Distributed robust optimization scheduling method and system of solar photovoltaic photo-thermal comprehensive utilization system

Abstract

The invention discloses a distributed robust optimization scheduling method and system for a solar photovoltaic photo-thermal comprehensive utilization system, and belongs to the field of comprehensive utilization of renewable energy sources. The method comprises the steps of obtaining an electrical load and a thermal load based on an integrated load prediction model; obtaining typical sunlight photovoltaic and photo-thermal based on scene clustering analysis; based on photovoltaic, photo-thermal, electrical load and thermal load, an objective function and a constraint condition for optimal scheduling are set for the solar photovoltaic photo-thermal comprehensive utilization system, and based on the objective function and the constraint condition, a two-stage distribution robust optimal scheduling model is obtained by combining a comprehensive demand response mechanism; and solving the two-stage distribution robust optimization scheduling model by adopting a column and constraint generation algorithm to obtain a two-stage distribution robust optimization scheduling result. The method obviously improves the utilization rate of renewable energy sources, reduces the operation cost and carbon emission of the system, and is suitable for building energy systems in solar energy-enriched areas.

Description

Technical Field

[0001] The present invention relates to the technical field of comprehensive utilization of renewable energy, and particularly to a distributed robust optimal scheduling method and system for a solar photovoltaic-thermal comprehensive utilization system. Background Art

[0002] To achieve the development goal of "carbon neutrality", all industries are working hard to accelerate the pace of reducing carbon emissions and guide green technology innovation; as a major source of carbon emissions, the carbon emissions from the operation part of the construction industry account for 28% of the global emissions. The International Energy Agency pointed out in its "Renewable Energy Report" in 2024 that renewable energy has important resource value in alleviating the environmental challenges and resource shortages brought about by fossil fuel consumption. In the future energy system, the importance of solar energy may be further enhanced. Tibet is located in a solar-rich area on the Qinghai-Tibet Plateau of China, and has endowment conditions for realizing energy transformation and large-scale utilization of solar energy. Improving the application ratio of solar photovoltaic and solar thermal in the buildings in this area is the key means to achieve the carbon neutrality goal in this region, and has important practical significance.

[0003] The solar photovoltaic-thermal comprehensive utilization system is organically coupled by power and thermal subsystems, providing a new solution for realizing the collaborative complementarity of energy in solar-rich areas and optimizing the energy structure. However, the multiple uncertainties in the system pose major challenges to the safe and stable operation of the system. The existing mainstream methods for solving uncertainties are stochastic optimization and robust optimization. Stochastic optimization is the most commonly used technique, but it is difficult to obtain an accurate distribution in practical applications, and stochastic optimization also depends on the number of scenarios, which often leads to an excessive computational workload. Robust optimization ignores the distribution information and only focuses on the worst-case scenario, so it usually occurs with a particularly low probability. In many cases, the optimal solution may be conservative and uneconomical. At the same time, it is also challenging to describe the distribution information with only one uncertainty set. To alleviate the system's energy supply pressure and meet user needs, a large number of existing studies and practices stimulate the interaction between demand-side resources and renewable energy generation through the application of demand response. However, demand response is only applicable to single-energy carrier systems and can only be implemented in the presence of adjustable or curtailable loads, which may sacrifice the comfort of energy users. In addition, since the data used in system scheduling may come from multiple channels, it is difficult to guarantee the standardization of the data, which reduces the accuracy of load forecasting. Summary of the Invention

[0004] To solve the problems of conservative scheduling and poor economy caused by uncertainties and improve the renewable energy consumption capacity, the present invention provides a distributed robust optimal scheduling method, system, device and storage medium for a solar photovoltaic-thermal comprehensive utilization system. The present invention is applicable to the optimal scheduling of building energy systems in solar-rich areas.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The present invention provides a distributed robust optimal scheduling method for a solar photovoltaic and solar thermal integrated utilization system, including:

[0007] Obtaining the electrical load and the thermal load based on an integrated load prediction model; obtaining the typical daily photovoltaic and solar thermal based on scenario clustering analysis;

[0008] Based on the photovoltaic, solar thermal, electrical load and thermal load, setting the objective function and constraint conditions for the optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system, and based on the objective function and constraint conditions, combining with the integrated demand response mechanism, obtaining a two-stage distributed robust optimal scheduling model;

[0009] Solving the two-stage distributed robust optimal scheduling model by using the column and constraint generation algorithm to obtain the two-stage distributed robust optimal scheduling result.

[0010] The objective function of the optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system is as follows:

[0011]

[0012] In the formula, p s is the probability of the scenario s after clustering; C PV-PT is the operation and maintenance cost of the photovoltaic and solar thermal components; is the cost of purchasing electricity from the main power grid; C loss is the cost of light curtailment and heat rejection; C EES and C PCSD are the operation costs of the energy storage system and the heat storage device respectively; is the carbon dioxide emission fine cost; C IDR is the cost of the integrated demand response; is the profit from the power grid electricity sales.

[0013] The constraint conditions include the operation constraints of the solar photovoltaic and solar thermal equipment, the interactive power constraints of the system and the distribution network, the operation constraints of the air source heat pump, the operation constraints of the energy storage system, the operation constraints of the heat storage device, the power balance constraint and the thermal balance constraint.

[0014] Based on the objective function and constraint conditions, combining with the integrated demand response mechanism, obtaining the two-stage distributed robust optimal scheduling model, specifically:

[0015] The two stages in the two-stage distributed robust optimal scheduling model are: the first stage formulates a scheduling plan based on the power purchase and sale plan of the main power grid and the charging and discharging of the energy storage; the second stage formulates a scheduling plan after the uncertainty of the renewable energy and the user energy consumption appears;

[0016] The two-stage distributionally robust optimization scheduling model is as follows:

[0017]

[0018] s.t. Bx ≤ a

[0019] Cy s ≤ b

[0020] Dy s =c

[0021] Ex + Fy s ≤ d

[0022] In the formula, s represents the scenario; p s is the probability of the scenario s after clustering; Ω p is the set interval of the probability distribution that p s satisfies, representing the confidence set constrained by the 1-norm and ∞-norm; N s is the total number of scenarios after clustering; A - F and a - d are constant coefficient matrices; y s represents the second-stage variable; x represents the first-stage variable.

[0023] The first-stage variable x included in the first stage is expressed as:

[0024]

[0025] The second-stage variable y included in the second stage is expressed as follows:

[0026]

[0027] In the formula, and are the power purchase and sale status between the system and the main power grid at time t, and are the operating status of the energy storage system at time t, and are the operating status of the heat storage device at time t, P PV,s,t is the electric power of the photovoltaic at time t; Q PT,s,t is the solar thermal power at time t; and are the electricity purchased from the power grid and sold to the power grid by the system during time t respectively; P ASHP,s,t is the electric power of the air source heat pump at time t; Q ASHP,s,t is the heat output power of the air source heat pump at time t; and are the charging and discharging powers of the energy storage system respectively; and are the heat storage and heat release powers of the heat storage device respectively; is the replaceable electric power at time t, is the replaceable heat power at time t, is the shifted load in time period t, is the interruptible load in time period t; and P IDR,t are the electric loads before and after the integrated demand response respectively, and Q IDR,t tables are the heat loads before and after the integrated demand response respectively.

[0028] The p s satisfies the probability distribution set interval Ω p To conform to the actual operation data and fluctuate within a reasonable range, a set centered on the initial probability distribution values of each scenario and restricted by the 1-norm and ∞-norm is constructed to limit the probability distribution values of the photovoltaic and solar thermal scenarios, expressed as:

[0029]

[0030] In the formula, p s0 is the initial probability value of the s-th discrete scenario screened from the available historical data of the dispatching system; θ 1 and θ ∞ are the probability allowable deviation limits under the 1-norm and ∞-norm constraint conditions respectively;

[0031] The probability distribution p of the clustered scenario s s satisfies the following confidence level:

[0032]

[0033] In the formula, Pr{·} is the probability operator. The probability allowable deviation values under the two norm constraint conditions can be obtained from the above two formulas, and then used to limit the fluctuation range of the scenario probability distribution; Ns is the total number of clustered scenarios.

[0034] The column and constraint generation algorithm is used to solve the two-stage distributionally robust optimization scheduling model to obtain the two-stage distributionally robust optimization scheduling result. Specifically:

[0035] The column and constraint generation algorithm divides the solution of the two-stage distributionally robust optimization scheduling model into the master problem MP and the sub-problem SP;

[0036] Set the lower bound value LB = -∞, the upper bound value UB = +∞, the convergence accuracy is ε, set the iteration times m = 1, and the initial probability distribution of the second stage is p s0 ;

[0037] Solve the master problem to obtain the optimal solution (x * , η * ), and update the lower bound value LB = η* ;

[0038] The main problem is expressed as follows:

[0039]

[0040] Where m represents the number of iterations, M represents the maximum number of iterations, and η is an introduced auxiliary variable;

[0041] Based on solving the main problem to obtain the optimal solution x * , solve the sub-problem to obtain the probability value in the worst-case scenario and the objective function value f(x * ), and at the same time update the upper bound value min{UB, f(x * )};

[0042] The sub-problem is expressed as follows:

[0043]

[0044] Where represents the two-stage decision variable in the m-th iteration, represents the solution result of the internal "min" problem, represents the probability value in the m-th iteration;

[0045] When UB - LB < ε, stop the iteration and return the optimal solution x * ; Otherwise, update the worst-case probability distribution of the main problem and add new variables and constraint conditions to the main problem; update m = m + 1, and return to the main problem-solving for iteration.

[0046] The present invention also provides a distributionally robust optimization scheduling system for a solar photovoltaic-thermal comprehensive utilization system, characterized by including:

[0047] A data acquisition module for obtaining electrical load and thermal load based on an integrated load prediction model; obtaining typical day photovoltaic and solar thermal based on scenario clustering analysis;

[0048] A scheduling model construction module for setting the objective function and constraint conditions for the optimal scheduling of the solar photovoltaic-thermal comprehensive utilization system based on photovoltaic, solar thermal, electrical load, and thermal load, and obtaining a two-stage distributionally robust optimization scheduling model based on the objective function and constraint conditions in combination with the comprehensive demand response mechanism;

[0049] A scheduling result solving module for solving the two-stage distributionally robust optimization scheduling model by using the column and constraint generation algorithm to obtain the two-stage distributionally robust optimization scheduling result.

[0050] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the distributed robust optimization scheduling method of the above-mentioned solar photovoltaic and thermal comprehensive utilization system are implemented.

[0051] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the distributed rod optimization scheduling method of the above-mentioned solar photovoltaic and thermal comprehensive utilization system are implemented.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The distributed robust optimization scheduling method for solar photovoltaic and thermal comprehensive utilization system proposed in the present invention takes into account data privacy and standardization, proposes an integrated load forecasting model, and analyzes the load characteristics of users; considering that the precise probability distribution of data is usually difficult to obtain, a two-stage distributed robust optimization scheduling model considering the comprehensive norm is proposed to solve the operation scheduling and source-load uncertainty problems of solar photovoltaic and thermal comprehensive utilization systems in solar energy-rich areas, broadening the research framework of distributed energy systems. Taking system operation and maintenance costs, carbon emissions and other indicators as optimization targets, the column and constraint generation algorithm is used to solve the proposed model without dualization processing, and the data obtained from the monitoring of the solar photovoltaic and thermal comprehensive utilization system are used to verify the feasibility of the optimization scheduling method proposed in the present invention. The optimization scheduling method proposed in the present invention uses the complementarity of the solar photovoltaic and thermal comprehensive utilization system to enable users to actively participate in the response plan by converting electricity into heat during peak hours, which takes into account transferable loads, interruptible loads and replaceable loads. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0055] Figure 1 A flow chart of a distributed blue rod optimization scheduling method for a solar photovoltaic and thermal comprehensive utilization system provided by the present invention;

[0056] Figure 2 This is a structural diagram of the solar photovoltaic and thermal comprehensive utilization system of the present invention;

[0057] Figure 3 A flow chart for solving the two-stage distributed robust optimization scheduling model provided by the present invention;

[0058] Figure 4 It is the feature importance ranking diagram extracted in the embodiments of the present invention, where (a) is the feature importance ranking diagram of the electrical load, and (b) is the feature importance ranking diagram of the thermal load;

[0059] Figure 5 It is the box plot of the relative errors of the five models in the embodiments of the present invention;

[0060] Figure 6 It is the hourly solar irradiance outdoors in the embodiments of the present invention;

[0061] Figure 7 It is the value of different numbers of clusters in the embodiments of the present invention;

[0062] Figure 8 It is the typical scenario of the photovoltaic output in the embodiments of the present invention;

[0063] Figure 9 It is the typical scenario of the solar thermal output in the embodiments of the present invention;

[0064] Figure 10 It is the electro-thermal scheduling result of the solar photovoltaic-thermal comprehensive utilization system in the embodiments of the present invention;

[0065] Figure 11 It is the scheduling result after the integrated demand response participates in the scheduling of the solar photovoltaic-thermal comprehensive utilization system in the embodiments of the present invention;

[0066] Figure 12 It is the schematic diagram of the distributed robust optimal scheduling system structure of the solar photovoltaic-thermal comprehensive utilization system in the preferred embodiments of the present invention;

[0067] Figure 13 It is the schematic diagram of the electronic device structure in the preferred embodiments of the present invention. Detailed implementation manners

[0068] To enable those skilled in the art to understand the features and effects of the present invention, the following provides a general description and definition of the terms and phrases mentioned in the specification and claims. Unless otherwise specified, all technical and scientific terms used herein shall have the ordinary meaning understood by those skilled in the art for the present invention. In case of conflict, the definition in this specification shall prevail.

[0069] The theories or mechanisms described and disclosed herein, whether correct or incorrect, shall not limit the scope of the present invention in any way, that is, the content of the present invention can be implemented without being limited by any specific theory or mechanism.

[0070] In this text, all features defined in the form of numerical ranges or percentage ranges, such as numerical values, quantities, contents, and concentrations, are only for the sake of brevity and convenience. Accordingly, the description of numerical ranges or percentage ranges should be regarded as having covered and specifically disclosed all possible sub-ranges and individual numerical values (including integers and fractions) within the ranges.

[0071] In this text, unless otherwise specified, the terms "comprise", "include", "contain", "have", or similar terms cover the meanings of "consist of" and "consist essentially of". For example, "A comprises a" covers the meanings of "A comprises a and others" and "A consists only of a".

[0072] In this text, for the sake of concise description, not all possible combinations of all technical features in each embodiment or example are described. Therefore, as long as there is no contradiction in the combination of these technical features, the technical features in each embodiment or example can be combined arbitrarily, and all possible combinations should be considered as the scope described in this specification.

[0073] The overall structure of the solar photovoltaic and solar thermal integrated utilization system is as Figure 2 shown. The solar photovoltaic and solar thermal integrated utilization system of the present invention consists of two energy subsystems. Among them, electric energy is provided by distributed photovoltaics and the main power grid, while thermal energy comes from the solar collector field and the air source heat pump (ASHP). The networks of the electric and thermal systems are coupled through the air source heat pump, providing an energy flow path to convert excess electricity into heat. The loads mainly include the electricity loads and heat loads of residential buildings, office buildings, and public buildings. To ensure the safe, flexible, and economic operation of the system, the system is also equipped with buffer devices such as an energy storage system (ESS) and a heat storage device (PCSD).

[0074] As Figure 1 shown, the present invention provides a distributionally robust optimal scheduling method for a solar photovoltaic and solar thermal integrated utilization system, including:

[0075] Obtaining the electricity load and heat load based on an integrated load forecasting model; obtaining the typical day's photovoltaics and solar thermal based on scenario clustering analysis;

[0076] Based on the photovoltaics, solar thermal, electricity load, and heat load, setting the objective function and constraint conditions for the optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system, and based on the objective function and constraint conditions, combining with the integrated demand response mechanism, obtaining a two-stage distributionally robust optimal scheduling model;

[0077] Using the column and constraint generation algorithm to solve the two-stage distributionally robust optimal scheduling model to obtain the two-stage distributionally robust optimal scheduling result.

[0078] Specifically, the electrical load and heat load are obtained through an integrated load forecasting model; the construction of the integrated load forecasting model incorporates the SHAP method, and the steps for obtaining the integrated load forecasting model mainly include data preprocessing, feature analysis, model construction, model training, model prediction, and evaluation, specifically as follows:

[0079] The obtained historical data is preprocessed to obtain a data set. The data preprocessing includes processing missing values using the interpolation method, processing outliers in the data using the 3σ method, and normalizing the data through Min-max normalization;

[0080] The data set is subjected to feature analysis to obtain a feature data set for training the model. The feature analysis uses the SHAP method, which is a model interpretability method based on the Shapley value of game theory and is used to quantify the contribution of each feature to the prediction result of the prediction model.

[0081] The calculation formula of the Shapley value is as follows:

[0082]

[0083] In the formula, N represents the set of all features; S represents the subset that does not contain feature i; |S| represents the number of features in subset S; f(S∪{i}) represents the prediction result of the model after adding feature i to subset S; f(S) represents the prediction result of the model under the feature subset S; |S|!(n - |S| - 1)! / n! is the weight for all possible feature combinations and represents the average weight of the marginal contribution of feature i in all permutations.

[0084] In some embodiments, 12 common features are selected for the heat load and electrical load respectively from three dimensions: time features, meteorological features, and historical data. Among them, the time features include time and date; the meteorological data is obtained from a local meteorological station and mainly includes air temperature, humidity, precipitation, ground wind speed, wind direction, surface horizontal radiation, direct solar radiation, diffuse solar radiation, and total solar radiation; the historical data includes the total solar radiation at the previous moment, the heat load and electrical load at the previous moment.

[0085] The feature data set is divided into a training set and a test set, and the preset integrated load forecasting model is trained and evaluated to obtain the integrated load forecasting model;

[0086] The construction of the preset integrated load forecasting model is based on the XGBoost algorithm for integrated load forecasting. The core idea of ensemble learning is that by combining the predictions of multiple models, the generalization error can be reduced, thereby enhancing the robustness and accuracy of the model. XGBoost belongs to the boosting framework algorithm in ensemble learning. The tree model used is the CART regression tree model, and the prediction result of the final model is the sum of the prediction results of all base learners.

[0087] The model of XGBoost is as follows:

[0088]

[0089] In the formula, represents the predicted value of the i-th sample, q i represents the feature value of the i-th sample, represents the prediction results of the previous W - 1 trees, represents the prediction result of the W-th tree.

[0090] Specifically, the training set is used to train the preset integrated load forecasting model. The base learner constructed by the SHAP-XGBoost model provided by the present invention is a tree model. Hyperparameter tuning is performed using grid search. The number of the final optimized tree models is determined to be 100, the learning rate is 0.3, and the depth is 6; the proportions of the training set, validation set, and test set are 70%, 20%, and 10% respectively, and the model is trained using 5-fold cross-validation;

[0091] RMSE, R 2 and CV are used to predict and evaluate the trained integrated load forecasting model, and the specific expressions are as follows:

[0092]

[0093] In the formula and l i are the actual and predicted building energy consumptions respectively, is the average predicted building energy consumption. n is the total number of samples in the given test set, and i is the number of samples in the test set. Among them, the smaller RMSE and CV are, the better the prediction effect is. When R 2 is very close to 1, better prediction performance can be guaranteed.

[0094] Based on the cluster analysis of photovoltaic and solar thermal data, typical daily photovoltaic and solar thermal data are obtained. The Davies-Bouldin Index (BDI) and Silhouette Coefficient (SC) are used to evaluate the clustering effect. BDI takes into account the tightness of sample points within each cluster and the separation between different clusters. When the DBI reaches its minimum value, it indicates a better clustering effect. SC involves the distance between each data point and other data points within its cluster, as well as the distance between this data point and other clusters. The value range of SC is [-1, 1], and the closer it is to 1, the better the clustering effect.

[0095] Specifically, the objective function for the optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system is as follows:

[0096]

[0097] C PV-PT = c PV-PT (P PV,s,t + Q PT,s,t )

[0098]

[0099] In the formula, C PV-PT is the operation and maintenance cost of the photovoltaic and solar thermal components; is the cost of purchasing electricity from the main power grid; C loss is the cost of curtailment of photovoltaic power and heat; C EES and C PCSD are the operation costs of the energy storage system and the heat storage device respectively; is the carbon dioxide emission fine cost; C IDR is the cost of integrated demand response; is the profit from the sale of electricity by the power grid; c PV-PT is the operation and maintenance cost coefficient per unit power of the photovoltaic and solar thermal; and are the unit prices of purchasing and selling electricity from / to the power grid by the system respectively; and are the electricity purchased from and sold to the power grid by the system during period t respectively; c loss is the unit fine cost of curtailment of photovoltaic power and heat; and are the predicted electricity output and heat output values of the photovoltaic and solar thermal during period t respectively; P PV,s,t is the electric power of the photovoltaic at time t; Q PT,s,t is the solar thermal power at time t; c EES and c PCSD are the converted unit charge-discharge and heat energy costs respectively; η EES and η PCSDis the charging and discharging efficiency of the energy storage system and the heat storage device; and are the charging and discharging powers of the energy storage system respectively; and are the heat storage and heat release powers of the heat storage device respectively; is the fine price per unit of CO 2 emission; k grid is the CO 2 emission per unit of electric power generated by the power grid; ω ESL is the compensation price for the transferable electric load; ω EIL is the compensation price for the interruptible electric load, ω EHSL is the compensation price for the alternative load.

[0100] Specifically, the constraint conditions include the operation constraints of solar photovoltaic-thermal equipment, the interactive power constraints of the system and the distribution network, the operation constraints of the air source heat pump, the energy storage device constraints, and the system power balance constraints; the energy storage device constraints include the operation constraints of the energy storage system and the operation constraints of the heat storage device, and the system power balance constraints include the electric power balance constraint and the thermal balance constraint.

[0101] The operation constraints of the solar photovoltaic-thermal equipment are as follows:

[0102]

[0103] In the formula, is the upper limit of the photovoltaic output electric power; is the upper limit of the solar thermal output heat power.

[0104] It should be noted that when the photovoltaic and energy storage systems cannot meet the electric load demand during discharging, electricity needs to be purchased from the main power grid; when the photovoltaic power generation is surplus, the redundant electric energy can be sold to the main power grid to reduce the system operation cost.

[0105] Specifically, the interactive power constraints of the system and the distribution network are as follows:

[0106]

[0107] In the formula: and are the power purchase and sale status of the system and the main power grid at time t. When and take the value of 1, it means power purchase and sale from the main power grid. When the value is 0, it means that the main power grid neither purchases nor sells electricity. and are the upper limits of the interactive power between the main power grid and the system respectively.

[0108] It should be noted that an air source heat pump is an energy-saving device that uses electric energy to make low-temperature heat source air flow to a high-temperature heat source, and can convert low-temperature heat energy that cannot be directly utilized in the air into high-temperature heat energy that can be directly utilized.

[0109] Specifically, the operating constraints of the air source heat pump are as follows:

[0110]

[0111] Q ASHP,s,t =COP ASHP P ASHP,s,t

[0112] In the formula, P ASHP,s,t is the electric power of the air source heat pump at time t, is the rated power of the air source heat pump, Q ASHP,s,t is the heat output power of the air source heat pump at time t, and COP ASHP is the energy efficiency coefficient of the air source heat pump.

[0113] It should be noted that the energy storage system and the heat storage device can realize the cross-time transfer of energy. When the system is in a period of low electricity price or high renewable energy output, the energy storage system and the heat storage device store energy. When the electricity price is high or the renewable energy output is low, the energy storage system and the heat storage device release energy to meet the load demand, realizing the low-cost and stable operation of the system.

[0114] Specifically, the operating constraints of the energy storage system are as follows:

[0115]

[0116] In the formula, and are the operating states of the energy storage system at time t. When and take the value of 1, it means that the energy storage system is in the charging and discharging states. When the value is 0, it means that the main power grid is neither charging nor discharging. and are the maximum chargeable and dischargeable powers of the energy storage system respectively, and η ESS is the charge-discharge efficiency of the energy storage system, T is the scheduling period, taking 24h, Δt is the scheduling step length, taking 1h, and E ESS,0 is the initial scheduling capacity, and are the minimum and maximum remaining capacities of the energy storage system.

[0117] Specifically, the operating constraints of the heat storage device are as follows:

[0118]

[0119] In the formula, and is the operating state of the heat storage device at time t. When and take the value of 1, it means that the heat storage device is in the state of storing and releasing heat. When the value is 0, it means that the heat storage device neither stores nor releases heat. and are the maximum storable and releasable powers of the heat storage device respectively, and η PCSD is the charging and discharging efficiency of the heat storage device, T is the scheduling period, taking 24 h, Δt is the scheduling step, taking 1 h, and E HSD,0 is the initial scheduling capacity. and are the minimum and maximum remaining capacities of the heat storage device.

[0120] Specifically, the power balance constraint is as follows:

[0121]

[0122] In the formula, P IDR,t is the electrical load after integrated demand response.

[0123] Specifically, the thermal balance constraint is as follows:

[0124]

[0125] In the formula, Q IDR,t is the thermal load after integrated demand response.

[0126] It should be noted that considering the accommodation of distributed energy, the integrated demand response (IDR) mechanism proposed in the present invention includes the regulation of electrical load and the transfer regulation between electrical and thermal loads. According to the response characteristics of electrical load, it is divided into electricity shiftable load (ESL) and electricity interruptible load (EIL). Considering that the temperature in the research area is cold and users cannot accept the reduction of thermal load over time, therefore, the thermal load is regarded as a rigid load directly supplied by the solar thermal system and the air source heat pump, as well as the electricity-to-heat load (EHSL) that can be increased. The models of electrical load and thermal load are as follows.

[0127]

[0128] In the formula, and are the electrical loads before and after implementing integrated demand response respectively, and Q IDR,t tables are the thermal loads before and after integrated demand response respectively.

[0129] Specifically, the electricity shiftable load model is as follows:

[0130]

[0131] In the formula, is the shifted load in time period t, and are respectively the upper and lower limits of the adjustable shifted load in time period t.

[0132] Specifically, the interruptible load of electricity is as follows:

[0133]

[0134] In the formula, is the interruptible load in time period t, is the maximum allowable interruptible load within time period t.

[0135] In some embodiments, it is set that is 10% of the power load demand for each time period.

[0136] Specifically, the heat - electricity substitution load is as follows

[0137]

[0138] In the formula, is the replaceable electric power at time t, is the replaceable heat power at time t, is the upper limit of the replaceable load in time period t, and η ASHP is the conversion coefficient of the air - source heat pump.

[0139] In some embodiments, it is set that is 10% of the heat load demand for each time period.

[0140] It should be noted that considering the multiple uncertainties of the photovoltaic, solar thermal and load demand of the solar photovoltaic - thermal comprehensive utilization system and the consumption of renewable energy, a two - stage distributionally robust optimization scheduling model is adopted. The first stage is to make decisions on the power purchase and sale plan of the main power grid and the charge and discharge of energy storage. The variables x in the first stage include discrete decision variables for determining the power purchase and sale of the main power grid, charge and discharge state variables of the energy storage system, and heat storage and release state variables of the heat storage device. The second stage includes formulating corresponding scheduling plans after the uncertainties of renewable energy and user energy consumption appear, including the output of photovoltaic - thermal and air - source heat pumps, purchasing electricity and selling electricity from the main power grid, etc. The variables y in the second stage include continuous variables such as the charge and discharge of the energy storage system and heat storage and release of the heat storage device, the actual output of photovoltaic and solar thermal, the electricity purchased and sold from the main power grid, and the output of the air - source heat pump. The first - stage variable x and the second - stage variable y are represented as follows:

[0141]

[0142] Based on the setting of the objective function and constraint conditions for the optimal scheduling of the above-mentioned solar photovoltaic-thermal comprehensive utilization system, after transforming the objective function and constraint function, the compact form of the two-stage distributionally robust optimization model can be expressed as follows:

[0143]

[0144] s.t. Bx ≤ a

[0145] Cy s ≤ b

[0146] Dy s =c

[0147] Ex + Fy s ≤ d

[0148] where s represents the scenario; p s is the probability distribution of the scenario s after clustering; Ω p is for p s to satisfy the probability distribution set interval, representing the confidence set constrained by the 1-norm and ∞-norm; Ns is the total number of scenarios after clustering; A - F and a - d are constant coefficient matrices.

[0149] It should be noted that theoretically, Ω p can be any range. However, to make it more in line with the actual operation data and ensure its fluctuation within a reasonable range, a set centered on the initial probability distribution value of each scenario and restricting the probability distribution value of the photovoltaic and solar thermal scenarios through the 1-norm and ∞-norm is constructed, which is expressed as follows:

[0150]

[0151] where p s0 is the initial probability value of the s-th discrete scenario screened from the available historical data of the scheduling system; θ 1 and θ ∞ are the probability allowable deviation limits under the 1-norm and ∞-norm constraint conditions respectively.

[0152] Specifically, the probability distribution p s satisfies the following confidence levels:

[0153]

[0154] where Pr{·} is the probability operator. From the above two equations, the probability allowable deviation values under the two norm constraint conditions can be obtained, and then used to limit the fluctuation range of the scenario probability distribution; let the right sides of the above equations be the uncertainty probability confidence levels α 1 and α ∞ respectively, then there is:

[0155]

[0156] Since the comprehensive norm confidence interval contains absolute value constraints, it is necessary to perform a linear equivalent decomposition on it by introducing 0-1 auxiliary variables, perform an equivalent conversion on the absolute value constraints, and the final obtained constraints are:

[0157]

[0158] In the formula, and are the positive offset and negative offset of the probability distribution p s of scenario s relative to p s0 respectively; and are the 0-1 flag variables that cause positive and negative offsets to p s respectively. Finally, p s is the actual scenario probability distribution.

[0159] The column and constraint generation algorithm is used to solve the two-stage distributionally robust optimization scheduling model to obtain the two-stage distributionally robust optimization scheduling result, specifically:

[0160] For the convenience of solution, for the compact form of the two-stage distributionally robust optimization model, the column and constraint generation (C&CG) algorithm is used to divide the original problem into a master problem (MP) and a sub-problem (SP) for solution.

[0161] The master problem is expressed as follows:

[0162]

[0163] In the formula, m represents the number of iterations, M represents the maximum number of iterations, and η is the introduced auxiliary variable.

[0164] By solving the master problem, the variable x * is obtained, and a lower bound value is provided for the compact form of the two-stage distributionally robust optimization model.

[0165] The sub-problem is to find the worst-case probability distribution under real-time operation given the first-stage variable x * in the master problem, return it to the master problem, and provide an upper bound value for the compact form of the two-stage distributionally robust optimization model. The sub-problem is expressed as follows:

[0166]

[0167] Since the probability value of scenario s and the second-stage variables are independent of each other in the sub-problem, the sub-problem can be solved in two steps. First, solve the inner minimum problem in the sub-problem, and then solve the outer problem in the sub-problem. Therefore, the solution result of the internal "min" problem can be expressed as

[0168]

[0169] Therefore, the formula can be reformulated as a mixed-integer linear programming problem in the following form.

[0170]

[0171] According to the solution of SP, execute the following program.

[0172] The specific solution process is as Figure 3 shown:

[0173] Step 1: Set the lower bound LB = -∞, the upper bound UB = +∞, the convergence accuracy is ε, set the iteration number m = 1, and the initial probability distribution of the second stage is p s0 ;

[0174] Step 2: Solve MP to obtain the optimal solution (x * , η * ), and update the lower bound value LB = η * ;

[0175] Step 3: Based on the given first-stage variable x * , solve the sub-problem to obtain the worst-case probability value and the objective function value f(x * ), and at the same time update the upper bound value min{UB, f(x * )};

[0176] Step 4: If UB - LB < ε, stop the iteration and return the optimal solution x * ; Otherwise, update the worst-case probability distribution of the master problem and add new variables and related constraint conditions in the master problem; update m = m + 1 and return to Step 2.

[0177] In summary, considering data privacy and standardization, the present invention proposes an integrated load forecasting model incorporating the SHAP method and analyzes the load characteristics of users. Considering that it is usually difficult to obtain the exact probability distribution of data, a two-stage distributionally robust optimization scheduling model considering the comprehensive norm is proposed to solve the operation scheduling and source-load uncertainty problems of the solar photovoltaic-thermal comprehensive utilization system in solar-rich areas, broadening the research framework of distributed energy systems. At the same time, taking system operation and maintenance costs, carbon emissions and other indicators as optimization objectives, the C&CG method is used to solve the proposed model without the need for dualization. The feasibility of the method is verified using data obtained from the monitoring of the solar photovoltaic-thermal comprehensive utilization system. The optimization scheduling method proposed by the present invention utilizes the complementarity of the solar photovoltaic-thermal comprehensive utilization system, enabling users to actively participate in the response plan by converting electricity into heat during peak hours, where shiftable load, interruptible load and alternative load are considered.

[0178] Example 1

[0179] As Figure 4 shown in (a) and (b) therein, after analysis using the SHAP method, the importance rankings of the electrical load characteristics and thermal load characteristics of users in solar-rich areas are presented. For the characteristics affecting the electrical load, the previous moment's electrical load and time contribute relatively large values to the user's electrical load, which corresponds to the energy usage habits of users in this area. For the characteristics affecting the thermal load, time and the previous moment's total radiation contribute relatively large values to the user's thermal load, which conforms to the characteristics of people's demand for heat. Since most of the objects under study are residential buildings, therefore, whether it is the electrical load or the thermal load, time has a relatively large impact on the load.

[0180] In this example, five very popular prediction methods in the current prediction field are selected for comparison with the proposed integrated prediction model. Tables 1 and 2 show the comparison results of the proposed integrated prediction method and other methods. It can be seen from the tables that for both electrical load forecasting and thermal load forecasting, the proposed method has the best results in all evaluation indicators. Its RMSE values are reduced by 10.19%-32.72% and 11.35%-27.15% respectively. Its RMSE value is reduced by 10.19% and 11.35% respectively compared to RF, which ranks second, indicating the superiority of the proposed method. In addition, it can also be found from the prediction results that as an excellent representative of ensemble learning, RF also has better prediction accuracy than single prediction models.

[0181] Table 1 Comparison of the prediction accuracy of different models for electrical load

[0182]

[0183] Table 2 Comparison of the prediction accuracy of different models for thermal load

[0184]

[0185] Taking the power load prediction error as an example, the relative errors of 5 prediction models for each sampling point are plotted as a box plot, and the results are as Figure 5 shown. It can be seen that, compared with the other 4 models, the maximum error of the SHAP-XGBoost model in power load prediction is significantly lower than that of other models. The model not only has a smaller error distribution range, but also has a significantly smaller maximum error, further verifying the advantages of the proposed model in terms of prediction accuracy and robustness.

[0186] As Figure 6 shown, what is shown is the solar radiation intensity per hour. From Figure 7 it can be seen that for photovoltaic output, the optimal number of clusters is 3, and 3 typical scenarios are obtained through clustering. Since the solar radiation intensity is the decisive factor determining the photovoltaic output power and is also the main factor affecting the heat gain of solar collectors, therefore, the heat collection amount of the solar collector corresponding to the date when the typical photovoltaic power generation scenario occurs is used as the typical scenario of solar thermal for research. Figure 8 And Figure 9 respectively show the typical scenarios after clustering, and then the data and initial probabilities of the typical scenarios are provided to the proposed distributionally robust optimization model for solution.

[0187] The results of the power and heat dispatch of the solar photovoltaic-thermal integrated utilization system under typical scenarios are respectively as Figure 10 shown. From Figure 10It can be seen that during periods when renewable energy output is sufficient, the electrical load and thermal load are completely supplied by renewable energy (from 9:00 to 18:00). In terms of electrical energy, during peak load periods such as 8:00 - 11:00 and 20:00 - 23:00, the energy storage system preferentially releases electrical energy, but its capacity is limited, so the power grid is also needed to meet the power gap. According to the dispatching results, on the basis of giving priority to the consumption of renewable energy, the power generation and purchase costs of the integrated photovoltaic and solar thermal utilization system are reduced, and the two-way energy flow of the energy storage system also reduces the power supply pressure, thus effectively achieving a more economical system operation. In terms of heat energy, the heat storage device and the solar collector field contribute the most to the thermal load. This is because the heat storage device consists of 6 phase change heat storage tanks to fully absorb solar energy. During most of the day, the heat load is mainly borne by the solar collector field. It can be seen that the air source heat pump also participates in the heat supply because the excess photovoltaic power generation of the system is converted into heat supply. At night, the heat load is basically provided by the heat storage device. Given that the electrical load is at a low ebb, but the heat load is at a peak, solar collection cannot be applied, and the air source heat pump is put into operation at this stage to make up for the power gap of RG. It should be noted that this stage happens to be in the low electricity price period (from 4:00 to 7:00 and from 14:00 to 17:00). In addition, the energy storage system and the heat storage device, as buffer devices, also play an important role in the balance of power and heat supply and demand. From Figure 10 In the heat of Scenario 3, it can also be found that the capacity of the heat storage device becomes a limiting factor for the heat storage device system to play its role.

[0188] Figure 11 shows the electrical load and thermal load curves before and after implementing integrated demand response under typical scenarios. From Figure 11 It can be seen that for the electrical load, after applying integrated demand response, the electrical load curve tends to be smooth, achieving peak shaving and valley filling. During peak electricity consumption periods (8:00 - 12:00 and 18:00 - 21:00), under the influence of electricity prices, users tend to reduce electricity consumption by transferring and interrupting the electrical load. During low electrical load periods (4:00 - 7:00, 14:00 - 17:00), the lower electricity price leads to an increase in electricity consumption. In addition, from Figure 10 and Figure 11 it can be observed that during this stage, part of the photovoltaic power generation participates in the integrated demand response by being converted into heat. For the thermal load, it is worth noting that during 0:00 - 3:00, 8:00 - 9:00, and 19:00 - 23:00, although this stage does not belong to the low electricity price period, the system conducts an alternative load conversion response. This is mainly because during this stage, electrical load transfer response and electrical load interruption response are carried out. It can be seen from the figure that under the condition of the lowest total system cost, the purpose of increasing the heat supply at night can be achieved through dispatching, providing more heat for users.

[0189] As Figure 12 shown, another object of the present invention is to provide a distributionally robust optimal scheduling system for a solar photovoltaic and solar thermal integrated utilization system, including:

[0190] A data acquisition module, configured to obtain an electrical load and a thermal load based on an integrated load prediction model; obtain typical day photovoltaics and solar thermals based on scenario clustering analysis;

[0191] A scheduling model construction module, configured to set an objective function and constraint conditions for optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system based on photovoltaics, solar thermals, electrical load, and thermal load, and obtain a two-stage distributionally robust optimal scheduling model based on the objective function and constraint conditions in combination with a comprehensive demand response mechanism;

[0192] A scheduling result solving module, configured to solve the two-stage distributionally robust optimal scheduling model by using a column and constraint generation algorithm to obtain a two-stage distributionally robust optimal scheduling result.

[0193] As Figure 13 shown, the third object of the present invention is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the steps of the distributionally robust optimal scheduling method for the solar photovoltaic and solar thermal integrated utilization system are implemented.

[0194] The distributionally robust optimal scheduling method for the solar photovoltaic and solar thermal integrated utilization system includes:

[0195] Obtaining an electrical load and a thermal load based on an integrated load prediction model; obtaining typical day photovoltaics and solar thermals based on scenario clustering analysis;

[0196] Setting an objective function and constraint conditions for optimal scheduling of the solar photovoltaic and solar thermal integrated utilization system based on photovoltaics, solar thermals, electrical load, and thermal load, and obtaining a two-stage distributionally robust optimal scheduling model based on the objective function and constraint conditions in combination with a comprehensive demand response mechanism;

[0197] Solving the two-stage distributionally robust optimal scheduling model by using a column and constraint generation algorithm to obtain a two-stage distributionally robust optimal scheduling result.

[0198] The fourth object of the present invention is to provide a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the distributionally robust optimal scheduling method for the solar photovoltaic and solar thermal integrated utilization system are implemented.

[0199] The distributionally robust optimal scheduling method for the solar photovoltaic and solar thermal integrated utilization system includes:

[0200] Obtain the electrical load and thermal load based on the integrated load forecasting model; obtain the typical daily photovoltaic and solar thermal based on scenario clustering analysis;

[0201] Based on the photovoltaic, solar thermal, electrical load and thermal load, set the objective function and constraint conditions for the optimal scheduling of the solar photovoltaic and solar thermal comprehensive utilization system. Based on the objective function and constraint conditions, combined with the integrated demand response mechanism, obtain the two-stage distributionally robust optimal scheduling model;

[0202] Use the column and constraint generation algorithm to solve the two-stage distributionally robust optimal scheduling model to obtain the two-stage distributionally robust optimal scheduling result.

[0203] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0204] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0205] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device realizes the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0206] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide means for realizing the functions specified in one process Figure 1steps of one or more processes and / or blocks Figure 1 steps of functions specified in one or more blocks

[0207] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific implementation manners of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A distributed rod optimization scheduling method for a solar photovoltaic and thermal comprehensive utilization system, characterized in that: include: Obtain electrical and thermal loads based on an integrated load forecasting model; Obtain typical day photovoltaic and solar thermal data based on scene cluster analysis; Based on photovoltaic, solar thermal, electric load and heat load, the objective function and constraints of optimal scheduling are set for the solar photovoltaic and solar thermal comprehensive utilization system. Based on the objective function and constraints, combined with the comprehensive demand response mechanism, a two-stage distributed robust optimization scheduling model is obtained; The column and constraint generation algorithm is used to solve the two-stage distributed robust optimization scheduling model to obtain the two-stage distributed robust optimization scheduling results.

2. The distributed blue rod optimization scheduling method of a solar photovoltaic and thermal comprehensive utilization system according to claim 1 is characterized in that: The objective function of the optimal scheduling of the solar photovoltaic and thermal comprehensive utilization system is as follows: In the formula, p s is the probability of scene s after clustering; C PV-PT is the operation and maintenance cost of photovoltaic thermal components; is the cost of purchasing electricity from the main grid; C loss is the cost of abandoned solar and heat; C EES and C PCSD are the operating costs of the energy storage system and the thermal storage device; C CO2 is the penalty cost for carbon dioxide emissions; C IDR is the cost of comprehensive demand response; It is the profit of power grid electricity sales.

3. The distributed rod optimization scheduling method of a solar photovoltaic and thermal comprehensive utilization system according to claim 2 is characterized in that: The constraints include solar photovoltaic and thermal equipment operation constraints, interactive power constraints between the system and the distribution network, air source heat pump operation constraints, energy storage system operation constraints, thermal storage device operation constraints, power balance constraints and thermal balance constraints.

4. The distributed rod optimization scheduling method of a solar photovoltaic and thermal comprehensive utilization system according to claim 3 is characterized in that: Based on the objective function and constraints, combined with the comprehensive demand response mechanism, a two-stage distributed robust optimization scheduling model is obtained, specifically: The two stages of the two-stage distributed robust optimization scheduling model are: the first stage formulates a scheduling plan based on the main grid's power purchase and sales plan and energy storage charging and discharging; the second stage formulates a scheduling plan after the uncertainty of renewable energy and user energy consumption emerges; The two-stage distributed robust optimization scheduling model is: stBx≤a Yes s ≤b Of s =c Ex+Fy s ≤d In the formula, s represents the scene; p s is the probability of scene s after clustering; Ω p For p s Satisfies the probability distribution set interval, representing the confidence set constrained by the 1-norm and ∞-norm; N s is the total number of scenes after clustering; A~F and a~d are constant coefficient matrices; y s represents the second-stage variable; x represents the first-stage variable.

5. The distributed and robust optimization scheduling method for a solar photovoltaic and thermal comprehensive utilization system according to claim 4 is characterized in that: The first stage variable x included in the first stage is expressed as: The second stage variable y included in the second stage is expressed as follows: In the formula, and is the power purchase and sale status between the system and the main power grid at time t, and is the operating state of the energy storage system at time t, and is the operating state of the heat storage device at time t, P PV,s,t is the photovoltaic power during period t; Q PT,s,t is the solar thermal power in period t; and are the electricity purchased from and sold to the grid by the system during period t; P ASHP,s,t is the electric power of the air source heat pump at time t; Q ASHP,s,t is the heat power output of the air source heat pump at time t; and are the charging and discharging power of the energy storage system, respectively; and are the heat storage and heat release powers of the heat storage device respectively; is the replaceable electric power at time t, is the replaceable thermal power at time t, is the time-shifted load in period t, is the interruptible load in period t; and P IDR,t are the electric loads before and after the integrated demand response, and Q IDR,t The tables show the heat loads before and after integrated demand response.

6. The distributed blue rod optimization scheduling method of a solar photovoltaic and thermal comprehensive utilization system according to claim 4 is characterized in that: The p s Satisfy the probability distribution set interval Ω p In order to conform to the actual operation data and fluctuate within a reasonable range, a probability distribution value centered on the initial probability distribution value of each scenario is constructed, and the probability distribution value of photovoltaic and solar thermal scenarios is restricted by a set of 1-norm and ∞-norm, which is expressed as: In the formula, p s0 is the initial probability value of the sth discrete scenario obtained by screening the available historical data of the scheduling system; θ1 and θ ∞ are the probability allowable deviation limits under 1-norm and ∞-norm constraints respectively; The probability distribution p of the clustered scene s s The following confidence levels are met: Where Pr{·} is the probability operator. The above two formulas can be used to obtain the probability allowable deviation value under the two norm constraints, which is then used to limit the fluctuation range of the scene probability distribution; Ns is the total number of scenes after clustering.

7. The distributed rod optimization scheduling method of a solar photovoltaic and thermal comprehensive utilization system according to claim 4 is characterized in that: The column and constraint generation algorithm is used to solve the two-stage distributed robust optimization scheduling model to obtain the two-stage distributed robust optimization scheduling results, specifically: The column and constraint generation algorithm is used to divide the solution of the two-stage distributed robust optimization scheduling model into the main problem MP and the sub-problem SP; Set the lower bound LB = -∞, the upper bound UB = +∞, the convergence accuracy to ε, the iteration number m = 1, and the initial probability distribution of the second stage to p s0 ; Solve the main problem and get the optimal solution (x * ,η * ), and update the lower bound LB = η * ; The main problem is expressed as follows: In the formula, m represents the number of iterations, M represents the maximum number of iterations, and η is an introduced auxiliary variable; Based on solving the main problem, we can get the optimal solution x * , solve the subproblem and obtain the probability value in the worst case And the objective function value f(x * ), and update the upper bound min{UB,f(x * )}; The sub-problem is expressed as follows: In the formula, represents the two-stage decision variable in the mth iteration, represents the solution to the internal "min" problem, represents the probability value of the mth iteration; When UB-LB < ε, stop iteration and return the optimal solution x * ; Conversely, update the worst probability distribution of the main problem And add new variables to the main problem and constraints; update m=m+1 and return to solving the main problem for iteration.

8. A distributed blue rod optimization scheduling system for a solar photovoltaic and thermal comprehensive utilization system, characterized in that: include: A data acquisition module, used for acquiring electric load and thermal load based on an integrated load prediction model; Obtain typical day photovoltaic and solar thermal data based on scene cluster analysis; The scheduling model building module is used to set the objective function and constraints of the optimal scheduling of the solar photovoltaic and thermal comprehensive utilization system based on photovoltaic, thermal, electrical and thermal loads. Based on the objective function and constraints, combined with the comprehensive demand response mechanism, a two-stage distributed robust optimization scheduling model is obtained; The scheduling result solving module is used to solve the two-stage distributed robust optimization scheduling model by adopting the column and constraint generation algorithm to obtain the two-stage distributed robust optimization scheduling result.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the distributed robust optimization scheduling method for the solar photovoltaic and thermal comprehensive utilization system as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the distributed rod optimization scheduling method for the solar photovoltaic and thermal comprehensive utilization system according to any one of claims 1 to 7.

Images