Urban heat supply system boundary condition optimization method and device based on multi-energy complementation

Abstract

The invention provides an urban heat supply system boundary condition optimization method and device based on multi-energy complementation, and the method comprises the steps: constructing a seed scene library containing scene feature vectors based on the obtained thermal load characteristics, meteorological conditions and energy prices of an urban heat supply system; based on the seed scene library, analyzing and determining a coding dimension and a variable range, and constructing a multi-objective optimization coding scheme comprising an economical efficiency objective function, an environmental protection objective function and a system stability objective function; calculating an initial value of each objective function based on a multi-objective optimization coding scheme, and performing iterative calculation by adopting a multi-objective optimization algorithm to obtain an optimization scheme; and based on the optimization scheme, constructing a boundary condition verification model for simulation verification, and dynamically correcting the boundary condition according to a verification result. According to the method, the limitation that a traditional method depends on a large amount of historical data is broken through, the optimization efficiency of the multi-energy complementary heat supply system is improved, and effective balance among multiple targets of economical efficiency, environmental protection performance and system stability is achieved.

Description

Technical Field

[0001] The present invention relates to the field of urban heating systems, and particularly to an optimization method and device for the boundary conditions of an urban heating system based on multi - energy complementarity. Background Art

[0002] The urban heating system is an important part of urban infrastructure, and its operating efficiency and economy directly affect the quality of life of urban residents and energy consumption. With the diversified development of the energy structure, the multi - energy complementary heating system has gradually become the main development direction of urban heating.

[0003] Currently, the common urban heating systems mainly include traditional coal - fired boiler heating and natural gas boiler heating. These single - energy heating systems have problems such as low energy utilization efficiency and serious environmental pollution. In some areas, renewable energy sources such as heat pumps and solar energy have begun to be used as supplementary heat sources, but there is a lack of systematic multi - energy complementary optimization schemes.

[0004] The more advanced multi - energy complementary heating systems adopt the method of coordinated heating with multiple energy sources and optimize the system operation by establishing a mathematical model. Such systems usually build an optimization model based on historical data and use traditional numerical optimization methods to solve the system boundary conditions.

[0005] However, the existing technologies have the following problems when dealing with the optimization of the boundary conditions of the multi - energy complementary heating system: 1) The optimization process requires a large amount of historical data support, and it is often difficult to obtain complete historical operation data in actual projects; 2) Traditional optimization methods are difficult to effectively handle the trade - off between multiple objectives, especially when considering multiple objectives such as economy, environmental protection, and system stability.

[0006] Therefore, a new optimization method is needed, which can effectively balance multiple optimization objectives under the condition of limited historical data and provide the optimal boundary conditions for the multi - energy complementary urban heating system. Summary of the Invention

[0007] The purpose of the present invention is to provide an optimization method and device for the boundary conditions of an urban heating system based on multi - energy complementarity, so as to solve the technical problems that the optimization process in the existing technology requires a large amount of historical data support and traditional optimization methods are difficult to effectively handle the trade - off between multiple objectives.

[0008] To achieve the above purpose, the present invention provides an optimization method for the boundary conditions of an urban heating system based on multi - energy complementarity, including the following steps: Based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, construct a seed scenario library containing scenario feature vectors; Based on the seed scenario library, analyze and determine the coding dimension and variable range, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function; Based on the multi-objective optimization coding scheme, calculate the initial values of each objective function, and use a multi-objective optimization algorithm for iterative calculation to obtain an optimization scheme; Based on the optimization scheme, construct a boundary condition verification model for simulation verification, and dynamically correct the boundary conditions according to the verification results; The boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; the operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperature, flow rate, and pipe network pressure.

[0009] Preferably, based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, construct a seed scenario library including scenario feature vectors, including: According to the heat load characteristics, meteorological conditions, and energy prices of the urban heating system, obtain key parameters; Based on the key parameters, use an improved K-means clustering algorithm to classify heating conditions and identify typical heating conditions; Extract features from the typical heating conditions, and establish a scenario feature vector including load features, environmental features, and economic features; Based on the scenario feature vector, construct a seed scenario library, and establish a scenario evaluation system including representative indicators, completeness indicators, and independence indicators to obtain the seed scenario library including scenario feature vectors; Among them, the representative indicator is used to evaluate the representation ability of the scenario to the actual operation state, the completeness indicator is used to evaluate the coverage of the scenario library to the system operation space, and the independence indicator is used to evaluate the information redundancy degree between scenarios.

[0010] Preferably, based on the seed scenario library, analyze and determine the coding dimension and variable range, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function, including: Analyze the scenario features in the seed scenario library including scenario feature vectors, and determine the coding dimension and variable range; Based on the coding dimension and variable range, design a coding structure suitable for the characteristics of the multi-energy complementary system, including the operation parameters of each energy subsystem; Based on the coding structure, construct an economic objective function, an environmental protection objective function, and a system stability objective function; Standardize the economic objective function, environmental protection objective function, and system stability objective function, and allocate weights to form a multi-objective optimization coding scheme including the economic objective function, environmental protection objective function, and system stability objective function.

[0011] Preferably, based on the coding structure, construct the economic objective function, including: Based on the coding structure, calculate the initial investment cost, operating cost, and maintenance cost of the system to obtain cost data; Based on the cost data, perform annualized processing to obtain a cost index with a unified dimension; Based on the cost index with a unified dimension, construct a calculation model including time-of-use electricity price and seasonal natural gas price to obtain the economic objective function.

[0012] Preferably, based on the coding structure, construct the environmental protection objective function, including: Based on the coding structure, calculate the carbon emissions and pollutant emissions of different energy forms to obtain environmental load data; Based on the environmental load data, perform equivalent carbon emissions conversion to obtain an environmental impact index; Based on the environmental impact index, construct a calculation model including carbon emissions and pollutant emissions to obtain the environmental protection objective function.

[0013] Preferably, based on the coding structure, construct the system stability objective function, including: Based on the coding structure, calculate the heating temperature fluctuation, heating continuity, and equipment start-stop frequency to obtain stability data; Based on the stability data, perform risk assessment to obtain a system reliability index; Based on the system reliability index, construct a calculation model including heating reliability and equipment life to obtain the system stability objective function.

[0014] Preferably, based on the multi-objective optimization coding scheme, calculate the initial values of each objective function, and perform iterative calculations using a multi-objective optimization algorithm, including: Set the population size, upper limit of the number of iterations, crossover probability, and mutation probability to obtain optimization parameters; Randomly generate an initial population, and calculate the initial values of each objective function based on the multi-objective optimization coding scheme; Based on the initial values of each objective function and the optimization parameters, perform iterative calculations using an improved NSGA-III algorithm, where the improvement includes introducing a local search strategy based on seed scenarios, designing adaptive crossover and mutation operators, and introducing an elite retention strategy to obtain a non-dominated solution set; Update the non-dominated solution set to generate the Pareto optimal solution set; Based on the actual requirements of the system, screen out the optimization plan from the Pareto optimal solution set.

[0015] Preferably, screening out the optimization plan from the Pareto optimal solution set based on the actual requirements of the system includes: Calculate the similarity index and crowding degree index of the solutions in the Pareto optimal solution set to obtain solution set evaluation data; Based on the solution set evaluation data, adopt a hierarchical screening strategy to eliminate solutions with a similarity greater than a preset threshold and solutions that do not meet the engineering implementation conditions to obtain candidate solutions; Based on the candidate solutions, select the optimization plan according to the decision maker's preferences for economy, environmental protection, and stability.

[0016] Preferably, based on the optimization plan, construct a boundary condition verification model for simulation verification, and dynamically correct the boundary conditions according to the verification results, including: Based on the boundary condition verification model, set static indicators and dynamic indicators to obtain verification parameters; Based on the verification parameters, adopt a hierarchical simulation strategy to perform steady-state simulation and dynamic simulation to obtain simulation data; Based on the simulation data, perform statistical analysis and sensitivity analysis to obtain boundary conditions that do not meet the engineering constraints; Based on the boundary conditions that do not meet the engineering constraints, adjust the constraint range and safety margin to obtain correction parameters; Based on the correction parameters, perform iterative optimization and balance the system reliability and economy to obtain a correction plan; Based on the correction plan, dynamically correct the boundary conditions.

[0017] The present invention also provides a boundary condition optimization device for a multi-energy complementary urban heating system, including: A construction module for constructing a seed scenario library containing scenario feature vectors based on the heat load characteristics, meteorological conditions, and energy prices of the obtained urban heating system; An analysis module for analyzing and determining the coding dimension and variable range based on the seed scenario library, and constructing a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function; A calculation module for calculating the initial values of each objective function based on the multi-objective optimization coding scheme, and performing iterative calculations using a multi-objective optimization algorithm to obtain an optimization plan; A correction module for constructing a boundary condition verification model for simulation verification based on the optimization plan, and dynamically correcting the boundary conditions according to the verification results; The boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; the operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperatures, flow rate, and pipe network pressure.

[0018] The beneficial effects of the present invention are as follows: 1. The optimization method based on seed scenarios proposed by the present invention breaks through the limitation of traditional methods relying on a large amount of historical data. By constructing a representative seed scenario library, effective optimization of the multi-energy complementary heating system is achieved under the condition of limited data. 2. The present invention constructs a seed scenario library containing scenario feature vectors based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, breaking through the limitation of traditional methods requiring a large amount of historical data. The effective characterization of system characteristics is realized through representative scenarios, improving the optimization efficiency. 3. The present invention constructs a multi-objective function system including economy, environmental protection, and system stability, comprehensively considering various factors such as the investment cost, operation cost, carbon emissions, pollutant emissions, heating reliability, and equipment life of the system. 4. The present invention constructs a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function, and performs iterative calculations through a multi-objective optimization algorithm to achieve an effective balance among the multi-objectives of economy, environmental protection, and system stability. 5. The present invention introduces a dynamic correction mechanism, conducts simulation verification by constructing a boundary condition verification model, and dynamically corrects the boundary conditions according to the verification results to ensure that the optimization results meet the actual engineering requirements. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those skilled in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of the method for optimizing the boundary conditions of the urban heating system provided by the embodiment of the present invention; Figure 2 It is a flowchart of the execution of the iterative optimization process provided by the embodiment of the present invention; Figure 3 It is a structural block diagram of the device for optimizing the boundary conditions of the urban heating system provided by the embodiment of the present invention. Detailed Embodiments

[0021] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0022] Please refer to Figure 1 , Figure 1 , which is a flowchart of the boundary condition optimization method for an urban heating system based on multi - energy complementarity provided by an embodiment of the present invention. As Figure 1 shown, the method includes the following steps: Step S101: Based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, construct a seed scenario library containing scenario feature vectors.

[0023] In this step, first, key parameter information of the urban heating system needs to be collected, including basic data such as heat load characteristics, meteorological conditions, and energy prices. Heat load characteristic data mainly includes daily load curves, peak - valley ratios, load change rates, etc., which can be obtained through on - site measurements or historical operation records; meteorological condition data includes environmental parameters such as outdoor temperature, humidity, wind speed, etc., which can be obtained from meteorological departments or collected through on - site weather stations; energy price data includes electricity prices, natural gas prices, heating prices, etc., which can be obtained from energy suppliers or queried through policy documents. For example, the operation data of typical days (such as the coldest day, transitional season day, etc.) during the heating season can be selected as the basic data set to ensure the representativeness and integrity of the data.

[0024] After obtaining the basic data, an improved K - means clustering algorithm is used to classify the heating conditions and identify representative typical heating conditions. Key factors such as heat load magnitude, outdoor temperature range, and energy price combination are mainly considered during the clustering process, and similar operating conditions are grouped into one category. The improved K - means algorithm improves the accuracy and efficiency of clustering by optimizing the selection of the initial clustering center and introducing an adaptive distance metric. Through clustering analysis, typical condition categories such as "high load - low temperature - peak - time electricity price" and "medium load - normal temperature - valley - time electricity price" may be obtained, and these categories can represent the operating states of the system under different conditions.

[0025] Feature extraction is performed on the identified typical conditions to establish scenario feature vectors. The feature vectors contain key information characterizing the operating state of the heating system, mainly including load characteristics (such as load magnitude, change rate), environmental characteristics (such as temperature, humidity), economic characteristics (such as energy price combination), equipment characteristics (such as equipment efficiency, operating constraints), etc. These features are stored in a standardized numerical form for subsequent optimization calculations. The feature extraction process uses dimensionality reduction techniques such as principal component analysis (PCA) to retain the most representative features and reduce data redundancy.

[0026] Finally, based on the feature vectors, a seed scenario library is constructed, and a scenario evaluation index system is established. The seed scenario library is a structured data set, and each scenario contains complete feature vector information. At the same time, an evaluation index system is established, including the representativeness index of the scenario (reflecting the representation ability of the scenario for the actual operating state), the completeness index (reflecting the coverage of the scenario library for the system operation space), the independence index (reflecting the information redundancy degree between scenarios), etc. Through these indexes, the quality of the seed scenario library can be evaluated, and the optimization and update of the scenario library can be guided. For example, the completeness can be evaluated by calculating the scenario coverage rate (the ratio of the covered operating conditions range to the total operating conditions range), and the independence can be evaluated by calculating the similarity matrix between scenarios. The construction process of the seed scenario library is iterative and needs to be continuously adjusted and optimized according to the evaluation indexes until the preset quality standard is met.

[0027] Step S102: Based on the seed scenario library, analyze and determine the coding dimension and variable range, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function.

[0028] In this step, it is first necessary to analyze the scenario characteristics in the seed scenario library to determine the coding dimension and variable range of the optimization problem. In a multi-energy complementary heating system, the variable dimensions that need to be coded usually include the operating parameters of each energy subsystem, the configuration parameters of the system, and the control strategy parameters. Specifically, the variable dimensions include: the heating source configuration dimension (such as the capacity of heat pumps, boilers, heat storage devices, etc.), the operation control dimension (such as the start-stop strategy of each heat source, the load distribution ratio, the heat storage and release strategy, etc.), and the system parameter dimension (such as the supply and return water temperature, flow rate, pipe network pressure, etc.). When defining the variable range for each dimension, it is necessary to comprehensively consider the physical constraints of the equipment, the actual engineering limitations, and the economic rationality. For example, the capacity range of the heat pump needs to consider the specifications provided by the manufacturer and the actual installation space limitations; the supply and return water temperature needs to consider the user comfort requirements, the pipe network pressure bearing capacity, and the heat source outlet temperature limitations; the load distribution ratio needs to consider the part-load characteristics of the equipment and the system stability requirements.

[0029] Based on the determined coding dimensions and variable ranges, a coding structure adapted to the characteristics of the multi-energy complementary system is designed. Considering that the multi-energy complementary system includes discrete variables and continuous variables, a hybrid coding structure is adopted: integer coding is used for discrete variables (such as equipment switch states, operation mode selections, etc.); real number coding is used for continuous variables (such as load distribution ratios, temperature set values, etc.). When designing the coding structure, special attention is paid to the correlation between variables, and closely related variables are coded in adjacent positions to improve the search efficiency of the optimization algorithm. For example, the start-stop state of a certain heat pump unit and its load distribution ratio should be coded adjacent to each other because these two variables have a direct logical correlation. In addition, the coding structure also needs to consider the implicit expression of constraint conditions. For example, the heat supply balance constraint is naturally satisfied through the coding method to avoid generating invalid solutions. The design of the coding structure directly affects the efficiency and result quality of the subsequent optimization algorithm, so it needs to be repeatedly tested and adjusted to ensure its adaptability and effectiveness.

[0030] Next, an economic objective function is constructed based on the coding structure. The economic objective function is a comprehensive cost function that needs to consider the life-cycle cost of the system. Specifically, it includes: initial investment cost (including the purchase cost, installation cost, auxiliary facility cost, etc. of various heating equipment), operation cost (including energy consumption cost, maintenance cost, labor cost, etc.), depreciation cost, and scrap disposal cost. These cost items are unified in dimension through annualization processing. For example, the initial investment is converted into an annualized cost according to the service life of the equipment and the discount rate. The construction of the objective function adopts the method of weighted summation, and the weight coefficients are determined through expert evaluation and sensitivity analysis. It should be particularly noted that the dynamic change characteristics of different energy prices also need to be considered in the economic objective function. For example, factors such as time-of-use electricity price and seasonal natural gas price are introduced to make the optimization results more in line with the actual operation situation. The mathematical expression of the economic objective function can adopt indicators such as net present value (NPV) or equivalent uniform annual cost (EUAC) to ensure the scientificity and accuracy of economic evaluation.

[0031] Meanwhile, construct an environmental protection objective function. The environmental protection objective function mainly considers the impact of system operation on the environment, including direct and indirect environmental loads. It mainly includes three aspects: First, the carbon emission index, which needs to consider the life-cycle carbon emissions of different energy forms, including direct emissions (such as emissions during the operation of gas boilers) and indirect emissions (such as emissions caused by using electricity); Second, the conventional pollutant emission index, including nitrogen oxides, sulfur oxides, soot, etc. These indexes need to be dynamically calculated according to the emission characteristics and operating conditions of different equipment; Third, the environmental impact index, including noise pollution, heat pollution, etc. These indexes need to be normalized through weight coefficients to form a unified environmental protection evaluation index. For example, carbon emission equivalents can be used as the conversion benchmark to convert different types of environmental impacts into equivalent carbon emissions. The construction of the environmental protection objective function needs to refer to the latest environmental protection standards and emission factors to ensure the scientificity and timeliness of the evaluation.

[0032] In addition, construct a system stability objective function. Stability is a key indicator to measure the reliable operation ability of the heating system, including the following dimensions: heating reliability (used to evaluate the ability of the system to meet the heating demand of users, including heating temperature fluctuations, heating continuity, etc.), equipment life (considering the impact of equipment start-stop frequency and load change rate on the service life of equipment), and system safety (evaluating the safety margin of system operation, including the fluctuation range of key parameters such as pressure and temperature). When constructing the stability objective function, a risk assessment-based method is adopted to quantify various risk factors into computable indexes. For example, the impact on the service life of equipment can be evaluated through parameters such as the start-stop times of equipment and the change speed of load rate, and the heating stability can be evaluated through the standard deviation of heating parameters. These indexes also need to be normalized through weight coefficients to form a unified stability evaluation index. The construction of the stability objective function needs to combine actual engineering experience and equipment characteristics to ensure the practicality and accuracy of the evaluation.

[0033] Finally, integrate each objective function to form a unified multi-objective optimization coding scheme. The integration process adopts a multi-level structure: The first layer is the standardization of the objective function, which converts the objective functions with different dimensions into dimensionless standard scores; the second layer is the weight allocation of the objective function, and the weight coefficients of each objective function are determined through the Analytic Hierarchy Process (AHP); the third layer is the processing of constraint conditions, encoding various hard constraints (such as equipment operation boundaries) and soft constraints (such as priority requirements) into the optimization scheme. The finally formed multi-objective optimization coding scheme is a structured mathematical description, which not only includes the calculation methods of each objective function, but also includes various constraint conditions and priority rules that need to be considered during the optimization process. The quality of the multi-objective optimization coding scheme directly affects the effectiveness of the subsequent optimization results. Therefore, it needs to be fully verified and tested to ensure that it can accurately reflect the actual characteristics and optimization objectives of the system.

[0034] Step S103: Based on the multi-objective optimization coding scheme, calculate the initial values of each objective function, and use the multi-objective optimization algorithm for iterative calculation to obtain the optimization scheme.

[0035] In this step, it is first necessary to initialize the optimization parameters and set the iteration conditions. Set the key parameters for the multi-objective optimization algorithm: the population size is usually set to 2-3 times the coding dimension to ensure sufficient coverage of the search space. For example, for a 30-dimensional coding space, the population size can be set to 60-90; the upper limit of the number of iterations is estimated based on the problem size and convergence speed, usually set to 500-1000 times, and more iterations may be required for complex problems; the crossover probability is set between 0.7-0.9 to ensure population diversity; the mutation probability is set between 0.1-0.2 to avoid falling into local optima. The setting of the convergence condition includes two aspects: one is that the change amplitude of the optimal solution for several consecutive generations (such as 20 generations) is less than the preset threshold (such as 0.1%); the other is that the advancement speed of the Pareto front is lower than a certain threshold. The initial values of these parameters can be adaptively adjusted based on the characteristics of the seed scenario. For example, for scenarios with high complexity, the population size and the number of iterations can be increased to ensure that the algorithm can maintain good performance in different scenarios.

[0036] Next, randomly generate the initial population, and calculate the initial values of each objective function based on the coding scheme. The generation of the initial population needs to consider the constraint range and mutual relationship of the variables to ensure the feasibility of the initial solution. For each individual (i.e., a complete operation scheme of the heating system), calculate its economic objective function value (including investment cost and operation cost), environmental protection objective function value (including carbon emissions and pollutant emissions), and stability objective function value (including reliability index and life index). The calculation process needs to consider the specific conditions in the seed scenario, such as load characteristics, meteorological conditions, etc. For example, when calculating the economic index, it is necessary to calculate the energy consumption cost according to the energy price and load characteristics in the scenario; when calculating the environmental protection index, it is necessary to calculate the emissions according to the operating conditions in the scenario; when calculating the stability index, it is necessary to evaluate the system response ability according to the load change characteristics in the scenario. These initial values will be used as the benchmark points for subsequent iterative optimization and are also used to evaluate the convergence performance of the optimization algorithm.

[0037] Then, an improved multi-objective optimization algorithm is used for iterative calculation. In this step, the improved NSGA-III algorithm (Non-dominated Sorting Genetic Algorithm III) is adopted, which is particularly suitable for dealing with optimization problems with three or more objectives. The improvement is mainly reflected in three aspects: First, a local search strategy based on seed scenarios is introduced to preferentially explore the solution space similar to the seed scenarios in each iteration, improving the search efficiency; Second, adaptive crossover and mutation operators are designed to dynamically adjust parameters according to the population diversity. For example, when the population diversity decreases, the mutation probability is increased; Third, an elite retention strategy is introduced to ensure that high-quality solutions will not be lost during the evolution process. In each iteration, the algorithm updates the non-dominated solution set and maintains the diversity of the population through the reference point method. During the iterative process, the algorithm continuously evaluates the quality of the current solution and adjusts the search direction and step size according to the evaluation results, gradually approaching the Pareto optimal front.

[0038] As the iteration progresses, a Pareto optimal solution set is generated. The algorithm continuously updates and maintains a non-dominated solution set, and each solution in this solution set represents different trade-off schemes for the three objectives of economy, environmental protection, and stability. To ensure the quality of the solution set, solution screening and evaluation are required: First, solutions with too high similarity are removed to maintain the diversity of the solution set. For example, a similarity threshold can be set, and when the similarity between two solutions exceeds the threshold, only one of them is retained; Then, the crowding degree index of each solution is calculated to ensure that the solutions are evenly distributed in the objective space and avoid the solution set being too concentrated in certain regions; Finally, the practicability of the solutions is evaluated, and solutions that are mathematically feasible but difficult to implement in engineering practice are excluded. For example, solutions with frequent equipment start-stop or excessive load changes may need to be excluded. Through these screening and evaluation steps, a high-quality Pareto optimal solution set can be obtained, providing sufficient alternative solutions for subsequent solution selection.

[0039] Finally, according to the actual requirements of the system, the final optimized solution is screened from the Pareto solution set. This step needs to consider the preferences of decision-makers and actual engineering constraints. The screening process adopts a hierarchical screening strategy: The first layer is screening based on hard constraints, excluding solutions that do not meet key engineering requirements, such as solutions that exceed the equipment operation range or do not meet safety standards; The second layer is screening based on soft constraints, considering factors such as the implementation difficulty and maintenance convenience of the solutions. For example, solutions with simple control logic and low maintenance costs may be preferred; The third layer is screening based on decision-making preferences, and the final solution is selected according to the importance attached by the decision-maker to different objectives. For example, if the decision-maker attaches more importance to economy, a solution with the optimal economic index can be selected on the premise of meeting the basic environmental protection and stability requirements; If more importance is attached to environmental protection, a solution with the lowest carbon emissions can be selected. The screening process can adopt multi-criteria decision-making methods such as TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) or PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluations) to assist decision-makers in selecting the most suitable optimized solution from numerous alternative solutions.

[0040] Step S104: Based on the optimization scheme, construct a boundary condition verification model for simulation verification, and dynamically correct the boundary conditions according to the verification results; the boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; the operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperatures, flow rates, and pipe network pressures.

[0041] In this step, first, it is necessary to construct a boundary condition verification model and set verification indicators. The construction of the boundary condition verification model needs to consider constraints at multiple levels: physical constraints (such as the laws of thermodynamics, energy conservation, etc.), equipment constraints (such as equipment operation range, start-stop characteristics, etc.), and system constraints (such as hydraulic balance and thermal balance of the heating network, etc.). The verification model should be able to accurately reflect the physical characteristics and operation laws of the actual system, and professional thermal system simulation software or self-built mathematical models can be used. The verification index system includes two categories: static indicators and dynamic indicators: Static indicators mainly verify the performance of the system during steady-state operation, such as equipment load rate (usually required to be between 30% - 90%), system energy efficiency (such as primary energy utilization rate, COP value, etc.), economic indicators (such as operation cost, payback period, etc.); Dynamic indicators focus on the response characteristics of the system when the operating conditions change, such as temperature adjustment rate (usually required that the room temperature change does not exceed 2°C / hour), pressure fluctuation range (usually required not to exceed ±10% of the design pressure), equipment start-stop frequency (usually required that the main equipment such as heat pumps start and stop no more than 6 times a day), etc. Clear calculation methods and judgment criteria need to be established for these indicators to form a quantifiable verification system.

[0042] Next, conduct simulation verification on the optimization plan. This step adopts a hierarchical simulation strategy: First, conduct steady-state simulation to verify the performance indicators of the system under various typical operating conditions, such as the operating parameters and performance at the design condition, partial load condition, and extreme condition; then conduct dynamic simulation, focusing on the transient characteristics of the system during the operating condition transition, such as the switching process from low load to high load, the conversion process from one energy mode to another energy mode, etc. Pay special attention to the following points during the simulation: First, consider the start-stop process and switching losses of the equipment, including start-up time, preheating time, switching energy consumption, etc.; second, simulate the system response under extreme conditions, such as the system performance in extremely cold weather or equipment failure; third, verify the feasibility of the control strategy to ensure that the control logic can be effectively executed in the actual system. The simulation uses professional thermal system simulation software, such as TRNSYS, EnergyPlus, etc., and sets simulation parameters in combination with actual engineering experience. For example, when verifying the dynamic characteristics of the heating system, extreme situations such as the outdoor temperature dropping suddenly from -5°C to -15°C within 24 hours or the user demand increasing by 30% in a short period of time can be simulated to observe the adaptability and stability of the system.

[0043] Then, analyze the verification results and identify the boundary conditions that do not meet the engineering constraints. This step first conducts a systematic analysis of the simulation results to identify the boundary conditions that may have problems. The analysis methods include: statistical analysis (calculating the statistical characteristics of each index, such as mean, standard deviation, maximum value, minimum value, etc.), sensitivity analysis (studying the influence degree of each parameter change on the system performance and identifying key parameters), and limit analysis (exploring the boundary conditions of the system performance and determining the safety margin). Through these analyses, potential problems can be found, such as the equipment load being too high under certain operating conditions (such as the load rate of the heat pump exceeding 95% in extremely cold weather), the system response being too slow (such as the room temperature adjustment speed not meeting the user comfort requirements), the energy efficiency not meeting the standard (such as the COP value of the system decreasing significantly under partial load), etc. Pay special attention to those boundary conditions that are easily overlooked in actual engineering, such as the partial load characteristics of the equipment, the safety margin of the system, and the robustness of the control strategy. Through systematic analysis, the feasibility and potential risks of the optimization plan can be comprehensively evaluated, providing a basis for subsequent corrections.

[0044] Next, based on the verification results, the boundary conditions are dynamically corrected. This step adopts an iterative optimization method to correct each problem found one by one. The correction strategies include: adjusting the constraint range (such as relaxing or tightening the value range of certain parameters, for example, adjusting the maximum load rate of the heat pump from 95% to 85% to increase the safety margin), increasing the safety margin (such as considering more conservative design boundaries, for example, adding 10% of the standby capacity based on the design capacity), and optimizing the control strategy (such as improving the start-stop logic of the equipment, for example, introducing a predictive control algorithm to respond to load changes in advance). The correction process needs to balance multiple objectives: ensuring the safety and reliability of the system while maintaining the economy of the optimization plan. For example, if it is found that the equipment load rate is too high under certain working conditions, it can be solved by increasing the standby capacity or optimizing the load distribution strategy, but at the same time, the impact of these modifications on the overall system performance needs to be evaluated to ensure that the investment cost is not significantly increased or the operating efficiency is not reduced. The correction process is iterative, and after each correction, the simulation verification needs to be carried out again until all indicators meet the requirements.

[0045] Finally, output the final optimized boundary condition scheme. The final scheme needs to include a complete technical document, which details: the optimized system configuration parameters (such as the capacity and operating range of each device), the control strategy parameters (such as the start-stop threshold and adjustment curve), and the operating boundary conditions (such as the allowable working condition range and safety limits). At the same time, a feasibility demonstration of the scheme needs to be provided, including technical feasibility (such as the availability of equipment and the maturity of technology), economic feasibility (such as the investment payback period and operating cost analysis), and engineering implementation suggestions (such as construction key points and commissioning methods). Particularly important is to clearly state the applicable conditions and limitations of the scheme to provide clear guidance for subsequent engineering implementation. For example, it is necessary to explain the adaptability of the scheme under different climate conditions (such as cold regions and mild regions), different load characteristics (such as residential areas and commercial areas), and the possible commissioning and optimization suggestions. The final scheme should be fully verified and optimized, capable of being effectively implemented in actual projects and achieving the expected economic, environmental, and stability goals.

[0046] In this embodiment, step S101 specifically includes: Step S201: Obtain key parameters according to the heat load characteristics, meteorological conditions, and energy prices of the urban heating system.

[0047] Step S202: Based on the key parameters, use an improved K-means clustering algorithm to classify the heating working conditions and identify typical heating working conditions.

[0048] Step S203: Extract the characteristics of the typical heating working conditions and establish a scenario feature vector including load characteristics, environmental characteristics, and economic characteristics.

[0049] Step S204: Construct a seed scenario library based on the scenario feature vectors, and establish a scenario evaluation system including representative indicators, completeness indicators, and independence indicators to obtain the seed scenario library containing the scenario feature vectors.

[0050] Among them, the representative indicator is used to evaluate the representation ability of the scenario for the actual operating state, the completeness indicator is used to evaluate the coverage of the scenario library for the system operating space, and the independence indicator is used to evaluate the information redundancy degree among scenarios.

[0051] In this technical solution, a "seed scenario" refers to the smallest set of scenarios that can represent the typical operating state of the system. These scenarios, like "seeds", contain the key features of the system operation, and more operating scenarios can be deduced through them. The advantage of seed scenarios is that only a small number of representative scenarios are needed to achieve system optimization, greatly reducing data dependence.

[0052] Step S201 first collects the key parameters of the urban heating system. This includes basic data such as heat load characteristics (such as daily load curve, peak-valley ratio, etc.), meteorological conditions (such as outdoor temperature, humidity, wind speed, etc.), and energy prices (such as electricity price, natural gas price, heating price, etc.). These data are obtained through on-site collection, historical record query, energy price policy analysis, etc. For example, the operating data of typical days (such as the coldest day, transitional season day, etc.) in the heating season can be selected as the basic data set.

[0053] After obtaining the basic data in Step S202, the clustering analysis method is used to identify typical heating conditions. Specifically, the improved K-means clustering algorithm is adopted to classify similar operating conditions. When clustering, key factors such as the magnitude of the heat load, the range of outdoor temperature, and the combination of energy prices are mainly considered. For example, typical condition categories such as "high load - low temperature - peak-time electricity price" and "medium load - normal temperature - valley-time electricity price" may be obtained. The purpose of this step is to extract the most representative condition combinations from a large number of possible operating conditions.

[0054] The improved K-means clustering algorithm is used to classify the heating conditions and identify typical heating conditions. The improved K-means algorithm is mainly optimized in the following aspects: First, the selection of the initial clustering center is improved. In the traditional K-means algorithm, the initial center point is randomly selected, which easily leads to unstable clustering results and is prone to falling into local optima. The present invention adopts a density-weighted selection method, preferentially selecting points in data-dense regions as the initial centers. In specific implementation, first, the local density index of each sample point is calculated, which is calculated based on the reciprocal of the average distance of the K nearest neighbor points. For heating condition data, the value of K is usually set to 5% of the total number of samples. Then, points with higher local density and farther distances from each other are selected as the initial clustering centers. Practice shows that this method of selecting the initial center can significantly improve the stability and accuracy of clustering, and the average clustering accuracy is increased by about 15%.

[0055] An adaptive distance metric function is designed. Considering the characteristics of heating condition data, different features have different degrees of influence on the clustering results. The weighted Euclidean distance function is adopted, and different weights are assigned to the load characteristics, temperature parameters, and price factors. The weight coefficients are determined through sensitivity analysis. Generally, the weight of the load characteristics is set to 0.5, the weight of the temperature parameters is set to 0.3, and the weight of the price factors is set to 0.2. In practical applications, these weights can be fine-tuned according to specific situations. The adaptive distance metric significantly improves the sensitivity of clustering to key features, making the clustering results more in line with the actual operating characteristics of the heating system.

[0056] Thirdly, a dynamic clustering number determination mechanism is introduced. The traditional K-means requires the pre-specification of the clustering number K, while in practical applications, the optimal clustering number is often difficult to determine in advance. The present invention uses the Silhouette Coefficient to evaluate the clustering effects of different clustering numbers. By iteratively testing different values of K (usually between 3 and 10), the value of K with the largest Silhouette Coefficient is selected as the final clustering number. The calculation of the Silhouette Coefficient takes into account the within-cluster similarity and the between-cluster difference, and can better evaluate the effectiveness of clustering. In addition, an index of the proportion of within-cluster variance is introduced, requiring that the sum of the within-cluster variances does not exceed 30% of the total variance to ensure the compactness of the clustering results.

[0057] Finally, considering the seasonality and periodicity of heating condition characteristics, a time correlation constraint is introduced. During the clustering process, the continuity of the operating conditions in adjacent time periods is considered to avoid frequent switching of clustering categories in a short time. In specific implementation, by adding a time smoothing term to the objective function, the category changes in a short time are penalized, thereby obtaining a clustering result that is more in line with the actual operating rules. The time smoothing parameter is usually set between 0.1 and 0.3 and adjusted according to the dynamic characteristics of the system.

[0058] Through the above improvements, the K-means clustering algorithm shows better stability and accuracy in the classification of heating operating conditions. The average clustering accuracy has increased by more than 20%, providing a more reliable classification of typical operating conditions for subsequent optimization.

[0059] In step S203, feature extraction is performed on the identified typical operating conditions to establish a scene feature vector. The feature vector contains key information characterizing the operating state of the heating system, mainly including: load characteristics (such as load magnitude, change rate), environmental characteristics (such as temperature, humidity), economic characteristics (such as energy price combination), equipment characteristics (such as equipment efficiency, operating constraints), etc. These features are stored in standardized numerical form for subsequent optimization calculations.

[0060] In step S204, a seed scene library is constructed based on the feature vector, and a scene evaluation index system is established. The seed scene library is a structured data set, and each scene contains complete feature vector information. At the same time, an evaluation index system is established, including a representativeness index of the scene (reflecting the ability of the scene to represent the actual operating state), a completeness index (reflecting the coverage of the scene library for the system operating space), an independence index (reflecting the information redundancy degree between scenes), etc. Through these indexes, the quality of the seed scene library can be evaluated, and the optimization and update of the scene library can be guided.

[0061] A seed scene library is constructed based on the scene feature vector, and a scene evaluation system is established. The scene evaluation system includes a representativeness index, a completeness index, and an independence index. The specific calculation methods of these indexes are as follows: The representativeness index R is used to evaluate the ability of the scene to represent the actual operating state. For each scene S i , its representativeness index calculation formula is R(S i ) = Σ(w j ·c ij ) / Σw j . Where c ij represents the coverage of scene S i for the jth type of operating condition, and w j is the importance weight of this type of operating condition. The coverage is calculated by the cosine similarity between the scene feature vector and the center of the operating condition class: c ij =cos(θ ij ) = (V i ·C j ) / (||V i ||·||C j ||), where V i is the feature vector of scene S i , and C jis the central vector of the j-th type of operating condition. The value range of the representative index is 0-1, and the closer it is to 1, the stronger the representativeness. In practical applications, it is usually required that the representative index of the scenario is not less than 0.75 to ensure that the selected scenario can effectively represent the main operating conditions. The importance weight of the operating condition is determined according to its occurrence frequency and influence degree. For example, although the extremely cold operating condition has a low occurrence frequency, it has a significant impact, so its weight is relatively high.

[0062] The completeness index C is used to evaluate the coverage of the scenario library for the system operation space. Theoretically, the completeness calculation formula is C = Vc / Vt, where Vc is the volume of the feature space covered by the scenario library, and Vt is the total volume of the feature space. Since the dimension of the feature space is high and it is difficult to directly calculate the volume, a sampling method is used for approximate calculation: randomly generate N sample points (usually N = 1000) in the feature space. For each sample point, calculate its distance d to the nearest scenario. If d is less than the preset threshold ε (usually set to 10% of the diameter of the feature space), it is considered that this point is covered by the scenario library. Finally, the completeness index C = Nc / N, where Nc is the number of covered sample points. In practical applications, it is usually required that the completeness index is not less than 0.85 to ensure that the scenario library can cover most of the system operation space. For particularly important operating condition areas (such as extreme operating conditions or high-frequency operating conditions), the sampling density can be increased to ensure that these areas are fully covered.

[0063] The independence index I is used to evaluate the information redundancy degree among scenarios. Its calculation formula is I = 1 - Σ(Sim(S i ,S j )) / (N(N - 1) / 2), where Sim(S i ,S j ) is the similarity between scenario S i and S j , and N is the total number of scenarios. The similarity is calculated by the cosine similarity of the feature vectors: Sim(S i ,S j ) = (V i ·V j ) / (||V i ||·||V j ||). The independence index also has a value range of 0-1. The closer the value is to 1, the lower the redundancy degree among scenarios and the more independent the information. In practical applications, it is usually required that the independence index is not less than 0.7 to ensure that the selected scenarios have sufficient differences and can express different operating characteristics. If it is found that the independence index is too low, methods such as clustering or principal component analysis can be used to screen the scenarios to remove redundant scenarios with high similarity.

[0064] To comprehensively evaluate the quality of the scenario library, the above three indicators are weighted and combined to form an overall evaluation indicator \(Q = w_1\cdot R+w_2\cdot C + w_3\cdot I\), where \(w_1\), \(w_2\), and \(w_3\) are the weights of representativeness, completeness, and independence, usually set as 0.4, 0.4, and 0.2 respectively. The construction of the scenario library is an iterative optimization process. By adjusting scenario selection and parameter settings, the evaluation indicators are continuously improved until the preset quality standards are met.

[0065] In actual engineering, the scenario library can also be manually adjusted and supplemented in combination with expert experience, especially for some important but low-frequency extreme working conditions, to ensure the integrity and reliability of the scenario library. The finally constructed seed scenario library usually contains 10 - 30 scenarios, which not only ensures sufficient expressive power but also controls the computational complexity of subsequent optimization.

[0066] Through the above steps, a concise but highly representative seed scenario library is established, laying a foundation for subsequent multi-objective optimization. The selection of seed scenarios directly affects the quality of the optimization results. Therefore, special attention is paid to the representativeness and completeness of scenarios during the construction process to ensure that the optimization results can better meet the actual operation requirements.

[0067] In this embodiment, step S102 specifically includes: Step S301: Analyze the scenario features in the seed scenario library containing scenario feature vectors, and determine the coding dimension and variable range.

[0068] Step S302: Based on the coding dimension and variable range, design a coding structure suitable for the characteristics of the multi-energy complementary system, including the operating parameters of each energy subsystem.

[0069] Step S303: Based on the coding structure, construct an economic objective function, an environmental protection objective function, and a system stability objective function.

[0070] Step S304: Perform standardization processing and weight assignment on the economic objective function, the environmental protection objective function, and the system stability objective function to form a multi-objective optimization coding scheme including the economic objective function, the environmental protection objective function, and the system stability objective function.

[0071] In step S301, first analyze the scenario features in the seed scenario library to determine the coding dimensions and variable ranges. In a multi-energy complementary heating system, the variable dimensions to be coded usually include the operating parameters of each energy subsystem, the configuration parameters of the system, and the control strategy parameters. Specifically, the variable dimensions include: the heating source configuration dimension (such as heat pump capacity, boiler capacity, heat storage device capacity, etc.), the operation control dimension (such as the start-stop strategy of each heat source, the load distribution ratio, the heat storage and release strategy, etc.), and the system parameter dimension (such as the supply and return water temperatures, flow rates, pipeline pressures, etc.). When defining the variable range for each dimension, it is necessary to consider the physical constraints of the equipment, the actual engineering limitations, and the economic rationality. For example, the capacity range of the heat pump needs to consider the specifications that the manufacturer can provide, and the supply and return water temperatures need to consider the user comfort requirements and the pipeline pressure-bearing capacity.

[0072] In step S302, design a coding structure suitable for the multi-energy complementary system. Considering the characteristics of the multi-energy complementary system, a hybrid coding structure is adopted: integer coding is used for discrete variables (such as equipment switch states, operation mode selections, etc.), and real number coding is used for continuous variables (such as load distribution ratios, temperature set values, etc.). When designing the coding structure, special attention is paid to the correlation between variables. Variables that are closely related are coded adjacent to each other to improve the search efficiency of the optimization algorithm. For example, the start-stop state of a certain heat pump unit and its load distribution ratio should be coded adjacent to each other because these two variables have a direct logical relationship. The coding structure also needs to consider the implicit expression of constraint conditions. For example, the heat supply balance constraint is naturally satisfied through the coding method to avoid generating invalid solutions.

[0073] In step S303, construct an economic objective function, an environmental protection objective function, and a system stability objective function.

[0074] Based on the above coding structure, construct the economic objective function, including: Based on the above coding structure, calculate the initial investment cost, operation cost, and maintenance cost of the system to obtain cost data; Based on the cost data, perform annualized processing to obtain a cost index with a unified dimension; Based on the cost index with a unified dimension, construct a calculation model including time-of-use electricity price and seasonal natural gas price to obtain the economic objective function.

[0075] The economic objective function is a comprehensive cost function that needs to consider the life-cycle cost of the system. Specifically, it includes: initial investment cost (including the purchase cost, installation cost, and auxiliary facility cost of various heating equipment), operation cost (including energy consumption cost, maintenance cost, labor cost, etc.), depreciation cost, and scrap disposal cost. These cost items need to be annualized to achieve a unified dimension. The objective function is constructed by weighted summation, and the weight coefficients are determined through expert evaluation and sensitivity analysis. It should be noted that the dynamic change characteristics of different energy prices also need to be considered in the economic objective function. For example, factors such as time-of-use electricity price and seasonal natural gas price can be introduced to make the optimization results more in line with the actual operation situation.

[0076] Based on the above coding structure, construct the environmental protection objective function, including: Based on the above coding structure, calculate the carbon emissions and pollutant emissions of different energy forms to obtain environmental load data; Based on the above environmental load data, perform equivalent carbon emissions conversion to obtain environmental impact indicators; Based on the above environmental impact indicators, construct a calculation model including carbon emissions and pollutant emissions to obtain the environmental protection objective function.

[0077] The environmental protection objective function mainly considers the impact of operation on the environment, including direct and indirect environmental loads. It mainly includes three aspects: First, the carbon emission index, which needs to consider the life-cycle carbon emissions of different energy forms, including direct emissions (such as emissions during the operation of gas boilers) and indirect emissions (such as emissions caused by using electricity); second, the conventional pollutant emission index, including nitrogen oxides, sulfur oxides, soot, etc. These indexes need to be dynamically calculated according to the emission characteristics and operation conditions of different equipment; third, the environmental impact index, including noise pollution, heat pollution, etc. These indexes need to be normalized through weight coefficients to form a unified environmental protection evaluation index. For example, carbon emission equivalents can be used as the conversion benchmark to convert different types of environmental impacts into equivalent carbon emissions.

[0078] Based on the above coding structure, construct the system stability objective function, including: Based on the above coding structure, calculate the heating temperature fluctuation, heating continuity, and equipment start-stop frequency to obtain stability data; Based on the above stability data, perform risk assessment to obtain system reliability indicators; Based on the above system reliability indicators, construct a calculation model including heating reliability and equipment life to obtain the system stability objective function.

[0079] Stability is a key indicator to measure the reliable operation ability of a heating system, including the following dimensions: heating reliability (used to evaluate the ability of the system to meet the heating demand of users, including heating temperature fluctuations, heating continuity, etc.), equipment life (considering the impact of equipment start-stop frequency and load change rate on the service life of equipment), and system safety (evaluating the safety margin of system operation, including the fluctuation range of key parameters such as pressure and temperature). When constructing the stability objective function, a risk assessment-based method is adopted to quantify various risk factors into computable indicators. For example, the impact on the service life of equipment can be evaluated through parameters such as the start-stop times of equipment and the change speed of load rate, and the heating stability can be evaluated through the standard deviation of heating parameters. These indicators also need to be normalized through weight coefficients.

[0080] Step S304 integrates each objective function to form a unified multi-objective optimization coding scheme. The core of this step is to organically integrate the economic, environmental, and stability objective functions constructed previously. The integration process adopts a multi-level structure: the first level is the standardization of the objective function, which converts the objective functions with different dimensions into dimensionless standard scores; the second level is the weight allocation of the objective function, and the weight coefficients of each objective function are determined through the Analytic Hierarchy Process (AHP); the third level is the processing of constraint conditions, encoding various hard constraints (such as equipment operation boundaries) and soft constraints (such as priority requirements) into the optimization scheme. The finally formed multi-objective optimization coding scheme is a structured mathematical description, which not only contains the calculation methods of each objective function, but also contains various constraint conditions and priority rules that need to be considered in the optimization process.

[0081] Please refer to Figure 2 , Figure 2 which is the flowchart of the execution of the iterative optimization process provided by the embodiment of the present invention. As Figure 2 shown, step S103 specifically includes: Step S401: Set the population size, the upper limit of the number of iterations, the crossover probability, and the mutation probability to obtain the optimization parameters.

[0082] Step S402: Randomly generate an initial population, and calculate the initial values of each objective function based on the multi-objective optimization coding scheme.

[0083] Step S403: Based on the initial values of each objective function and the optimization parameters, use the improved NSGA-III algorithm for iterative calculation, where the improvement includes introducing a local search strategy based on seed scenarios, designing adaptive crossover and mutation operators, and introducing an elite retention strategy, to obtain a non-dominated solution set.

[0084] Step S404: Update the non-dominated solution set to generate the Pareto optimal solution set.

[0085] Step S405: Based on the actual requirements of the system, screen out the optimization plan from the Pareto optimal solution set.

[0086] Step S401 initializes the optimization parameters and sets the iteration conditions. In this step, key parameters need to be set for the multi-objective optimization algorithm first: population size (usually set to 2-3 times the coding dimension to ensure full coverage of the search space), upper limit of the number of iterations (estimated based on the problem scale and convergence speed, usually set to 500-1000 times), crossover probability (set between 0.7-0.9 to ensure population diversity), and mutation probability (set between 0.1-0.2 to avoid falling into local optima). The setting of the convergence conditions includes two aspects: one is that the change range of the optimal solution for several consecutive generations is less than the preset threshold; the other is that the advancing speed of the Pareto front is lower than a certain threshold. The initial values of these parameters can be adaptively adjusted based on the characteristics of the seed scenario to ensure that the algorithm can maintain good performance in different scenarios.

[0087] Step S402 calculates the initial values of each objective function based on the coding scheme. In this step, an initial population needs to be randomly generated first, and each individual is a complete operation plan for the heating system. For each individual, calculate its economic objective function value (including investment cost and operation cost), environmental protection objective function value (including carbon emissions and pollutant emissions), and stability objective function value (including reliability index and life index) respectively. The calculation process needs to consider specific conditions in the seed scenario, such as load characteristics, meteorological conditions, etc. These initial values will be used as the benchmark points for subsequent iterative optimization and also for evaluating the convergence performance of the optimization algorithm.

[0088] Step S403 performs iterative calculations using an improved multi-objective optimization algorithm. In this step, the improved NSGA-III algorithm (Non-dominated Sorting Genetic Algorithm III) is used, which is particularly suitable for handling optimization problems with three or more objectives. The improvement is mainly reflected in three aspects: one is to introduce a local search strategy based on the seed scenario, and preferentially explore the solution space similar to the seed scenario in each iteration; the second is to design adaptive crossover and mutation operators, and dynamically adjust the parameters according to the population diversity; the third is to introduce an elite retention strategy to ensure that high-quality solutions will not be lost during the evolution process. In each iteration, the algorithm updates the non-dominated solution set and maintains the population diversity through the reference point method.

[0089] Perform iterative calculations using the improved NSGA-III algorithm. NSGA-III (Non-dominated Sorting Genetic Algorithm III) is an evolutionary algorithm suitable for multi-objective optimization problems. The present invention makes targeted improvements to it to improve the efficiency in the optimization of the multi-energy complementary heating system. The specific improvement contents are as follows: First, a local search strategy based on seed scenarios is introduced. Traditional NSGA-III does not utilize problem-specific prior knowledge during the search process, resulting in low search efficiency. The present invention constructs a seed scenario influence function Fi(x) to evaluate the similarity between the solution x and the seed scenario Si. The design of the influence function takes into account the Euclidean distance in the feature space and the correlation in the decision space, and is formed through weighted combination. For each seed scenario, its influence radius ri is defined, and the solutions within the radius will be guided by this scenario. During the algorithm iteration process, local search is performed with a certain probability (the initial probability is set to 0.3 and linearly decreases to 0.1 as the number of iterations increases). Specifically, solutions with high similarity to a certain seed scenario are selected from the current non-dominated solution set, and local mutation is performed along the direction guided by the scenario features. The step size of the local search is set to an adaptive value, initially large (10% of the variable range), and gradually decreases as the number of iterations increases (minimum to 1% of the variable range). Experiments show that after introducing the local search based on seed scenarios, the algorithm improves the speed of finding high-quality solutions by about 30%, especially in the case of complex and high-dimensional solution spaces, the effect is more obvious.

[0090] Secondly, adaptive crossover and mutation operators are designed. Traditional NSGA-III uses fixed crossover and mutation probabilities, which are difficult to adapt to the optimization requirements at different stages. The present invention introduces an adaptive mechanism based on population diversity, and calculates the population diversity index S d to dynamically adjust the operator parameters. S d is calculated based on the average Euclidean distance between individuals in the solution space and is normalized. When S d is lower than the threshold T1 (usually set to 0.3), it indicates that the population diversity is insufficient. At this time, the mutation probability is increased (up to 0.3 at most) to promote exploration; when S d is higher than the threshold T2 (usually set to 0.7), it indicates that the population dispersion is relatively high. At this time, the crossover probability is increased (up to 0.95 at most) to strengthen exploitation. The dynamic adjustment formulas for the crossover probability p c and the mutation probability p m are as follows: p c = + ( - ) * max(0, (S d - T1) / (T2 - T1)) p m = + ( - ) * max(0, (T2 - S d) / (T2 - T1)) where = 0.7, = 0.95, = 0.1, = 0.3。

[0091] In addition, considering the characteristics of the multi - energy complementary system, dedicated mutation operators are designed for key decision variables (such as heat pump capacity, load distribution ratio, etc.). These mutation operators take into account the constraint relationships and physical meanings between variables to ensure that the solutions after mutation are still within a reasonable range. For example, for the load distribution ratio variable, it is ensured that the sum of all ratios is 1 during mutation; for the equipment capacity variable, the standard specifications of the equipment are considered during mutation. The dedicated mutation operators significantly increase the generation ratio of effective solutions and reduce the computational overhead of invalid solutions.

[0092] And elitist retention and diversity maintenance strategies are introduced. To prevent high - quality solutions from being lost during the evolution process, the top m solutions (m is set to 10% of the population size) in the current non - dominated solution set are retained in each iteration and directly copied to the next generation. The selection of elite individuals is based on priority sorting, first considering the non - dominated level and then the crowding degree. To avoid premature convergence caused by elitist retention, slight mutation with a low probability (0.05) is also performed on elite individuals, and the mutation amplitude is controlled within a small range (within 2% of the variable range).

[0093] In addition, to maintain the diversity of the solution set, a reference - point adaptive adjustment technique is adopted. In the standard NSGA - III, the reference points are evenly distributed in the objective space, but this may not be suitable for the shape of the Pareto front of the actual problem. According to the distribution of the current non - dominated solutions, the density of the reference points is dynamically adjusted in this invention, increasing the reference points in the sparse area of solutions and reducing the reference points in the dense area of solutions. The reference points are updated every 10 generations, and the update ratio does not exceed 20% to maintain a certain stability.

[0094] Through the above improvements, the NSGA - III algorithm performs excellently in the multi - objective optimization problem of the multi - energy complementary heating system. The convergence speed is increased by about 40%, and the diversity and uniformity of the solution set are also significantly improved, providing richer and higher - quality optimization scheme options for decision - makers.

[0095] Step S404 generates the Pareto optimal solution set. As the iteration progresses, the algorithm continuously updates and maintains a non - dominated solution set. Each solution in this solution set represents different trade - off schemes for the three objectives of economy, environmental protection, and stability. To ensure the quality of the solution set, solution screening and evaluation are required: First, solutions with too high similarity are removed to maintain the diversity of the solution set; then, the crowding degree index of each solution is calculated to ensure that the solutions are evenly distributed in the objective space; finally, the practicality of the solutions is evaluated, and solutions that are mathematically feasible but difficult to implement in engineering practice are excluded.

[0096] Step S405 selects the final optimized solution from the Pareto solution set according to the actual system requirements. This step needs to consider the decision-maker's preferences and actual engineering constraints. The screening process adopts a hierarchical screening strategy: the first layer is screening based on hard constraints, eliminating solutions that do not meet the key engineering requirements; the second layer is screening based on soft constraints, considering factors such as the difficulty of implementation and the convenience of maintenance of the solutions; the third layer is screening based on decision-making preferences, selecting the final solution according to the decision-maker's emphasis on different objectives. For example, if the decision-maker attaches more importance to economy, a solution with the optimal economic indicators can be selected on the premise of meeting the basic environmental protection and stability requirements.

[0097] Selecting the optimized solution from the Pareto optimal solution set based on the actual system requirements includes: Calculating the similarity index and crowding degree index of the solutions in the Pareto optimal solution set to obtain solution set evaluation data; Based on the solution set evaluation data, adopting a hierarchical screening strategy, eliminating solutions with a similarity greater than a preset threshold and solutions that do not meet the engineering implementation conditions to obtain candidate solutions; Based on the candidate solutions, selecting the optimized solution according to the decision-maker's preferences for economy, environmental protection, and stability.

[0098] In this embodiment, step S104 specifically includes: Step S501: Based on the boundary condition verification model, set static indicators and dynamic indicators to obtain verification parameters.

[0099] Step S502: Based on the verification parameters, perform steady-state simulation and dynamic simulation using a hierarchical simulation strategy to obtain simulation data.

[0100] Step S503: Based on the simulation data, perform statistical analysis and sensitivity analysis to obtain boundary conditions that do not meet the engineering constraints.

[0101] Step S504: Based on the boundary conditions that do not meet the engineering constraints, adjust the constraint range and safety margin to obtain corrected parameters.

[0102] Step S505: Based on the corrected parameters, perform iterative optimization and balance the system reliability and economy to obtain a corrected solution.

[0103] Step S506: Based on the corrected solution, perform dynamic correction of the boundary conditions.

[0104] Step S501 constructs a boundary condition verification model and sets verification metrics. The construction of the boundary condition verification model needs to consider constraints at multiple levels: physical constraints (such as the laws of thermodynamics, energy conservation, etc.), equipment constraints (such as the operating range of equipment, start-stop characteristics, etc.), and system constraints (such as the hydraulic balance and thermal balance of the heating network, etc.). The verification metric system includes two categories: static metrics and dynamic metrics. Static metrics mainly verify the performance of the system during steady-state operation, such as equipment load rates, system energy efficiency, etc. Dynamic metrics focus on the response characteristics of the system when the operating conditions change, such as the temperature adjustment rate, pressure fluctuation amplitude, etc. Clear calculation methods and judgment criteria need to be established for these metrics to form a quantifiable verification system. For example, specific metrics such as setting the reasonable range of the equipment load rate to be 30% - 90% and the system response time not exceeding 30 minutes can be set.

[0105] Set static and dynamic metrics based on the boundary condition verification model. The boundary condition verification model is a comprehensive system simulation model, including a thermodynamic model, an equipment characteristic model, and a control strategy model, used to verify the feasibility and effectiveness of the optimization scheme in actual projects. The main mathematical expressions and specific metrics of this model are as follows: The boundary condition verification model is established based on the basic conservation laws of the thermal system. The mass conservation equation describes the flow of the working fluid in the system: , where m is the mass, ρ is the density, and V is the velocity. The energy conservation equation describes the transfer and conversion of energy in the system: , where cp is the specific heat capacity, T is the temperature, k is the thermal conductivity, and Q is the heat source term. The momentum conservation equation describes the pressure distribution in the system: , where p is the pressure, μ is the viscosity, and F is the external force.

[0106] Under the framework of the above basic equations, the model includes a variety of static and dynamic metrics for comprehensively evaluating system performance: Static metrics mainly evaluate the performance of the system during steady-state operation: (1) Equipment load rate LR: Defined as the ratio of the actual load of the equipment to the rated capacity, LR = Qactual / Qrated. For major equipment (such as heat pumps, boilers, etc.), it is required that LR is between 30% - 90% to avoid long-term low-load or over-load operation of the equipment. Too low a load rate will lead to a decrease in equipment efficiency, and too high a load rate may affect the equipment life and safety.

[0107] (2) System energy efficiency COP: Defined as the ratio of the output heat to the input energy, COP = Qout / Ein. The overall system energy efficiency is required to be not less than 85% of the reference value, and the reference value is determined according to the system type and regional standards. For heat pump systems, the COP is usually required to be not less than 3.0 under the design conditions; for boiler systems, the thermal efficiency is required to be not less than 92%.

[0108] (3) Economic Index EI: Defined as the ratio of the annualized cost to the annual heat supply, EI = Cy / Qy, with the unit of yuan / GJ. The annualized cost includes equipment depreciation, energy consumption, operation and maintenance costs, etc. This index is required not to exceed 110% of the benchmark value, and the benchmark value is determined according to local energy prices and heating standards.

[0109] (4) Primary Energy Utilization Rate PER: Defined as the ratio of the useful energy output by the system to the primary energy consumed, PER = Quseful / (Ep / ηp), where Ep is the primary energy consumption and ηp is the primary energy conversion efficiency. This index is required to be not less than 0.6, reflecting the efficient utilization of energy by the system.

[0110] Dynamic indicators mainly evaluate the response characteristics of the system under changing operating conditions: (1) Temperature Regulation Rate TR: Defined as the rate of change of temperature per unit time, TR =ΔT / Δt. To ensure user comfort and equipment safety, it is required that the change in the supply water temperature does not exceed 5°C / hour and the change in the indoor temperature does not exceed 2°C / hour.

[0111] (2) Pressure Fluctuation Range PF: Defined as the ratio of the pressure change to the design pressure, PF =ΔP / Pdesign. To ensure the safety of the pipe network, it is required that the pressure fluctuation does not exceed ±10% of the design pressure.

[0112] (3) Equipment Start / Stop Frequency SF: Defined as the number of start / stop times of the equipment per unit time. For major equipment (such as heat pumps, boilers, etc.), it is required that the number of start / stop times per day does not exceed 6 times to extend the equipment life.

[0113] (4) System Response Time RT: Defined as the time required for the system to reach a new stable state from the start of the load change. It is required that under a 30% step change in the load, the system response time does not exceed 30 minutes.

[0114] (5) Heating Reliability Index RI: Defined as the proportion of time that the system can meet the design heating requirements, RI =Tsatisfied / Ttotal. This index is required to be not less than 0.95, that is, the system can meet the heating demand for more than 95% of the time.

[0115] In addition to the above quantitative indicators, the verification model also considers some qualitative indicators, such as equipment compatibility, the feasibility of control strategies, and the convenience of maintenance, etc. These indicators are evaluated by expert scoring, and the average score is required to be not less than 7 points (out of 10).

[0116] The verification model adopts a hierarchical simulation strategy. First, a steady-state simulation is carried out to verify the performance of the system under the design conditions. Then, a typical-day dynamic simulation is carried out to verify the performance of the system under daily load variations. Finally, an extreme-condition simulation is carried out to verify the adaptability of the system under extreme conditions. During the simulation process, a variable-step integration algorithm is used, with the step size ranging from 10 seconds to 10 minutes, which is automatically adjusted according to the rate of change of the system state to ensure the balance between calculation accuracy and efficiency.

[0117] The simulation results are presented through statistical analysis and visualization, including time series graphs, load distribution maps, energy efficiency change graphs, etc., which intuitively reflect the operating characteristics and performance indicators of the system under different conditions. For the indicators that do not meet the requirements, the system will automatically mark them and provide optimization suggestions, which serve as the basis for dynamically modifying the boundary conditions.

[0118] Step S502 conducts a simulation verification of the optimization plan. This step adopts a hierarchical simulation strategy: First, a steady-state simulation is carried out to verify the performance indicators of the system under various typical conditions. Then, a dynamic simulation is carried out, focusing on the transition characteristics of the system during condition switching. The following points should be particularly noted during the simulation process: First, the start-stop process and switching losses of the equipment should be considered; second, the system response under extreme conditions should be simulated; third, the feasibility of the control strategy should be verified. The simulation uses professional thermal system simulation software and sets the simulation parameters in combination with actual engineering experience. For example, when verifying the dynamic characteristics of a heating system, extreme situations such as sudden changes in outdoor temperature and sudden changes in user demand need to be simulated to observe the adaptability of the system.

[0119] Step S503 analyzes the verification results and identifies the boundary conditions that do not meet the engineering constraints. This step first conducts a systematic analysis of the simulation results to identify the boundary conditions that may have problems. The analysis methods include: statistical analysis (calculating the statistical characteristics of each indicator, such as mean, standard deviation, etc.), sensitivity analysis (studying the degree of influence of parameter changes on system performance), and limit analysis (exploring the boundary conditions of system performance). Through these analyses, potential problems can be discovered, such as excessive equipment load, slow system response, and unqualified energy efficiency under certain conditions. Particular attention should be paid to those boundary conditions that are easily overlooked in actual engineering, such as the partial load characteristics of equipment and the safety margin of the system.

[0120] Step S504 dynamically modifies the boundary conditions according to the verification results. This step adopts an iterative optimization method to correct the problems found one by one. The correction strategies include: adjusting the constraint range (such as relaxing or tightening the value range of certain parameters), increasing the safety margin (such as considering more conservative design boundaries), and optimizing the control strategy (such as improving the start-stop logic of the equipment). The correction process needs to balance multiple objectives: ensuring the safety and reliability of the system while maintaining the economy of the optimization plan. For example, if it is found that the equipment load rate is too high under certain working conditions, it can be solved by increasing the standby capacity or optimizing the load distribution strategy, but at the same time, the impact of these modifications on the overall system performance needs to be evaluated.

[0121] Steps S505 and S506 output the final optimized boundary condition scheme. The final scheme needs to include a complete technical document, which details: the optimized system configuration parameters (such as the capacity and operating range of each device), control strategy parameters (such as start-stop thresholds, adjustment curves, etc.), and operating boundary conditions (such as the allowable working condition range, safety limits, etc.). At the same time, it is necessary to provide a feasibility demonstration of the scheme, including technical feasibility, economic feasibility, and engineering implementation suggestions. Particularly important is to clearly explain the applicable conditions and limitations of the scheme to provide clear guidance for subsequent engineering implementation. For example, it is necessary to explain the adaptability of the scheme under different climate conditions and different load characteristics, as well as the possible debugging and optimization suggestions.

[0122] Please refer to Figure 3 , Figure 3 which is the structural block diagram of the urban heating system boundary condition optimization device provided by the embodiment of the present invention. As Figure 3 shown, the device includes: A construction module 601, configured to construct a seed scenario library containing scenario feature vectors based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system; An analysis module 602, configured to analyze and determine the coding dimension and variable range based on the seed scenario library, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function; A calculation module 603, configured to calculate the initial values of each objective function based on the multi-objective optimization coding scheme, and perform iterative calculations using a multi-objective optimization algorithm to obtain an optimization scheme; A correction module 604, configured to construct a boundary condition verification model based on the optimization scheme for simulation verification, and dynamically correct the boundary conditions according to the verification results; the boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; the operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperatures, flow rates, and pipe network pressures.

[0123] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural transformation made under the inventive concept of the present invention by using the content of the specification and drawings of the present invention, or any direct / indirect application in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. A method for optimizing the boundary conditions of a multi - energy complementary urban heating system, characterized in that The steps include: Based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, construct a seed scenario library containing scenario feature vectors; Based on the seed scenario library, analyze and determine the coding dimension and variable range, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function; Based on the multi-objective optimization coding scheme, calculate the initial values of each objective function, and use the multi-objective optimization algorithm for iterative calculation to obtain an optimization scheme; Based on the optimization scheme, construct a boundary condition verification model for simulation verification, and dynamically correct the boundary conditions according to the verification results; The boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; The operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperatures, flow rates, and pipe network pressures.

2. The method according to claim 1, wherein Based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system, construct a seed scenario library containing scenario feature vectors, including: According to the heat load characteristics, meteorological conditions, and energy prices of the urban heating system, obtain key parameters; Based on the key parameters, use an improved K-means clustering algorithm to classify the heating conditions and identify typical heating conditions; Extract the features of the typical heating conditions, and establish a scenario feature vector containing load features, environmental features, and economic features; Based on the scenario feature vector, construct a seed scenario library, and establish a scenario evaluation system containing representative indicators, completeness indicators, and independence indicators to obtain the seed scenario library containing scenario feature vectors; Among them, the representative indicator is used to evaluate the representation ability of the scenario for the actual operation state, the completeness indicator is used to evaluate the coverage of the scenario library for the system operation space, and the independence indicator is used to evaluate the information redundancy degree between scenarios.

3. The method according to claim 1, wherein Based on the seed scenario library, analyze and determine the coding dimension and variable range, and construct a multi-objective optimization coding scheme including an economic objective function, an environmental protection objective function, and a system stability objective function, including: Analyze the scenario features in the seed scenario library containing scenario feature vectors to determine the coding dimension and variable range; Based on the coding dimension and variable range, design a coding structure suitable for the characteristics of the multi-energy complementary system, including the operation parameters of each energy subsystem; Based on the coding structure, construct an economic objective function, an environmental protection objective function, and a system stability objective function; Perform standardization processing and weight assignment on the economic objective function, environmental protection objective function, and system stability objective function to form the multi-objective optimization coding scheme including the economic objective function, environmental protection objective function, and system stability objective function.

4. The method according to claim 3, characterized in that, Based on the coding structure, construct the economic objective function, including: Based on the coding structure, calculate the system's initial investment cost, operation cost, and maintenance cost to obtain cost data; Based on the cost data, perform annualization processing to obtain a cost indicator with a unified dimension; Based on the cost index with unified dimension, a calculation model including time-of-use electricity price and seasonal natural gas price is constructed to obtain the economic objective function.

5. The method according to claim 3, wherein Based on the coding structure, the environmental protection objective function is constructed, including: Based on the coding structure, the carbon emissions and pollutant emissions of different energy forms are calculated to obtain environmental load data; Based on the environmental load data, equivalent carbon emissions conversion is performed to obtain environmental impact indicators; Based on the environmental impact indicators, a calculation model including carbon emissions and pollutant emissions is constructed to obtain the environmental protection objective function.

6. The method according to claim 3, wherein Based on the coding structure, the system stability objective function is constructed, including: Based on the coding structure, the heating temperature fluctuation, heating continuity, and equipment start-stop frequency are calculated to obtain stability data; Based on the stability data, risk assessment is performed to obtain system reliability indicators; Based on the system reliability indicators, a calculation model including heating reliability and equipment life is constructed to obtain the system stability objective function.

7. The method according to claim 1, wherein Based on the multi-objective optimization coding scheme, the initial values of each objective function are calculated, and a multi-objective optimization algorithm is used for iterative calculation to obtain an optimization scheme, including: Set the population size, upper limit of iteration times, crossover probability, and mutation probability to obtain optimization parameters; Randomly generate an initial population, and calculate the initial values of each objective function based on the multi-objective optimization coding scheme; Based on the initial values of each objective function and the optimization parameters, an improved NSGA-III algorithm is used for iterative calculation, where the improvement includes introducing a local search strategy based on seed scenarios, designing adaptive crossover and mutation operators, and introducing an elite retention strategy, to obtain a non-dominated solution set; Update the non-dominated solution set to generate a Pareto optimal solution set; Based on the actual requirements of the system, the optimization scheme is screened from the Pareto optimal solution set.

8. The method according to claim 7, wherein Based on the actual requirements of the system, the optimization scheme is screened from the Pareto optimal solution set, including: Calculate the similarity index and crowding degree index of the solutions in the Pareto optimal solution set to obtain solution set evaluation data; Based on the solution set evaluation data, a hierarchical screening strategy is used to eliminate solutions with similarity greater than a preset threshold and solutions that do not meet the engineering implementation conditions to obtain candidate solutions; Based on the candidate solutions, according to the decision-maker's preferences for economy, environmental protection, and stability, the optimization scheme is selected.

9. The method according to claim 1, characterized in that, Based on the optimization scheme, a boundary condition verification model is constructed for simulation verification, and the boundary conditions are dynamically corrected according to the verification results, including: Based on the boundary condition verification model, static indicators and dynamic indicators are set to obtain verification parameters; Based on the verification parameters, a hierarchical simulation strategy is used to perform steady-state simulation and dynamic simulation to obtain simulation data; Based on the simulation data, statistical analysis and sensitivity analysis are performed to obtain boundary conditions that do not meet engineering constraints; Based on the boundary conditions that do not meet engineering constraints, the constraint range and safety margin are adjusted to obtain correction parameters; Based on the correction parameters, iterative optimization is performed and the system reliability and economy are balanced to obtain a correction scheme; Based on the correction scheme, the boundary conditions are dynamically corrected.

10. An optimization device for the boundary conditions of an urban heating system based on multi-energy complementarity, characterized in that, Including: A construction module, configured to construct a seed scenario library containing scenario feature vectors based on the obtained heat load characteristics, meteorological conditions, and energy prices of the urban heating system; An analysis module, configured to analyze and determine the coding dimension and variable range based on the seed scenario library, and construct a multi-objective optimization coding scheme containing an economic objective function, an environmental protection objective function, and a system stability objective function; A calculation module, configured to calculate the initial values of each objective function based on the multi-objective optimization coding scheme, and perform iterative calculations using a multi-objective optimization algorithm to obtain an optimization scheme; A correction module, configured to construct a boundary condition verification model based on the optimization scheme for simulation verification, and dynamically correct the boundary conditions according to the verification results; The boundary conditions include: heat source configuration parameters, operation control parameters, and system parameters, where the heat source configuration parameters include heat pump capacity, boiler capacity, and heat storage device capacity; The operation control parameters include heat source start-stop strategy, load distribution ratio, and heat storage and release strategy; the system parameters include supply and return water temperature, flow rate, and pipe network pressure.

Images