System configuration method and device based on probability prediction and credible regulation potential quantification

Abstract

The invention relates to a system configuration method and device based on probability prediction and credible regulation potential quantification, and the method comprises the steps: inputting the source load data of a typical scene into a trained QRLSTM model, obtaining a source load probability prediction result, describing the time sequence correlation of photovoltaic and load through a small number of scenes, and achieving the effective quantification of prediction uncertainty; flexible resources of the heat supply network and the cold supply network are aggregated in sequence by adopting a robust optimization algorithm, dispatching decoupling of thermoelectricity, heat cooling and cold electricity is realized, the maximum upward / downward adjustable potential and credible adjustment potential of the heat supply network and the cold supply network to the power grid are accurately quantified, and the problem of strong information barriers among energy systems is solved; according to the method, the source load probability prediction and the flexible resource credible adjustment potential are taken as boundaries, the cost minimization is taken as a target, the double-layer multi-scene collaborative optimization configuration model is constructed, the uncertainty of the source load is accurately described, the adjustment capability of the flexible resource is quantified, and the economy and the flexibility of planning operation of the integrated energy system are integrally improved.

Description

Technical Field

[0001] The present invention relates to the technical field of optimized configuration of electric-heat-cold integrated energy systems, and in particular to a system configuration method and device based on probability prediction and quantification of credible regulation potential. Background Art

[0002] With the rapid development of integrated energy systems, the system covers multiple energy forms such as electricity, heat, and cooling, forming a complex and highly coupled network. In actual dispatching and operation, it is particularly difficult to obtain the global parameters of the electricity-heat-cooling network at the same time due to the information barriers between the energy systems. In addition, the large-scale access to new energy sources such as photovoltaics has significantly increased the uncertainty within the system and put forward higher requirements for the flexibility of the system. This increase in uncertainty and flexibility has brought huge challenges to the safe and stable operation of the integrated energy system.

[0003] At present, the deterministic short-term prediction method is mainly used for source load prediction. This method can provide prediction information to a certain extent, but it is difficult to fully reflect the changing characteristics and uncertainty of source load. In addition, the deterministic short-term prediction method mainly serves the daily operation level of the system and is difficult to be directly applied to the long-term planning scenario of the integrated energy system. At the same time, most of the research on the flexibility of the integrated energy system is based on deterministic prediction results, lacking in-depth research on the uncertainty of source load prediction, which makes it difficult to accurately quantify the credible regulation potential of multi-energy synergy.

[0004] Therefore, how to carry out source-load probability prediction and quantitative research on the credible regulation potential of multi-energy coordination in the medium and long term has become an urgent problem to be solved in the planning of integrated energy systems. Summary of the invention

[0005] In view of this, it is necessary to provide a system configuration method and device based on probability prediction and quantification of credible regulation potential to solve the above-mentioned defects of the prior art.

[0006] In order to solve the above problems, in a first aspect, an embodiment of the present invention provides a system configuration method based on probability prediction and quantification of credible adjustment potential, including:

[0007] Collect historical data of source load and corresponding meteorological data, and perform data preprocessing;

[0008] Build a QRLSTM model that combines quantile regression and long short-term memory neural network, and use the preprocessed data to train and test the model;

[0009] Generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result;

[0010] Identify flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling;

[0011] Based on the integrated energy system network model of electricity-heat-cold, a robust optimization algorithm is used to aggregate the flexibility resources of the heat network and the cold network, realize the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid;

[0012] With the goal of minimizing the cost of the integrated energy system, a two-layer multi-scenario collaborative optimization configuration model is constructed;

[0013] According to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, the double-layer optimization configuration model is solved and the equipment configuration results of the integrated energy system are output.

[0014] Furthermore, the data preprocessing includes:

[0015] Data cleaning is performed on the source load historical data and the corresponding meteorological data: the Z-score algorithm is used to detect outliers in the data and the detected outliers are replaced with zero; the linear interpolation method is used to fill in the missing values.

[0016] Furthermore, the establishment of a QRLSTM model combining quantile regression and long short-term memory neural network, and using the preprocessed data to train and test the model specifically includes:

[0017] In the Tensorflow framework, the QRLSTM model is constructed by combining quantile regression and long short-term memory neural network;

[0018] The preprocessed data samples are divided into training set and test set, and the quantile loss function is defined. The goal is to minimize the quantile loss function, and the QRLSTM model is trained using the training set.

[0019] The test set is used to test whether the probability prediction accuracy of the trained QRLSTM model meets the preset accuracy requirements. If so, the model training is completed; if not, the above steps are repeated until the preset accuracy requirements are met.

[0020] Furthermore, the method generates a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, inputs the source-load data of the typical scenario into the trained QRLSTM model, and obtains the source-load probability prediction result, which specifically includes:

[0021] By characterizing the temporal correlation between sources and loads, a photovoltaic-load joint scenario is generated;

[0022] The number of typical scenes is determined based on the intra-cluster error quadratic sum method and the silhouette coefficient method, and the typical scenes and their corresponding probabilities are determined based on the DTW dynamic time warping algorithm;

[0023] The source-load data corresponding to the typical scenario is input into the trained QRLSTM model to obtain the source-load probability prediction results.

[0024] Furthermore, the determination of the flexibility resources in the integrated energy system and the establishment of an integrated energy system network model including electricity, heat and cooling specifically include:

[0025] The flexibility resources in the integrated energy system include CCHP combined cooling, heating and power units, GB gas boiler equipment, HS thermal energy storage equipment, EC electric refrigerator equipment, CS cold energy storage equipment, PV photovoltaic units, and ES electric energy storage equipment;

[0026] The heat network and cold network models are established by mass regulation. The water supply flow in the model is constant, and the flow into the node is equal to the flow out of the node. For the heat network, the heat provided by the heating equipment enters the thermal system through the heat exchange station. The expression of this process is:

[0027]

[0028] In the formula, and are the heating powers of CCHP and GB at time t, and are the heat release and charging power of HS at time t, c w is the specific heat capacity of water, is the flow injected into node k at time t, and are the heat source heating temperature and return water network temperature of node k at time t respectively;

[0029] The expression for the dynamic change of water supply network temperature is:

[0030]

[0031] In the formula, and are the average temperatures of the water supply pipe and the return pipe p at time t, and are the outlet temperature and inlet temperature of the water supply pipe p at time t, and are the outlet temperature and inlet temperature of the return pipe p at time t, Δt is the time interval, S p , L p and λ p are the cross-sectional area, length and heat transfer coefficient of the pipe p, T tW is the ambient temperature of the pipe, m p,t is the flow rate of pipe p at time t;

[0032] When hot water flows from multiple pipes to a node or from a node to multiple pipes, the mixed temperature of the node should satisfy the node temperature constraint. The expression of the node mixed temperature is:

[0033]

[0034] In the formula, is the mixed water temperature of node k in the water supply network at time t, and are the sets of pipes in the water supply / return network with node k as the end and start point respectively;

[0035] The expression of heat exchange process on the load side is:

[0036]

[0037] In the formula, H k,t is the heat load of node k at time t, is the regenerative temperature of node k at time t;

[0038] For the cold network, the cold air provided by the cooling equipment enters the cooling system through the cooling station. The expression of this process is:

[0039]

[0040] In the formula, and are the cooling powers of CCHP and EC at time t, and are the cooling and charging powers of CS at time t, and are the cooling source supply temperature and return water network temperature of node k at time t respectively;

[0041] The modeling process of the cold network is similar to that of the hot network, so it will not be repeated here. The expression of the cold exchange process of the load side cold network is:

[0042]

[0043] In the formula, C k,t is the cooling load of node k at time t, is the cooling temperature of node k at time t.

[0044] Furthermore, the integrated energy system network model based on electricity, heat and cold adopts a robust optimization algorithm to aggregate the flexibility resources of the heat network and the cold network, realizes the scheduling decoupling of heat and electricity, heat and cold, and cold and electricity, and quantifies the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid, specifically including:

[0045] A robust optimization algorithm is used to establish a heat network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated CCHP output under the worst scenario:

[0046]

[0047] In the formula, and H CCHP are the vectors consisting of the upper and lower limits of CCHP output at each moment, ξ t represents the CCHP output uncertainty at time t, U is the set of CCHP output uncertainties at each time; x(ξ t ) is the vector of other decision variables except CCHP output at time t, including GB output, HS charging power, heat release power and HS charging and discharging state;

[0048] Solve the heat network flexibility aggregation model to obtain the optimization result after heat network power aggregation Using CCHP electrical-thermal coupling relationship Decoupling to obtain the regulation potential range provided by the heat network to the power grid Using CCHP heat-cold coupling relationship Decoupling to obtain the flexible power support range provided by the heating network to the cooling network

[0049] A robust optimization algorithm is used to establish a cold network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated EC output in the worst scenario:

[0050]

[0051] In the formula, and C EC are the vectors consisting of the feasible upper and lower bounds of EC output at each moment, ζ t represents the uncertainty of EC output at time t, D is the set of EC output uncertainties at each time; x(ζ t ) is a vector of decision variables other than the EC output at time t, including the CCHP output after heat-cold decoupling, the CS charging and cooling power, the CS cooling power, and the CS charging and discharging state;

[0052] Solve the cold network flexibility aggregation model to obtain the optimization result after cold network power aggregation Using EC electrical-cold coupling relationship Decoupling to obtain the power regulation range of the cooling network

[0053] The flexibility resources in the integrated energy system can provide upward / downward flexible adjustment capabilities, which can be expressed as:

[0054]

[0055]

[0056] In the formula, F t CCHP,up 、F t ES,up 、F t EC,up and F t PV,up are the upward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, respectively, t CCHP,dn 、F t ES,dn 、F t EC,dn and F t PV,dn are the downward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, r CCHP and r EC are the ramp rates of CCHP and EC respectively, and are the charging efficiency and discharging efficiency of ES respectively, and are the maximum charging power and maximum discharging power of ES, E ES is the capacity of ES, SOC t is the energy storage state of ES at time t, and SOC ES are the maximum energy storage state and the minimum energy storage state of ES, respectively, t PV is the PV output at time t, is the maximum output of PV at time t, is the minimum output of PV at time t;

[0057] The flexibility resources in the integrated energy system can provide the maximum upward adjustable potential F t m,up The expression is:

[0058] F tm,up =F t CCHP,up +F t ES,up +F t PV,up +F t EC,up

[0059] The flexibility resources in the integrated energy system can provide the maximum downward adjustable potential F t m,dn The expression is:

[0060] F t m,dn =F t CCHP,dn +F t ES,dn +F t PV,dn +F t EC,dn

[0061] Since photovoltaic power is an uncontrollable power source, its output is volatile and uncontrollable. Flexible resources in the integrated energy system can provide upward credible regulation potential. t b,up The expression is:

[0062] F t b,up =F t CCHP,up +F t ES,up -F t PV,dn +F t EC,up

[0063] Flexible resources in an integrated energy system can provide credible downward regulation potential. t b,dn The expression is:

[0064] F t b,dn =F t CCHP,dn +F t ES,dn -F t PV,up +F t EC,dn .

[0065] Furthermore, the two-layer multi-scenario collaborative optimization configuration model includes an upper layer model and a lower layer model;

[0066] The upper model aims to minimize the total system investment cost and annual operating cost, and the upper objective function is:

[0067] min(C inv +C ope )

[0068] In the formula, C inv is the total investment cost, C ope is the annual operating cost of the system;

[0069] The lower model aims to minimize the annual operating cost of the system, and the objective function of the lower layer is:

[0070] minC ope

[0071] The two-layer multi-scenario collaborative optimization configuration model consisting of the upper model and the lower model is:

[0072]

[0073] In the formula, s is a typical scene, S is a typical scene set, π s is the dual variable of scenario s;

[0074] Total investment cost of integrated energy system C inv The expression is:

[0075]

[0076] Where j is the jth device among CCHP, GB, EC, ES, HS, CS and PV, k j is the discounted years of the jth equipment, M j is the installed capacity of the jth type of equipment, is the unit capacity investment cost of the jth type of equipment;

[0077] System annual operating cost C ope The expression is:

[0078]

[0079] In the formula, ρ s is the probability of a typical scenario s, and are the system electricity purchase cost, gas purchase cost and operation and maintenance cost of flexible resources under scenario s, is the penalty cost of insufficient system flexibility under scenario s, a and b are the weight coefficients of economic cost and penalty cost of insufficient flexibility respectively;

[0080] The penalty cost of insufficient system flexibility is expressed as:

[0081]

[0082] In the formula, F t def,up 、F t def,dn and F t def They are the upward, downward and total flexibility of the system at time t, respectively, t load,up and F t load,dn are the upward and downward regulation requirements of the electric load at time t, c def is the penalty cost coefficient.

[0083] In a second aspect, an embodiment of the present invention provides a system configuration device based on probability prediction and quantification of credible adjustment potential, including:

[0084] Data collection and preprocessing module, used to collect source load historical data and corresponding meteorological data, and perform data preprocessing;

[0085] Model training module, used to build a QRLSTM model combining quantile regression and long short-term memory neural network, and use preprocessed data to train and test the model;

[0086] The source-load probability prediction module is used to generate typical light-load joint time series scenarios based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenarios into the trained QRLSTM model, and obtain the source-load probability prediction results;

[0087] System modeling module, used to determine the flexibility resources in the integrated energy system and establish a network model of the integrated energy system including electricity, heat and cooling;

[0088] The dispatch decoupling module is used for the integrated energy system network model based on electricity-heat-cold. It uses a robust optimization algorithm to aggregate the flexibility resources of the heat network and the cold network, realizes the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantifies the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid;

[0089] Cost optimization module, which is used to build a two-layer multi-scenario collaborative optimization configuration model with the goal of minimizing the cost of the integrated energy system;

[0090] The collaborative optimization configuration module is used to solve the double-layer optimization configuration model according to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid, and preset constraints, and output the equipment configuration results of the integrated energy system.

[0091] In a third aspect, the present invention further provides an electronic device, including a memory and a processor, wherein:

[0092] The memory is used to store programs;

[0093] The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the system configuration method based on probability prediction and quantification of credible adjustment potential as described in the embodiment of the first aspect of the present invention.

[0094] In a fourth aspect, the present invention further provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the system configuration method based on probability prediction and quantification of reliable adjustment potential as described in the embodiment of the first aspect of the present invention.

[0095] The present invention generates typical light-load joint time series scenarios based on the DTW algorithm, inputs the source-load data of the typical scenarios into the trained QRLSTM model, and obtains the source-load probability prediction results. It can not only describe the time series correlation between photovoltaic and load through a small number of scenarios, but also realize the effective quantification of prediction uncertainty, and provide key data support for the analysis and decision-making of the integrated energy system; the robust optimization algorithm is used to successively aggregate the flexibility resources of the heating network and the cold network, realize the dispatch decoupling of thermal power, heat and cold, and cold power, and accurately quantify the maximum adjustable potential and credible regulation potential of the heating network and the cold network to the power grid upward / downward, which can not only overcome the problem of strong information barriers between energy systems, but also quantify the credible regulation potential of multi-energy collaboration, and take into account the economy and flexibility of system operation; the present invention takes the source-load probability prediction and the credible regulation potential of flexible resources as the boundaries, and constructs a double-layer multi-scenario collaborative optimization configuration model with the goal of minimizing costs, accurately describes the uncertainty of source and load, quantifies the regulation capacity of flexible resources, and improves the economy and flexibility of the planning and operation of the integrated energy system as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 A flow chart of a system configuration method based on probability prediction and quantification of credible regulation potential provided by the present invention;

[0097] Figure 2 A schematic diagram of load probability prediction results under different confidence levels of the QRLSTM model provided by the present invention;

[0098] Figure 3 A diagram of a comprehensive energy system model provided by the present invention;

[0099] Figure 4 A structural block diagram of a system configuration device based on probability prediction and quantification of credible regulation potential provided by the present invention;

[0100] Figure 5 This is a structural block diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0101] 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.

[0102] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0103] At present, the deterministic short-term prediction method is mainly used for source load prediction. This method can provide prediction information to a certain extent, but it is difficult to fully reflect the changing characteristics and uncertainty of source load. In addition, the deterministic short-term prediction method mainly serves the daily operation level of the system and is difficult to be directly applied to the long-term planning scenario of the integrated energy system. At the same time, most of the research on the flexibility of the integrated energy system is based on deterministic prediction results, lacking in-depth research on the uncertainty of source load prediction, which makes it difficult to accurately quantify the credible regulation potential of multi-energy synergy.

[0104] In view of this, the present invention provides a system configuration method based on probability prediction and quantification of credible regulation potential, with source-load probability prediction and flexible resource credible regulation potential as the boundary, and a two-layer multi-scenario collaborative optimization configuration model is constructed with the goal of minimizing cost, which accurately depicts the uncertainty of source load, quantifies the regulation capacity of flexible resources, and improves the economy and flexibility of integrated energy system planning and operation as a whole. The following will be explained and introduced through multiple embodiments.

[0105] Figure 1 The flowchart of the system configuration method based on probability prediction and quantification of credible regulation potential provided by the present invention. Figure 1 As shown, the system configuration method based on probability prediction and quantification of credible regulation potential includes at least the following steps:

[0106] Collect historical data of source load and corresponding meteorological data, and perform data preprocessing.

[0107] In this embodiment, "source load" refers to the general term for energy supply (source) and energy demand (load). Here, energy supply mainly refers to equipment or systems that can provide electricity, heat or cold, such as photovoltaic units (PV), combined cooling, heating and power units (CCHP), etc.; energy demand refers to various loads, such as power load, heat load and cooling load.

[0108] Specifically, first, collect the historical data of source load and the corresponding meteorological data. The historical data of source load is the basis for establishing the prediction model. By collecting historical data, we can understand the changing rules and trends of source load and provide data support for subsequent prediction and configuration. Meteorological data has a significant impact on source load. For example, meteorological factors such as light intensity and temperature will directly affect the power generation and load demand of photovoltaic units. Therefore, collecting the corresponding meteorological data and combining it with the historical data of source load can improve the accuracy and reliability of the prediction model.

[0109] Then, the historical data of source load and the corresponding meteorological data are cleaned: the Z-score algorithm is used to detect outliers in the data and the detected outliers are replaced with zero; the linear interpolation method is used to fill the missing values. It is understandable that the original data may have problems such as outliers and missing values, which will affect the performance and accuracy of the prediction model. Through data preprocessing, such as cleaning outliers and filling missing values, the quality and availability of the data can be improved, thereby further improving the accuracy and stability of the prediction model.

[0110] A QRLSTM model combining quantile regression and long short-term memory neural network is established, and the model is trained and tested using the preprocessed data.

[0111] In a preferred embodiment of the present invention, the following steps may be specifically included:

[0112] In the Tensorflow framework, the QRLSTM model is constructed by combining quantile regression and long short-term memory neural network.

[0113] Here, Tensorflow is a widely used deep learning framework that provides a rich set of neural network layers and optimization algorithms. Quantile regression is a statistical method used to estimate the value of conditional quantiles, such as conditional medians or conditional quartiles. In the QRLSTM model, the quantile regression layer is used to output the probability distribution of the source load data, that is, the predicted values ​​at different quantiles. The Long Short-Term Memory (LSTM) neural network has the ability to memorize long sequence information. In the QRLSTM model, the LSTM neural network is used to capture the time series characteristics of the source load data and pass the historical information to the quantile regression layer. By combining the quantile regression layer and the LSTM neural network, the QRLSTM model can simultaneously consider the probability distribution and time series characteristics of the source load data, thereby improving the accuracy of the prediction.

[0114] The preprocessed data samples are divided into training set and test set, and the quantile loss function is defined. The training set is used to train the QRLSTM model with the goal of minimizing the quantile loss function.

[0115] In this embodiment, the preprocessed data samples are divided into a training set and a test set. The training set is used to train the model, that is, to adjust the parameters of the model to minimize the loss function; the test set is used to test the model, that is, to evaluate the performance of the model. The quantile loss function is a function that measures the difference between the model's predicted value and the actual value. In the QRLSTM model, the quantile loss function is used to measure the difference between the model's predicted value at different quantiles and the actual source load data. By minimizing the quantile loss function, the model can learn the probability distribution characteristics of the source load data and output more accurate prediction values.

[0116] The test set is used to test whether the probability prediction accuracy of the trained QRLSTM model meets the preset accuracy requirements. If so, the model training is completed; if not, the above steps are repeated until the preset accuracy requirements are met.

[0117] Specifically, after the training is completed, a test set is needed to evaluate the performance of the model. In this embodiment, the performance evaluation of the model is achieved by calculating the difference between the predicted value and the actual value of the QRLSTM model on the test set. If the prediction accuracy of the model meets the accuracy requirements (such as the mean square error, mean absolute error and other indicators reach the preset threshold), it means that the model training is complete; otherwise, it is necessary to repeat the process described in the above steps, that is, readjust the model parameters or increase the training data, etc., to improve the prediction accuracy of the model.

[0118] In this embodiment, by combining quantile regression and LSTM neural network, the QRLSTM model can simultaneously consider the probability distribution and time series characteristics of the source load data, thereby improving the accuracy of the prediction. By dividing the data samples into training sets and test sets, and defining the quantile loss function for training, the QRLSTM model can learn the inherent laws and characteristics of the source load data and has good generalization ability.

[0119] Generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result;

[0120] In a preferred embodiment of the present invention, the following steps may be specifically included:

[0121] By characterizing the temporal correlation between source and load, a photovoltaic-load joint scenario is generated.

[0122] Specifically, source-load data can include photovoltaic output data and load data, which are temporally correlated. For example, photovoltaic output is higher during the day and lower at night, and load data may also show a similar daily variation pattern. By analyzing the temporal correlation of source-load data, a series of photovoltaic-load joint scenarios can be generated, which represent possible combinations of photovoltaic output and load at different time points.

[0123] The number of typical scenes is determined based on the intra-cluster error quadratic sum method and the silhouette coefficient method, and the typical scenes and their corresponding probabilities are determined based on the DTW dynamic time warping algorithm.

[0124] Specifically, the Sum of Squared Errors within Clusters (SSE) is an indicator to measure the clustering effect. It calculates the sum of the squares of the distances from the data points in each cluster to the centroid of the cluster. The smaller the SSE, the better the clustering effect.

[0125] The Silhouette Coefficient method is used to evaluate the rationality of clustering. It takes into account the compactness within the cluster and the separation between clusters. The higher the Silhouette Coefficient value, the better the clustering effect.

[0126] In this embodiment, by combining the intra-cluster error square sum method and the silhouette coefficient method, an appropriate number of typical scenarios can be determined so that these scenarios can fully represent the variation characteristics of the source load data while avoiding excessive scenarios leading to increased computational complexity.

[0127] Furthermore, the DTW (Dynamic Time Warping) algorithm is an algorithm for measuring the similarity between two time series, which allows the time series to be stretched and retracted on the time axis to find the best alignment. In this embodiment, the DTW algorithm is used to select the most representative typical scenarios from the generated photovoltaic-load joint scenarios. These typical scenarios can reflect the main change patterns and characteristics of source-load data. In addition, a corresponding probability is assigned to each typical scenario, which indicates the possibility of the scenario occurring in actual situations. This helps to consider the uncertainty of different scenarios in subsequent prediction and optimization processes.

[0128] The source-load data corresponding to the typical scenario is input into the trained QRLSTM model to obtain the source-load probability prediction results.

[0129] Specifically, a QRLSTM model combining quantile regression and long short-term memory neural network was established, and preprocessed data was used for training and testing. The source load data of the selected typical scenarios were input into the trained QRLSTM model, and the model would output the source load probability prediction results under these scenarios.

[0130] In this embodiment, the source-load data of typical scenarios are input into the trained QRLSTM model to obtain the source-load probability prediction results. This can not only describe the temporal correlation between photovoltaics and loads through a small number of scenarios, but also achieve effective quantification of prediction uncertainty, providing key data support for integrated energy system analysis and decision-making.

[0131] Figure 2 This is a schematic diagram of load probability prediction results under different confidence levels of the QRLSTM model provided by the present invention. Taking load as an example, the trained QRLSTM model is tested based on the test set data to obtain load probability prediction results under different confidence levels.

[0132] Identify the flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling.

[0133] Figure 3 This is a model diagram of the integrated energy system provided by the present invention. Figure 3 The integrated energy system includes the main power grid, gas grid, energy hub, network model, electric cooling and heating loads, among which the network model includes the cooling network, heating network and distribution network model. The energy hub contains a variety of flexibility resources, and the flexibility resources in the integrated energy system include CCHP combined cooling, heating and power units, GB gas boiler equipment, HS thermal energy storage equipment, EC electric refrigerator equipment, CS cold energy storage equipment, PV photovoltaic units, and ES electric energy storage equipment.

[0134] Specifically, the heat network and cold network models are established by mass regulation. In the model, the water supply flow is constant, and the flow into the node is equal to the flow out of the node. For the heat network, the heat provided by the heating equipment enters the thermal system through the heat exchange station. The expression of this process is:

[0135]

[0136] In the formula, and are the heating powers of CCHP and GB at time t, and are the heat release and charging power of HS at time t, c w is the specific heat capacity of water, is the flow injected into node k at time t, and are the heat source heating temperature and return water network temperature of node k at time t respectively;

[0137] In order to reflect the time delay of heat transfer from the heat source to the load side, a quasi-dynamic model is used to represent the temperature changes of the water supply and return water networks in the heat network. The expression of the dynamic change of the water supply network temperature is:

[0138]

[0139] In the formula, and are the average temperatures of the water supply pipe and the return pipe p at time t, and are the outlet temperature and inlet temperature of the water supply pipe p at time t, and are the outlet temperature and inlet temperature of the return pipe p at time t, Δt is the time interval, S p , L p and λ p are the cross-sectional area, length and heat transfer coefficient of the pipe p, T t W is the ambient temperature of the pipe, m p,t is the flow rate of pipe p at time t;

[0140] When hot water flows from multiple pipes to a node or from a node to multiple pipes, the mixed temperature of the node should satisfy the node temperature constraint. The expression of the node mixed temperature is:

[0141]

[0142] In the formula, is the mixed water temperature of node k in the water supply network at time t, and are the sets of pipes in the water supply / return network with node k as the end and start point respectively;

[0143] The expression of heat exchange process on the load side is:

[0144]

[0145] In the formula, H k,t is the heat load of node k at time t, is the heat recovery temperature of node k at time t; for the cold network, the cold air provided by the cooling equipment enters the cooling system through the cold exchange station. The expression of this process is:

[0146]

[0147] In the formula, and are the cooling powers of CCHP and EC at time t, and are the cooling and charging powers of CS at time t, and are the cooling source supply temperature and return water network temperature of node k at time t respectively;

[0148] The modeling process of the cold network is similar to that of the hot network, so it will not be repeated here. The expression of the cold exchange process of the load side cold network is:

[0149]

[0150] In the formula, C k,t is the cooling load of node k at time t, is the cooling temperature of node k at time t.

[0151] Based on the integrated energy system network model of electricity, heat and cooling, a robust optimization algorithm is used to aggregate the flexibility resources of the heating network and the cooling network, realize the scheduling decoupling of thermal power, heat and cooling, and cold power, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cooling network to the power grid.

[0152] Specifically, firstly, a robust optimization algorithm is used to establish a heat network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated CCHP output under the worst scenario:

[0153]

[0154] In the formula, and H CCHP are the vectors consisting of the upper and lower limits of CCHP output at each moment, ξ t represents the CCHP output uncertainty at time t, U is the set of CCHP output uncertainties at each time; x(ξ t ) is the vector of other decision variables except CCHP output at time t, including GB output, HS charging power, heat release power and HS charging and discharging state;

[0155] Then, the heat network flexibility aggregation model is solved to obtain the optimization result after heat network power aggregation. Using CCHP electrical-thermal coupling relationship Decoupling to obtain the regulation potential range provided by the heat network to the power grid Using CCHP heat-cold coupling relationship Decoupling to obtain the flexible power support range provided by the heating network to the cooling network

[0156] Next, a robust optimization algorithm is used to establish a cold network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated EC output under the worst scenario:

[0157]

[0158] In the formula, and C EC are the vectors consisting of the feasible upper and lower bounds of EC output at each moment, ζ t represents the uncertainty of EC output at time t, D is the set of EC output uncertainties at each time; x(ζ t ) is a vector of decision variables other than the EC output at time t, including the CCHP output after heat-cold decoupling, the CS charging and cooling power, the CS cooling power, and the CS charging and discharging state;

[0159] Solve the cold network flexibility aggregation model to obtain the optimization result after cold network power aggregation Using EC electrical-cold coupling relationship Decoupling to obtain the power regulation range of the cooling network

[0160] The flexibility resources in the integrated energy system can provide upward / downward flexible adjustment capabilities, which can be expressed as:

[0161]

[0162] In the formula, F t CCHP,up 、F t ES,up 、F t EC,up and F t PV,up are the upward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, respectively, t CCHP,dn 、F t ES,dn 、F t EC,dn and F t PV,dn are the downward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, r CCHP and r EC are the ramp rates of CCHP and EC respectively, and are the charging efficiency and discharging efficiency of ES respectively, and are the maximum charging power and maximum discharging power of ES, E ES is the capacity of ES, SOC t is the energy storage state of ES at time t, and SOC ESare the maximum energy storage state and the minimum energy storage state of ES, respectively, t PV is the PV output at time t, is the maximum output of PV at time t, is the minimum output of PV at time t;

[0163] The flexibility resources in the integrated energy system can provide the maximum upward adjustable potential F t m,up The expression is:

[0164] F t m,up =F t CCHP,up +F t ES,up +F t PV,up +F t EC,up

[0165] The flexibility resources in the integrated energy system can provide the maximum downward adjustable potential F t m,dn The expression is:

[0166] F t m,dn =F t CCHP,dn +F t ES,dn +F t PV,dn +F t EC,dn

[0167] Since photovoltaic power is an uncontrollable power source, its output is volatile and uncontrollable. Flexible resources in the integrated energy system can provide upward credible regulation potential. t b,up The expression is:

[0168] F t b,up =F t CCHP,up +F t ES,up -F t PV,dn +F t EC,up

[0169] Flexible resources in an integrated energy system can provide credible downward regulation potential. t b,dn The expression is:

[0170] Ft b,dn =F t CCHP,dn +F t ES,dn -F t PV,up +F t EC,dn .

[0171] The embodiment of the present invention adopts a robust optimization algorithm to successively aggregate the flexibility resources of the heating network and the cooling network, realizes the scheduling decoupling of thermal power, heat and cooling, and cold power, and accurately quantifies the maximum adjustable potential and reliable adjustment potential of the heating network and the cooling network to the power grid upward / downward, which can not only overcome the problem of strong information barriers between energy systems, but also quantify the reliable adjustment potential of multi-energy collaboration, taking into account the economy and flexibility of system operation.

[0172] With the goal of minimizing the cost of the integrated energy system, a two-layer, multi-scenario collaborative optimization configuration model is constructed.

[0173] Specifically, the two-layer multi-scenario collaborative optimization configuration model includes an upper-layer model and a lower-layer model.

[0174] Among them, the upper model aims to minimize the total investment cost and annual operating cost of the system, and the upper objective function is:

[0175] min(C inv +C ope )

[0176] In the formula, C inv is the total investment cost, C ope is the annual operating cost of the system;

[0177] The lower model aims to minimize the annual operating cost of the system, and the objective function of the lower layer is:

[0178] minC ope

[0179] The two-layer multi-scenario collaborative optimization configuration model consisting of the upper model and the lower model is:

[0180]

[0181] In the formula, s is a typical scene, S is a typical scene set, π s is the dual variable of scenario s;

[0182] Total investment cost of integrated energy system C inv The expression is:

[0183]

[0184] Where j is the jth device among CCHP, GB, EC, ES, HS, CS and PV, k j is the discounted years of the jth equipment, M j is the installed capacity of the jth type of equipment, is the unit capacity investment cost of the jth type of equipment;

[0185] System annual operating cost C ope The expression is:

[0186]

[0187] In the formula, ρ s is the probability of a typical scenario s, and are the system electricity purchase cost, gas purchase cost and operation and maintenance cost of flexible resources under scenario s, is the penalty cost of insufficient system flexibility under scenario s, a and b are the weight coefficients of economic cost and penalty cost of insufficient flexibility respectively;

[0188] The penalty cost of insufficient system flexibility is expressed as:

[0189]

[0190] In the formula, F t def,up 、F t def,dn and F t def They are the upward, downward and total flexibility of the system at time t, respectively, t load,up and F t load,dn are the upward and downward regulation requirements of the electric load at time t, c def is the penalty cost coefficient.

[0191] According to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, the double-layer optimization configuration model is solved and the equipment configuration results of the integrated energy system are output.

[0192] In this embodiment, the double-layer optimization configuration model is solved according to the system operating parameters (such as equipment efficiency, cost coefficient, etc.), the source-load probability prediction results of typical scenarios (i.e., the predicted values ​​of photovoltaic and load in different scenarios), the maximum power support range of the heating and cooling networks to the power grid (i.e., the maximum regulation potential that the heating and cooling networks can provide), and the preset constraints (such as equipment capacity limitations, operation limitations, etc.).

[0193] Specifically, for the upper model of the two-layer optimization configuration model, since the number of decisions in the upper model is relatively small, a real number coded genetic algorithm is used for optimization. This genetic algorithm can accurately encode the type and number of flexible resources, and find the optimal equipment configuration plan through iterative optimization. For the lower model, based on the type and number of equipment determined by the given upper model, the lower model calls the Gurobi solver to solve the mixed integer linear programming model. Gurobi is an efficient mathematical optimization solver that can quickly find the optimal operation strategy that meets the constraints. Finally, according to the solution results of the two-layer model, the configuration results of each device in the system are output, including the type, number and operation strategy of the equipment. By solving the two-layer optimization configuration model, the optimal configuration plan for each device in the integrated energy system can be obtained. This plan not only takes into account the investment cost and operating cost of the equipment, but also fully considers the impact of the volatility of photovoltaic and load and the flexibility of power supply equipment on system operation. This plan can improve the flexibility and reliability of the system while ensuring the economy of the system.

[0194] The present invention generates typical light-load joint time series scenarios based on the DTW algorithm, inputs the source-load data of the typical scenarios into the trained QRLSTM model, and obtains the source-load probability prediction results. It can not only describe the time series correlation between photovoltaic and load through a small number of scenarios, but also realize the effective quantification of prediction uncertainty, and provide key data support for the analysis and decision-making of the integrated energy system; the robust optimization algorithm is used to successively aggregate the flexibility resources of the heating network and the cold network, realize the dispatch decoupling of thermal power, heat and cold, and cold power, and accurately quantify the maximum adjustable potential and credible regulation potential of the heating network and the cold network to the power grid upward / downward, which can not only overcome the problem of strong information barriers between energy systems, but also quantify the credible regulation potential of multi-energy collaboration, and take into account the economy and flexibility of system operation; the present invention takes the source-load probability prediction and the credible regulation potential of flexible resources as the boundaries, and constructs a double-layer multi-scenario collaborative optimization configuration model with the goal of minimizing costs, accurately describes the uncertainty of source and load, quantifies the regulation capacity of flexible resources, and improves the economy and flexibility of the planning and operation of the integrated energy system as a whole.

[0195] Figure 4 The structural block diagram of the system configuration device based on probability prediction and credible regulation potential quantification provided by the present invention is shown in FIG. Figure 4 The system configuration device 400 based on probability prediction and credible adjustment potential quantification includes:

[0196] The data collection and preprocessing module 401 is used to collect source load historical data and corresponding meteorological data and perform data preprocessing;

[0197] A model training module 402 is used to establish a QRLSTM model combining quantile regression and long short-term memory neural network, and use the preprocessed data to train and test the model;

[0198] The source-load probability prediction module 403 is used to generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result;

[0199] The system modeling module 404 is used to determine the flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling;

[0200] The dispatch decoupling module 405 is used for the integrated energy system network model based on electricity-heat-cold, and adopts a robust optimization algorithm to aggregate the flexibility resources of the heat network and the cold network, realize the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid;

[0201] Cost optimization module 406, used to build a two-layer multi-scenario collaborative optimization configuration model with the goal of minimizing the cost of the integrated energy system;

[0202] The collaborative optimization configuration module 407 is used to solve the double-layer optimization configuration model according to the system operation parameters, the source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, and output the equipment configuration results of the integrated energy system.

[0203] The system configuration device based on probability prediction and quantification of credible adjustment potential provided by the present invention executes the system configuration method based on probability prediction and quantification of credible adjustment potential provided by the above-mentioned embodiments through the above-mentioned modules. The system configuration method based on probability prediction and quantification of credible adjustment potential has been described in detail in the above-mentioned embodiments, and will not be repeated in this embodiment.

[0204] Figure 5 The structural block diagram of the electronic device provided by the present invention is as follows: Figure 5 As shown, the present invention also provides an electronic device, and the electronic device 500 can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer, and a server. The electronic device 500 includes a processor 501 and a memory 502, wherein the memory 502 stores a system configuration program 503 based on probability prediction and quantification of trustworthy adjustment potential.

[0205] In some embodiments, the memory 502 may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory 502 may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. Further, the memory 502 may also include both an internal storage unit of the computer device and an external storage device. The memory 502 is used to store application software and various types of data installed on the computer device, such as program codes for installing the computer device. The memory 502 can also be used to temporarily store data that has been output or is to be output. In one embodiment, when the system configuration program 503 based on probability prediction and quantification of credible adjustment potential is executed by the processor 501, the following steps are implemented:

[0206] Collect historical data of source load and corresponding meteorological data, and perform data preprocessing;

[0207] Build a QRLSTM model that combines quantile regression and long short-term memory neural network, and use the preprocessed data to train and test the model;

[0208] Generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result;

[0209] Identify flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling;

[0210] Based on the integrated energy system network model of electricity-heat-cold, a robust optimization algorithm is used to aggregate the flexibility resources of the heat network and the cold network, realize the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid;

[0211] With the goal of minimizing the cost of the integrated energy system, a two-layer multi-scenario collaborative optimization configuration model is constructed;

[0212] According to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, the double-layer optimization configuration model is solved and the equipment configuration results of the integrated energy system are output.

[0213] In some embodiments, the processor 501 can be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 502, such as executing a system configuration program based on probability prediction and quantification of trusted adjustment potential.

[0214] This embodiment further provides a computer-readable storage medium, on which a system configuration program based on probability prediction and quantification of credible adjustment potential is stored. When the system configuration program based on probability prediction and quantification of credible adjustment potential 503 is executed by a processor, the following steps are implemented:

[0215] Collect historical data of source load and corresponding meteorological data, and perform data preprocessing;

[0216] Build a QRLSTM model that combines quantile regression and long short-term memory neural network, and use the preprocessed data to train and test the model;

[0217] Generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result;

[0218] Identify flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling;

[0219] Based on the integrated energy system network model of electricity-heat-cold, a robust optimization algorithm is used to aggregate the flexibility resources of the heat network and the cold network, realize the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid;

[0220] With the goal of minimizing the cost of the integrated energy system, a two-layer multi-scenario collaborative optimization configuration model is constructed;

[0221] According to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, the double-layer optimization configuration model is solved and the equipment configuration results of the integrated energy system are output.

[0222] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0223] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A system configuration method based on probability prediction and quantification of credible regulation potential, characterized in that: include: Collect historical data of source load and corresponding meteorological data, and perform data preprocessing; Build a QRLSTM model that combines quantile regression and long short-term memory neural network, and use the preprocessed data to train and test the model; Generate a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenario into the trained QRLSTM model, and obtain the source-load probability prediction result; Identify flexibility resources in the integrated energy system and establish an integrated energy system network model including electricity, heat and cooling; Based on the integrated energy system network model of electricity-heat-cold, a robust optimization algorithm is used to aggregate the flexibility resources of the heat network and the cold network, realize the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantify the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid; With the goal of minimizing the cost of the integrated energy system, a two-layer multi-scenario collaborative optimization configuration model is constructed; According to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid and the preset constraints, the double-layer optimization configuration model is solved and the equipment configuration results of the integrated energy system are output.

2. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 1 is characterized in that: The data preprocessing comprises: Data cleaning is performed on the source load historical data and the corresponding meteorological data: the Z-score algorithm is used to detect outliers in the data and the detected outliers are replaced with zero; the linear interpolation method is used to fill in the missing values.

3. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 1 is characterized in that: The establishment of a QRLSTM model combining quantile regression and long short-term memory neural network, and training and testing the model using preprocessed data specifically include: In the Tensorflow framework, the QRLSTM model is constructed by combining quantile regression and long short-term memory neural network; The preprocessed data samples are divided into training set and test set, and the quantile loss function is defined. The goal is to minimize the quantile loss function, and the QRLSTM model is trained using the training set. The test set is used to test whether the probability prediction accuracy of the trained QRLSTM model meets the preset accuracy requirements. If so, the model training is completed; if not, the above steps are repeated until the preset accuracy requirements are met.

4. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 1, characterized in that: The method generates a typical light-load joint time series scenario based on the DTW dynamic time warping algorithm, inputs the source-load data of the typical scenario into the trained QRLSTM model, and obtains the source-load probability prediction result, specifically including: By characterizing the temporal correlation between sources and loads, a photovoltaic-load joint scenario is generated; The number of typical scenes is determined based on the intra-cluster error quadratic sum method and the silhouette coefficient method, and the typical scenes and their corresponding probabilities are determined based on the DTW dynamic time warping algorithm; The source-load data corresponding to the typical scenario is input into the trained QRLSTM model to obtain the source-load probability prediction results.

5. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 1 is characterized in that: The determination of the flexibility resources in the integrated energy system and the establishment of an integrated energy system network model including electricity, heat and cooling specifically include: The flexibility resources in the integrated energy system include CCHP combined cooling, heating and power units, GB gas boiler equipment, HS thermal energy storage equipment, EC electric refrigerator equipment, CS cold energy storage equipment, PV photovoltaic units, and ES electric energy storage equipment; The heat network and cold network models are established by mass regulation. The water supply flow in the model is constant, and the flow into the node is equal to the flow out of the node. For the heat network, the heat provided by the heating equipment enters the thermal system through the heat exchange station. The expression of this process is: In the formula, and are the heating powers of CCHP and GB at time t, and are the heat release and charging power of HS at time t, c w is the specific heat capacity of water, is the flow injected into node k at time t, and are the heat source heating temperature and return water network temperature of node k at time t respectively; The expression for the dynamic change of water supply network temperature is: In the formula, and are the average temperatures of the water supply pipe and the return pipe p at time t, and are the outlet temperature and inlet temperature of the water supply pipe p at time t, and are the outlet temperature and inlet temperature of the return pipe p at time t, Δt is the time interval, S p , L p and λ p are the cross-sectional area, length and heat transfer coefficient of the pipe p, T t W is the ambient temperature of the pipe, m p,t is the flow rate of pipe p at time t; When hot water flows from multiple pipes to a node or from a node to multiple pipes, the mixed temperature of the node should satisfy the node temperature constraint. The expression of the node mixed temperature is: In the formula, is the mixed water temperature of node k in the water supply network at time t, and are the sets of pipes in the water supply / return network with node k as the end and start point respectively; The expression of heat exchange process on the load side is: In the formula, H k,t is the heat load of node k at time t, is the regenerative temperature of node k at time t; For the cold network, the cold air provided by the cooling equipment enters the cooling system through the cooling station. The expression of this process is: In the formula, and are the cooling powers of CCHP and EC at time t, and are the cooling and charging powers of CS at time t, and are the cooling source supply temperature and return water network temperature of node k at time t respectively; The expression of the cooling capacity exchange process of the load side cooling network is: In the formula, C k,t is the cooling load of node k at time t, is the cooling temperature of node k at time t.

6. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 5, characterized in that: The integrated energy system network model based on electricity, heat and cold adopts a robust optimization algorithm to aggregate the flexibility resources of the heat network and the cold network, realizes the scheduling decoupling of heat and electricity, heat and cold, and cold and electricity, and quantifies the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid. Specifically, it includes: A robust optimization algorithm is used to establish a heat network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated CCHP output under the worst scenario: In the formula, and H CCHP are the vectors consisting of the upper and lower limits of CCHP output at each moment, ξ t represents the uncertainty of CCHP output at time t, U is the set of CCHP output uncertainties at each time; x(ξ t ) is the vector of other decision variables except CCHP output at time t, including GB output, HS charging power, heat release power and HS charging and discharging state; Solve the heat network flexibility aggregation model to obtain the optimization result after heat network power aggregation Using CCHP electrical-thermal coupling relationship Decoupling to obtain the regulation potential range provided by the heat network to the power grid Using CCHP heat-cold coupling relationship Decoupling to obtain the flexible power support range provided by the heating network to the cooling network A robust optimization algorithm is used to establish a cold network flexibility aggregation model with the goal of maximizing the difference between the feasible upper and lower bounds of the aggregated EC output in the worst scenario: In the formula, and C EC are the vectors consisting of the feasible upper and lower bounds of EC output at each moment, ζ t represents the uncertainty of EC output at time t, D is the set of EC output uncertainties at each time; x(ζ t ) is a vector of decision variables other than the EC output at time t, including the CCHP output after heat-cold decoupling, the CS charging and cooling power, the CS cooling power, and the CS charging and discharging state; Solve the cold network flexibility aggregation model to obtain the optimization result after cold network power aggregation Using EC electrical-cold coupling relationship Decoupling to obtain the power regulation range of the cooling network The flexibility resources in the integrated energy system can provide upward / downward flexible adjustment capabilities, which can be expressed as: In the formula, F t CCHP,up 、F t ES,up 、F t EC,up and F t PV,up are the upward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, respectively, t CCHP,dn 、F t ES,dn 、F t EC,dn and F t PV,dn are the downward flexible adjustment capabilities of CCHP, ES, EC and PV at time t, r CCHP and r EC are the ramp rates of CCHP and EC respectively, and are the charging efficiency and discharging efficiency of ES respectively, and are the maximum charging power and maximum discharging power of ES, E ES is the capacity of ES, SOC t is the energy storage state of ES at time t, and SOC ES are the maximum energy storage state and the minimum energy storage state of ES, respectively, t PV is the PV output at time t, is the maximum output of PV at time t, is the minimum output of PV at time t; The flexibility resources in the integrated energy system can provide the maximum upward adjustable potential F t m,up The expression is: F t m,up =F t CCHP,up +F t ES,up +F t PV,up +F t EC,up The flexibility resources in the integrated energy system can provide the maximum downward adjustable potential F t m,dn The expression is: F t m,dn =F t CCHP,dn +F t ES,dn +F t PV,dn +F t EC,dn Since photovoltaic power is an uncontrollable power source, its output is volatile and uncontrollable. Flexible resources in the integrated energy system can provide upward credible regulation potential. t b,up The expression is: F t b,up =F t CCHP,up +F t ES,up -F t PV,dn +F t EC,up Flexible resources in an integrated energy system can provide credible downward regulation potential. t b,dn The expression is: F t b,dn =F t CCHP,dn +F t ES,dn -F t PV,up +F t EC,dn 。 7. The system configuration method based on probability prediction and quantification of credible regulation potential according to claim 6 is characterized in that: The two-layer multi-scenario collaborative optimization configuration model includes an upper layer model and a lower layer model; The upper model aims to minimize the total system investment cost and annual operating cost, and the upper objective function is: min(C inv +C ope ) In the formula, C inv is the total investment cost, C ope is the annual operating cost of the system; The lower model aims to minimize the annual operating cost of the system, and the objective function of the lower layer is: minC ope The two-layer multi-scenario collaborative optimization configuration model consisting of the upper model and the lower model is: In the formula, s is a typical scene, S is a typical scene set, π s is the dual variable of scenario s; Total investment cost of integrated energy system C inv The expression is: Where j is the jth device among CCHP, GB, EC, ES, HS, CS and PV, k j is the discounted years of the jth equipment, M j is the installed capacity of the jth type of equipment, is the unit capacity investment cost of the jth type of equipment; System annual operating cost C ope The expression is: In the formula, ρ s is the probability of a typical scenario s, and are the system electricity purchase cost, gas purchase cost and operation and maintenance cost of flexible resources under scenario s, is the penalty cost of insufficient system flexibility under scenario s, a and b are the weight coefficients of economic cost and penalty cost of insufficient flexibility respectively; The penalty cost of insufficient system flexibility is expressed as: In the formula, F t def,up 、F t def,dn and F t def They are the upward, downward and total flexibility of the system at time t, respectively, t load,up and F t load,dn are the upward and downward regulation requirements of the electric load at time t, c def is the penalty cost coefficient.

8. A system configuration device based on probability prediction and quantification of credible regulation potential, characterized in that: include: Data collection and preprocessing module, used to collect source load historical data and corresponding meteorological data, and perform data preprocessing; Model training module, used to build a QRLSTM model combining quantile regression and long short-term memory neural network, and use preprocessed data to train and test the model; The source-load probability prediction module is used to generate typical light-load joint time series scenarios based on the DTW dynamic time warping algorithm, input the source-load data of the typical scenarios into the trained QRLSTM model, and obtain the source-load probability prediction results; System modeling module, used to determine the flexibility resources in the integrated energy system and establish a network model of the integrated energy system including electricity, heat and cooling; The dispatch decoupling module is used for the integrated energy system network model based on electricity-heat-cold. It uses a robust optimization algorithm to aggregate the flexibility resources of the heat network and the cold network, realizes the dispatch decoupling of heat and electricity, heat and cold, and cold and electricity, and quantifies the maximum adjustable potential and credible adjustment potential of the heat and cold network to the power grid; Cost optimization module, which is used to build a two-layer multi-scenario collaborative optimization configuration model with the goal of minimizing the cost of the integrated energy system; The collaborative optimization configuration module is used to solve the double-layer optimization configuration model according to the system operating parameters, source-load probability prediction results, the maximum power support range of the heat and cold network to the power grid, and preset constraints, and output the equipment configuration results of the integrated energy system.

9. An electronic device, It is characterized in that comprising a memory and a processor, wherein: The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the system configuration method based on probability prediction and quantification of credible adjustment potential as described in any one of claims 1 to 7 above.

10. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the system configuration method based on probability prediction and quantification of reliable adjustment potential as described in any one of claims 1 to 7.

Images