is given, such as 0.1.

[0159] The fuzzy complementary matrix is ​​as follows:

[0160] ;

[0161] Among them, n is the number of situational safety indicators, , the element values ​​on the matrix diagonal are 1.0, and the element values ​​at other positions are 0.1~0.9.

[0162] Element values ​​of fuzzy complementary matrix The quantitative scale is 0.1~0.9. Since the diagonal of the matrix is ​​the comparison of the indicator itself, the element value on the diagonal of the matrix is ​​1.0.

[0163] Based on the element values ​​of the fuzzy complementary matrix, the weight decision results of each expert on each situational safety index are calculated as follows:

[0164] ;

[0165] in, is the weight decision result of the expert group on indicator i, and any two indicators m and j among the n indicators satisfy , The discernment of decision makers.

[0166] The value of is between 0 and 1. A smaller value indicates that the decision maker has a weaker ability to identify the differences between indicators, and a larger value indicates that the decision maker is more able to distinguish the relative importance of indicators. The value of is used to improve the resolution of the weight distribution.

[0167] The expert group provides professionals who judge the importance of indicators, and the decision makers are the people who calculate the weights and make the final decision based on the expert opinions. The two can be the same group of people or different groups. In specific cases, the expert group is professionals in this field who provide subjective judgments on the relative importance of indicators. The decision makers are the people who calculate the weights and make the final decision based on the expert judgment. This division of labor helps to comprehensively consider the views of different professional backgrounds and experiences, and improve the objectivity and scientificity of the evaluation.

[0168] Step S22, cluster analysis is performed on the weight decision vector to obtain the weight of each situation safety index. Figure 2 shown.

[0169] The weights of each situational safety index obtained by clustering analysis of the weight decision vector include:

[0170] Performing cluster analysis on the weight decision vector using a k-means algorithm;

[0171] Select the number of clusters k, and randomly select k experts' weighted decision vectors for situational safety indicators as the initial clustering center of each cluster ;

[0172] in, is the weight decision vector of the i-th cluster center;

[0173] Calculate the Euclidean distance from each weighted decision vector to the cluster center of each cluster, and assign the weighted decision vector to the cluster represented by the nearest cluster center;

[0174] For each cluster, the mean of all weight decision vectors in the cluster is calculated, and the mean is used as the new cluster center. The clustering operation is terminated until the convergence distance of each cluster center is less than the predetermined convergence parameter or the predetermined number of iterations is reached.

[0175] Based on the number of weighted decision vectors contained in each cluster, the weight of the weighted decision vector of each cluster is calculated;

[0176] Based on the weight of the weight decision vector of each cluster and the cluster center of each cluster, the weight of each situation safety index is calculated.

[0177] (1) Initialization of k-means algorithm.

[0178] The number of clusters selected for classification is k. When performing the k-means algorithm, it is necessary to determine in advance how many clusters the data will be divided into. Choosing an appropriate number of clusters k requires some prior knowledge or determination through experiments. Different numbers of clusters lead to different clustering results. Too small a cluster may merge different entities that should be separated, while too large a cluster number may introduce noise.

[0179] The predetermined number of iterations V for the cluster analysis to end, or the convergence coefficient M for the cluster analysis to end.

[0180] Randomly select k experts’ weighted decision vectors for situational safety indicators as the initial clustering center of each cluster ;in, is the weight decision vector of the i-th cluster center;

[0181] (2) Calculate the Euclidean distance from each weighted decision vector to the cluster center of each cluster, and assign the weighted decision vector to the cluster represented by the nearest cluster center;

[0182] The Euclidean distance calculation from each weighted decision vector to the cluster center is shown in formula (15):

[0183] ;

[0184] Where z is the index or label of the cluster, is the cluster center of the zth cluster, l is the number of indicators, is the lth eigenvalue of the cluster center, A decision vector of weights for each data point.

[0185] The calculated Euclidean distance is used to measure the similarity between each weighted decision vector and the cluster center. Each weighted decision vector is assigned to the cluster represented by the nearest cluster center.

[0186] (3) For each cluster, calculate the mean of all weight decision vectors in the cluster and use the mean as the new cluster center.

[0187] New cluster centers for:

[0188] ;

[0189] in, express To classify into categories (i.e. clusters) The weighted decision vector of , L is the number of weighted decision vectors classified into this category.

[0190] (4) Iterate the clustering operation until the convergence distance of each cluster center is less than the predetermined convergence parameter M or reaches the predetermined number of iterations V, and then terminate the clustering operation.

[0191] If the number of iterations is equal to V, the clustering is terminated; otherwise, the convergence distance of each cluster in the iteration is calculated. If the convergence distance is less than the given parameter M, the iteration is terminated, otherwise the iteration continues.

[0192] The calculation formula for the mth iteration convergence distance is:

[0193] ;

[0194] (5) Based on the number of weight decision vectors contained in each cluster, the weight of the weight decision vector of each cluster is calculated.

[0195] The weight of the weight decision vector within each cluster is calculated as follows:

[0196] ;

[0197] in, is the weight of the lth cluster, is the number of weight decision vectors in the lth cluster, k is the number of clusters, is the number of all weight decision vectors in k clusters.

[0198] (6) Based on the weight of the weighted decision vector within each cluster, calculate the weight of the situational safety index.

[0199] The weight of the situational safety index is calculated as follows:

[0200] ;

[0201] in, is the weight of the i-th situational safety indicator, is the weight decision result of the i-th situational safety indicator in the weight decision vector corresponding to the cluster center of the l-th cluster.

[0202] Step S3, includes steps S31-S33, specifically.

[0203] Step S31, performing weighted summation on the situation safety index value at each time point based on the weight to obtain the situation awareness result at each time point.

[0204] The situation awareness results at each time point are calculated as follows:

[0205] ;

[0206] Among them, S, are the situation awareness result at each time point, the weight of the i-th situation safety index, and the situation safety index value.

[0207] Step S32: marking the situation safety index value at each time point according to the situation awareness result.

[0208] According to the situation awareness results, the situation security index values ​​at each time point are marked, including:

[0209] The security situation awareness results are judged according to the preset threshold to obtain the situation security level;

[0210] The situational safety level includes four safety levels: normal, recovery, critical and emergency;

[0211] The situation security index value is marked according to the situation security level corresponding to the situation security index value at each time point.

[0212] Based on the security situation assessment value of the integrated energy system, the security level of the integrated energy situation is divided into four levels: normal, recovery, critical, and emergency, as shown in Table 1.

[0213] Normal: After a disturbance, the power cyber-physical system that is already in a normal state can still meet all equations and inequality constraints of the operating state of key equipment. And the changes of all key parameters are within the allowable deviation;

[0214] Recovery: The state of a system after being disturbed for a short period of time, in which the system can recover by itself using centralized or distributed control methods.

[0215] Critical: There are problems such as energy quality fluctuations and stability degradation. At this time, the system operates under critical operating constraints and requires manual adjustment and recovery of the system using stable control measures;

[0216] Emergency: In this state, some key parameters in the integrated energy system cannot meet the constraints of safe operation. In this state, it is necessary to remove the faulty energy equipment or use the adjustment of controllable flexible energy equipment to maintain the stability of the entire system.

[0217] Table 1 Classification of security levels of integrated energy system situation

[0218]

[0219] For example,

[0220] If the situation awareness result is in the range of [0,0.25], it is marked as 0, indicating normal;

[0221] If the situation awareness result is in the range of [0.25, 0.5], it is marked as 1, indicating recovery;

[0222] If the situation awareness result is in the range of [0.5, 0.75], it is marked as 2, indicating critical;

[0223] If the situation awareness result is in the range of [0.75, 1.0], it is marked as 3, indicating emergency.

[0224] According to the situation awareness results, the situation safety index values ​​at each time point are marked to form a training sample set.

[0225] Step S33, using the training sample set, training the convolutional neural network to obtain a trained integrated energy system situation awareness model.

[0226] The relevant parameters of the convolutional neural network used in the present invention are shown in Table 2.

[0227] Table 2 Convolutional neural network model related parameters

[0228]

[0229] (1) Input layer: a sample set of station-level multivariate heterogeneous data for each energy site;

[0230] (2) Convolutional layer 1: The convolution kernel size is 10×10, 64 neurons are set, and the ReLU (Rectified Linear Unit) activation function is used;

[0231] (4) Pooling layer 1: Use the maximum pooling operation with a pooling window size of 1×1 to reduce the size of the feature map and retain the main features;

[0232] (5) Convolutional layer 2: The convolution kernel size is 10×10, 128 neurons are set, and the ReLU activation function is used; it further extracts high-level abstract features so that the network can understand the input data more deeply;

[0233] (6) Pooling layer 2: Use the maximum pooling operation with a pooling window size of 2×2; process the feature map generated by convolution layer 2 through downsampling and feature selection to further extract abstract features;

[0234] (7) Fully connected layer: Set 64×4 neurons and use the ReLU activation function; flatten the features extracted by the convolutional layer and connect them to the fully connected layer to further learn the abstract features of the data; Dropout is 0.5, which means that neurons are randomly dropped with a probability of 50% during training to prevent overfitting;

[0235] (8) Upsampling layer: Set 64 neurons and use the tanh activation function to map the learned features back to the original data space through upsampling operations;

[0236] (9) Multilayer perceptron layer: 256×4 neurons are set, and ReLU activation function is used. Dropout is 0.5. Multilayer perceptron layer is introduced for higher-level abstraction and feature learning.

[0237] The weights are continuously optimized through the back-propagation algorithm, and the model performance is improved by adjusting parameters such as learning rate and optimizer.

[0238] The comprehensive energy system situation awareness model outputs a security situation awareness result of the comprehensive energy system.

[0239] Step S4, specifically.

[0240] The electrical parameter data and system parameter data of each energy site collected dynamically in real time are used to calculate the situational safety index value, which is input into the comprehensive energy system situation awareness model to obtain the safety situation awareness results of the comprehensive energy system.

[0241] In this step, the situation security index value is calculated by referring to step S1; the weight is calculated by referring to step S1, which will not be repeated here.

[0242] Conduct situational awareness of the integrated energy system in real time, and dynamically and real-time feedback the security situational awareness results of the integrated energy system to relevant operation or maintenance personnel or monitoring systems to support timely decision-making and processing.

[0243] In order to verify the situation awareness method of a comprehensive energy system provided in an embodiment of the present invention, a simulation model is provided for verification:

[0244] The data in this paper is combined with the actual historical load data and measured environmental data such as temperature, wind speed, and light intensity from July 1 to December 31, 2018 in a certain area in the east, and the analysis is carried out based on the establishment of a comprehensive energy system simulation model as an example.

[0245] (1) Calculation of weights of security indicators of the comprehensive energy system situation

[0246] The weight coefficient of each indicator is calculated based on the fuzzy hierarchical clustering method. This paper obtains the fuzzy complementary matrix through expert evaluation, determines the weight coefficient of each indicator, and performs cluster analysis on experts. The weights of each situation security indicator are shown in Table 3.

[0247] Table 3 Weight values ​​of comprehensive energy system security situation indicators

[0248]

[0249] (2) Situational Awareness Computing

[0250] The status at an hour on a certain day is selected for situation awareness. According to the situation awareness results, the situation warning level is divided into 4 levels. The situation safety index values ​​are shown in Table 4.

[0251] Table 4 Comprehensive energy system situation security index values

[0252]

[0253] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0254] 1. The present invention provides a situation awareness method for an integrated energy system, which fully considers the characteristics of the operation of a multi-station system in an integrated energy system, obtains the weight decision vector of each expert on the situation safety index by introducing the fuzzy hierarchical analysis method, and combines the expert decision results to perform cluster analysis on the weight decision vector of the situation safety index using the k-means algorithm to accurately calculate the weight of each situation safety index;

[0255] 2. The convolutional neural network model is used to conduct a comprehensive analysis of the electrical parameter data and system parameter data of the integrated energy system collected in real time and dynamically, highlighting the comprehensive perception and scientific evaluation of the system security situation; the present invention provides a new, more accurate and reliable real-time situation perception method for the safety management of the integrated energy system;

[0256] 3. Determine the situational safety indicators of the integrated energy system, providing a comprehensive assessment framework that can cover all aspects of the integrated energy system, including static and dynamic safety, real-time monitoring of the system operation status, and timely perception of system problems and risks, which will help formulate safety strategies and measures, support the decision-making process, and prevent potential accidents.

[0257] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0258] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A situation awareness method for an integrated energy system, characterized in that: The steps include: Determine the situation safety index, collect the electrical parameter data and system parameter data of each energy site in the integrated energy system within the historical time period, and calculate the situation safety index value at each time point; The fuzzy hierarchical analysis method is used to obtain the weight decision vector of each expert on the situation safety index, and the weight decision vector is clustered to obtain the weight of each situation safety index; Based on the weights, weighted summation of the situation safety index values ​​at each time point is performed to obtain the situation awareness results at each time point, and the situation safety index values ​​at each time point are marked according to the situation awareness results to form a training sample set, and the convolutional neural network is trained to obtain a trained comprehensive energy system situation awareness model; According to the situation awareness results, the situation security index values ​​at each time point are marked, including: The security situation awareness results are judged according to the preset threshold to obtain the situation security level; The situational safety level includes four safety levels: normal, recovery, critical and emergency; Marking the situation security index value according to the situation security level corresponding to the situation security index value at each time point; If the situation awareness result is in the range of [0,0.25], it is marked as 0, indicating normal; If the situation awareness result is in the range of [0.25, 0.5], it is marked as 1, indicating recovery; If the situation awareness result is in the range of [0.5, 0.75], it is marked as 2, indicating critical; If the situation awareness result is in the range of [0.75, 1.0], it is marked as 3, indicating emergency; The situation awareness results at each time point are calculated as follows: Among them, S, ω i 、V i are the situation awareness result at each time point, the weight of the i-th situation safety index, and the situation safety index value; The convolutional neural network includes an input layer, a convolutional layer 1, a pooling layer 1, a convolutional layer 2, a pooling layer 2, a fully connected layer, an upsampling layer and a multi-layer perceptron layer in sequence; Based on the electrical parameter data and system parameter data of each energy site collected dynamically in real time, the situational safety index value is calculated and input into the comprehensive energy system situation awareness model to obtain the safety situation awareness result of the comprehensive energy system; The situational safety index includes the power grid flow distribution index F C , Functional quality qualification rate E SQ , Transient absolute safety margin K P , transient comprehensive safety margin K V , the grid-connected power quality of new energy points, the grid-connected node voltage qualification rate μ, the grid-connected power quality of new energy points, the grid-connected node current qualification rate σ, the grid-connected power quality of new energy points, the grid-connected node frequency qualification rate f, and the system load rate L f 、System overload level L g 、 Steady-state total adequacy A E , Steady-state component margin A e and transformer load factor L b ; The power grid flow distribution index F C ,as follows: Among them, F j，max is the maximum transmission capacity of line j in the system; n is the total number of system lines; F j is the active power flow of system line j, which is the actual power transmitted in the power system; The functional quality qualification rate E SQ ,as follows: Among them, q w is a 0-1 variable of energy m; N m is the number of output energy types; e, h, g, w, l represent electricity, heat, gas, wind, and light energy flows respectively; q i is a 0-1 variable for energy m and mass item i; esq i is the actual value of energy m and mass item i; Q m is the set of main mass items of energy m; The grid-connected power quality and grid-connected node voltage qualification rate μ of the new energy point is as follows: Among them, V is the total number of nodes connected to the grid of new energy points; The grid-connected power quality of the new energy point and the grid-connected node current qualification rate σ are as follows: Among them, A is the total number of nodes connected to the grid of new energy points; The grid-connected power quality and grid-connected node frequency qualification rate f of the new energy point is as follows: Among them, H is the total number of nodes connected to the grid of new energy points; The system load factor L f ,as follows: in, is the maximum transmission capacity allowed by the system line and, Transmit power for system lines and; The system overload level L g ,as follows: Where m is the number of overloaded lines in the system, and n is the total number of lines in the system; The steady-state total adequacy A E ,as follows: Among them, E out is the total output energy of the energy quality coefficient; E load is the total load of the energy quality coefficient; The steady-state component margin A e ,as follows: Among them, m e，i is the matching degree of output energy i, E(m e，i ) is m e，i The expected value, E out，i 、E load，i are the supply and demand of output energy i respectively; The transformer load factor L b ,as follows: Among them, F b is the active power output by transformer b in the system, S b is the rated capacity of transformer b, n0 is the number of transformers in the system; The comprehensive energy system includes energy storage stations, charging and swapping stations, substations, new energy stations, data center stations, 5G base stations and Beidou navigation stations; The method of using fuzzy analytic hierarchy process to obtain the weight decision vector of each expert on the situation safety index includes: Each expert compares the situation safety indicators in pairs to obtain fuzzy judgment results, and constructs a fuzzy complementary matrix based on the fuzzy judgment results, where each element value (i, j) of the matrix is ​​the fuzzy judgment result between situation safety indicators i and j; Based on the element values ​​of the fuzzy complementary matrix, the weighted decision results u of each expert for each situational safety index are calculated. Zi , where i = 1, 2, ..., n, and the weighted decision vector U of each expert on the situational safety index is obtained Z =(u Z1 ,u Z2 ,......,u Zn ), where Z = 1…N, and N represents the number of experts; The fuzzy complementary matrix is ​​as follows: F=(f ij ) n×n ; Among them, n is the number of situational safety indicators, f ij ∈[0, 1], 1≤i, j≤n, the element values ​​on the matrix diagonal are 1.0, and the element values ​​at other positions are 0.1~0.

9.

2. The method according to claim 1, characterized in that Based on the element values ​​of the fuzzy complementary matrix, the weight decision results of each expert on each situational safety index are calculated as follows: Among them, u Zi is the weight decision result of the expert group on indicator i, and any two indicators m and j among the n indicators satisfy f mj =log α u Zm -log α u Zj +0.5, α is the decision maker’s discrimination ability.

3. The method according to claim 2, characterized in that The weights of each situational safety index obtained by clustering analysis of the weight decision vector include: Performing cluster analysis on the weight decision vector using a k-means algorithm; Select the number of clusters k, randomly select k experts’ weighted decision vectors for situational safety indicators as the initial clustering centers {p1, p2, …, p i ,…,p k }; Among them, p i ={p i1 , p i2 , ..., p i10 } is the weight decision vector of the i-th cluster center; Calculate the Euclidean distance from each weighted decision vector to the cluster center of each cluster, and assign the weighted decision vector to the cluster represented by the nearest cluster center; For each cluster, the mean of all weight decision vectors in the cluster is calculated, and the mean is used as the new cluster center. The clustering operation is terminated until the convergence distance of each cluster center is less than the predetermined convergence parameter or the predetermined number of iterations is reached. Based on the number of weighted decision vectors contained in each cluster, the weight of the weighted decision vector of each cluster is calculated; based on the weight of the weighted decision vector of each cluster and the cluster center of each cluster, the weight of each situation safety index is calculated.

4. The method according to claim 3, characterized in that The weight of the weight decision vector of each cluster is calculated as follows: Among them, η l is the weight of the lth cluster, n l is the number of weight decision vectors in the lth cluster, k is the number of clusters, is the number of all weight decision vectors in k clusters.

5. The method according to claim 4, characterized in that The weight of the situational safety index is calculated as follows: Among them, ω i is the weight of the i-th situational safety indicator, p li is the weight decision result of the i-th situational safety indicator in the weight decision vector corresponding to the cluster center of the l-th cluster.

Images