A method and related devices for integrated energy dispatching in industrial parks

By transforming the partial Bruker chance constraint into a deterministic linear constraint and combining it with a multi-objective particle swarm optimization algorithm, the uncertainty handling and multi-objective optimization problems of the park's integrated energy system were solved, achieving safe and economical operation and multi-timescale collaborative optimization, reducing wind and solar curtailment rates and improving system efficiency.

CN123073971APending Publication Date: 2026-11-10XIDIAN SMART ENERGY TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611244837.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-17
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing technologies for optimizing and scheduling integrated energy systems in industrial parks are insufficient to simultaneously achieve safe and economical operation under uncertainty, rapid real-time solutions, multi-objective scientific trade-offs, and multi-timescale collaborative optimization. They are particularly lacking in handling source-load uncertainty, the fusion of heuristic algorithms and distributed bar chance constraints, multi-objective conflicts, and timescale coordination.

Method used

By transforming the random constraint based on Wasserstein distance fuzzy sets into a deterministic linear constraint, and combining it with a multi-objective particle swarm optimization algorithm, a mixed-integer linear programming model is constructed. The model is then corrected in real time through model predictive control, achieving multi-timescale collaborative optimization within the safe and feasible domain.

Benefits of technology

It achieves rapid solution under safety constraints, outputs Pareto optimal solution set, reduces wind and solar curtailment rate, improves system operation economy and reliability, and meets the real-time requirements of intraday rolling scheduling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123073971A_ABST
    Figure CN123073971A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of integrated energy system optimization and scheduling, and discloses a method and system for integrated energy scheduling in industrial parks. By establishing the balance equations and equipment constraints of the integrated energy system in the park, and constructing a Wasserstein distance fuzzy set based on historical wind and solar data, the power balance equation is transformed into a partial Bruker chance constraint. Using duality theory, this is equivalently transformed into a deterministic linear constraint, forming a safe and feasible scheduling region. A day-ahead mixed-integer linear programming model is constructed to solve the day-ahead schedule, minimizing the total system operating cost. Based on the day-ahead schedule, intraday rolling multi-objective particle swarm optimization is performed within the feasible region to obtain the Pareto optimal scheduling solution set. Real-time corrections and output commands are performed based on model predictive control. This method alleviates the contradiction between handling source-load uncertainty and economic efficiency, integrates partial Bruker optimization and heuristic algorithms to improve solution speed and safety, supports scientific trade-offs through multi-objective optimization, and enhances the economy and reliability of system operation through multi-timescale coordination.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system optimization and scheduling technology, and particularly relates to a method and related device for integrated energy scheduling in industrial parks. Background Technology

[0002] The integrated energy system of the industrial park is characterized by multi-energy complementarity, multi-source supply, and heterogeneous energy flow coupling. Through the coordinated management and optimized scheduling of various energy types such as electricity, gas, cooling, and heating, it has become an important platform for improving the overall efficiency of energy utilization and promoting the consumption of renewable energy. A typical integrated energy system in an industrial park is equipped with various types of energy equipment, including photovoltaic, wind power, gas turbines, electric boilers, power-to-gas conversion, and energy storage. Its optimized scheduling technology aims to rationally allocate the operating output of each device while meeting the system's load demand, thereby achieving multiple optimization goals such as economic operation and low-carbon emission reduction. It is a core technical link to ensure the efficient and stable operation of the integrated energy system in the industrial park.

[0003] Currently, there are still many technical shortcomings in the optimization and scheduling technology of the park's integrated energy system: First, there is an inherent contradiction between handling the uncertainty of source and load and the economic efficiency of operation. Deterministic optimization methods cannot effectively cope with the random fluctuations in wind and solar power output. Stochastic programming methods rely on precise probability distribution parameters, which are difficult to meet the conditions for engineering applications. Although traditional robust optimization based on worst-case scenarios can ensure operational safety, its excessive conservatism significantly sacrifices the economic efficiency of system operation. Although the existing sub-Benders robust optimization method based on Wasserstein distance to construct fuzzy sets can alleviate the conservatism problem, its solution relies heavily on mathematical programming methods such as column and constraint generation and Benders decomposition, which generally take a long time to solve and cannot meet the real-time requirements of intraday rolling scheduling. Second, heuristic algorithms cannot directly adapt to Pareto opportunistic constraints. While heuristic methods such as multi-objective particle swarm optimization have the advantage of fast solution speed, they are mainly applicable to simple or unconstrained optimization scenarios. Pareto opportunistic constraints, on the other hand, are complex probabilistic constraints. Directly using heuristic algorithms to solve them will generate a large number of infeasible solutions, failing to guarantee the safe operation of scheduling schemes under wind and solar fluctuations. In existing technologies, Pareto optimization and heuristic algorithms are mostly used independently, and no effective integration scheme has yet been formed. The core obstacle is that Pareto opportunistic constraints lack a simple deterministic equivalent transformation form that can be embedded in the framework of heuristic algorithms, making it impossible for heuristic algorithms to conduct efficient searches within the decision space that satisfies safety constraints. Third, there is a lack of effective trade-off mechanisms for multi-objective conflicts. System optimization involves multiple conflicting objectives such as economy, low carbon emissions, and reliability. Existing methods mostly use weighted summation methods, with weight settings relying on human experience, making it difficult to obtain a true Pareto optimal solution set and unable to flexibly adapt to external operating conditions such as carbon price fluctuations. Fourth, the multi-timescale coordination mechanism is imperfect. Existing scheduling schemes mostly adopt a single timescale, which cannot fully utilize the operational benefits brought by the improved accuracy of ultra-short-term forecasts, and may easily lead to an increase in wind and solar curtailment rates or increased system regulation costs.

[0004] It is evident that existing integrated energy system optimization and scheduling technologies for industrial parks cannot simultaneously achieve safe and economical operation under uncertainty, rapid real-time solution, multi-objective scientific trade-offs, and multi-timescale collaborative optimization. Summary of the Invention

[0005] This invention provides a method and related device for integrated energy dispatching in industrial parks. This method can simultaneously achieve safe and economical operation under uncertainty, rapid real-time solution, multi-objective scientific trade-off, and multi-timescale collaborative optimization.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A comprehensive energy dispatching method for industrial parks includes: Based on the equipment composition and operating parameters of the park's integrated energy system, establish the electrical power balance equation, the thermal power balance equation, and the operating constraints of each piece of equipment. Based on historical wind and solar power output data, an empirical distribution and Wasserstein distance fuzzy set are constructed. The power balance equation, which includes wind and solar prediction errors, is transformed into a split-Bruker chance constraint. It is then equivalently transformed into a deterministic linear constraint through dual theory. The power balance equation, thermal power balance equation, and the operating constraints of each device are combined to form a scheduling safe and feasible region. With minimizing the total system operating cost as the optimization objective, a day-ahead scheduling mixed integer linear programming model is constructed based on the scheduling safe and feasible region. By inputting the day-ahead predicted operating data, the day-ahead start-stop status and day-ahead output plan of each device are obtained by solving the model. Using the day-ahead start-stop state as a hard constraint and the day-ahead output plan as a benchmark, an initial population is generated. Intraday rolling multi-objective particle swarm optimization is performed within the scheduling safety and feasibility domain to obtain the Pareto optimal scheduling solution set. Based on real-time collected measured operation data, the system uses the state-space model of the park's integrated energy system as the prediction model, executes model predictive control within the safe and feasible scheduling domain, and makes real-time corrections with reference to the intraday scheduling scheme selected by the Pareto optimal scheduling solution set, outputting real-time control commands for each device.

[0007] Furthermore, the integrated energy system of the park adopts an integrated source-grid-load-storage architecture, including power supply, heating and gas supply subsystems. The equipment consists of photovoltaic power generation arrays, wind turbine generators, gas turbines, electric boilers, electric-to-gas conversion equipment, battery energy storage systems, thermal storage tanks and grid interconnection lines. The operating constraints of each piece of equipment include upper and lower limits of equipment output, gas turbine ramp rate constraints, state of charge constraints and charge / discharge mutual exclusion constraints of the battery energy storage system, and power interaction constraints of the grid interconnection lines. The power balance equation is: During any scheduling period t, the sum of the output of all generating equipment in the system equals the sum of the consumption of all consuming equipment.

[0008] in, For the power generation of gas turbines; and These refer to the actual output of photovoltaic and wind power, respectively. and These refer to the power consumption for online electricity purchases and the power consumption for retail sales, respectively. and These are the battery discharge and charging power, respectively. For electrical load; This refers to the electrical power consumption of the electric boiler. The power consumption of the electro-gas conversion equipment; The actual output of photovoltaic and wind power is expressed as the sum of the day-ahead forecast and the forecast error:

[0009]

[0010] in, and The day-ahead forecasts for solar and wind power are respectively provided. and This represents the prediction error; The heat power balance equation is: the heat supply of the source equipment equals the sum of the heat load and the heat charge of the storage tank.

[0011] in, The heat-to-electric ratio of a gas turbine; and These are the heat release and heat charging power of the thermal storage tank, respectively. This refers to the heat load.

[0012] Furthermore, the process of constructing the split-bar chance constraint is as follows: An empirical distribution is constructed based on historical wind and solar power output data for a preset number of days. The Wasserstein distance radius is calculated according to a preset confidence level. A fuzzy set of probability distribution is constructed with the empirical distribution as the center and the Wasserstein distance radius as the boundary. A preset violation probability is set to rewrite the power balance equation, which includes wind and solar prediction errors, into a split-Bruker chance constraint form. Then, it is equivalently transformed into a deterministic linear constraint through dual theory.

[0013] Furthermore, the day-ahead forecast operating data includes electricity load forecast curves, heat load forecast curves, photovoltaic output forecast curves, wind power output forecast curves, time-of-use electricity prices, and natural gas prices; A mathematical programming solver is used to solve the day-ahead scheduling mixed integer linear programming model, and the day-ahead start and stop states obtained from the solution are passed as hard constraints to the intraday optimization stage.

[0014] Furthermore, the velocity and position of the particles are updated using adaptive inertial weights, and an adaptive Gaussian mutation operation is applied to the updated particle positions. The mutation probability is positively correlated with the Wasserstein distance radius, and the mutation probability increases with the greater uncertainty of wind and solar power output. All particles satisfy the constraint boundary of the scheduling safe and feasible region, and boundary absorption processing is performed on particles that exceed the boundary.

[0015] Furthermore, when performing intraday rolling multi-objective particle swarm optimization, the total operating cost and total carbon emissions are calculated for each particle, all particles are sorted using the Pareto dominance relation, the Pareto optimal solution is stored in an external archive, and the size of the external archive is controlled using an adaptive mesh method.

[0016] Furthermore, in the process of real-time correction using the intraday scheduling scheme selected by the Pareto optimal scheduling solution set as a reference, the optimization objective is to minimize the weighted sum of squares of the deviation between the output of each device and the intraday reference plan. The initial state of the system is updated in each control cycle, and the constrained optimization problem is solved within the scheduling safety feasible region. Only the control instructions of the first control cycle are executed.

[0017] A comprehensive energy dispatching system for industrial parks includes: A module is established to create electrical power balance equations, thermal power balance equations, and operating constraints for each device based on the equipment composition and operating parameters of the park's integrated energy system. The constraint transformation module is used to construct an empirical distribution and Wasserstein distance fuzzy set based on historical wind and solar power output data. It transforms the power balance equation, which includes wind and solar prediction errors, into a partial Bruker chance constraint, and then transforms it into a deterministic linear constraint through dual theory. The module combines the power balance equation, the thermal power balance equation, and the operating constraints of each device to form a scheduling safe and feasible domain. The solution module is used to construct a day-ahead scheduling mixed integer linear programming model based on the scheduling safety feasible region with the optimization objective of minimizing the total system operating cost. It takes the day-ahead predicted operating data as input and solves to obtain the day-ahead start-up and shutdown status and day-ahead output plan of each device. The optimization module is used to generate an initial population based on the day-ahead start-stop status as a hard constraint and the day-ahead output plan as a benchmark, and to perform intraday rolling multi-objective particle swarm optimization within the scheduling safety and feasible domain to obtain the Pareto optimal scheduling solution set. The correction module is used to perform model predictive control within the safe and feasible domain of scheduling, based on the real-time collected measured operating data and the state space model of the park's integrated energy system as the prediction model. It uses the intraday scheduling scheme selected by the Pareto optimal scheduling solution set as a reference for real-time correction and outputs real-time control commands for each device.

[0018] A comprehensive energy dispatching device for industrial parks, comprising: Memory, used to store computer programs; A processor is used to implement the above-mentioned integrated energy dispatching method for the park when executing the computer program.

[0019] A computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the above-mentioned integrated energy dispatching method for industrial parks.

[0020] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a method for integrated energy dispatching in a park. It establishes the balance equations and equipment constraints of the park's integrated energy system, constructs a Wasserstein distance fuzzy set based on historical wind and solar data, transforms the power balance equations into partial Bruker chance constraints, and then uses duality theory to equivalently transform them into deterministic linear constraints, thus forming a safe and feasible dispatch region. Next, a day-ahead mixed-integer linear programming model is constructed to solve the day-ahead schedule, minimizing the total system operating cost. Then, based on the day-ahead schedule, intraday rolling multi-objective particle swarm optimization is performed within the feasible region to obtain the Pareto optimal dispatch solution set. Finally, real-time correction commands are output based on model predictive control. Through the deterministic equivalent transformation of partial Bruker chance constraints, complex probabilistic constraints can be embedded into a heuristic algorithm framework. Combined with multi-objective particle swarm optimization, the Pareto front is efficiently searched within the feasible region, and multi-timescale coordination is achieved. This method effectively alleviates the contradiction between source-load uncertainty handling and economic efficiency. The integration of partial Bruker optimization and heuristic algorithms improves solution speed and safety. Multi-objective optimization supports scientific trade-offs, and multi-timescale coordination enhances the economy and reliability of system operation. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the integrated energy system architecture for a park provided in an embodiment of the present invention; Figure 2 A flowchart illustrating the implementation of a comprehensive energy dispatching method for industrial parks, provided as an embodiment of the present invention; Figure 3 A core flowchart of a comprehensive energy dispatching method for industrial parks provided in this embodiment of the invention; Figure 4 This is a schematic diagram of a park integrated energy dispatching system provided in an embodiment of the present invention. Detailed Implementation

[0022] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0023] As described in the background section, The optimization scheduling method for the park's integrated energy system mainly has the following problems: First, there is a contradiction between handling source-load uncertainty and economic efficiency. Deterministic optimization cannot cope with fluctuations in wind and solar power output; stochastic programming relies on precise probability distributions, which are difficult to satisfy in engineering; traditional robust optimization is based on worst-case scenarios, ensuring safety but sacrificing economic efficiency excessively. Existing technologies use Wasserstein distance to construct fuzzy sets for partial robust optimization, which can alleviate conservatism, but its solution often relies on mathematical programming methods such as column and constraint generation or Benders decomposition, resulting in long solution times (usually tens of seconds), making it difficult to meet the real-time requirements of intraday rolling scheduling.

[0024] Secondly, heuristic algorithms struggle to directly handle split-Brow chance constraints. While heuristic methods like Multi-Objective Particle Swarm Optimization (MOPSO) offer fast solutions, they are primarily applicable to simple or unconstrained problems. Split-Brow chance constraints, however, are complex probabilistic constraints. Directly applying MOPSO would generate numerous infeasible solutions, failing to guarantee the safe operation of the scheduling scheme under wind and solar fluctuations. Currently, split-Brow optimization and MOPSO are applied independently, with no effective integration observed. The technical obstacle lies in the need to first convert split-Brow chance constraints into deterministic linear constraints before MOPSO can efficiently search within the decision space that satisfies safety constraints. However, existing literature and patents do not provide a simple form that can be embedded within the MOPSO framework for this conversion method.

[0025] Third, there is a lack of effective trade-off mechanisms for conflicting multiple objectives. System optimization involves multiple objectives such as economy, low carbon emissions, and reliability. Existing methods mostly use weighted summation, with weights relying on experience, making it difficult to obtain a truly Pareto optimal solution set and unable to flexibly adapt to carbon price fluctuations.

[0026] Fourth, coordination across multiple time scales is inadequate. Most systems employ single-time-scale scheduling, failing to fully leverage the benefits of improved ultra-short-term forecast accuracy, leading to high wind and solar curtailment rates or increased regulation costs.

[0027] To address the aforementioned issues, this embodiment provides a comprehensive energy scheduling method for industrial parks. This method achieves the following objectives: First, under the premise of ensuring safety constraints, it achieves efficient integration of sub-Bruker optimization and multi-objective particle swarm optimization, balancing solution speed and robustness; Second, it obtains Pareto optimal scheduling schemes that decision-makers can flexibly choose from; Third, it achieves coordinated optimization across multiple time scales, including day-ahead, intraday, and real-time.

[0028] For example, such as Figure 3 As shown, this embodiment provides a method for integrated energy dispatching in a park, including: Based on the equipment composition and operating parameters of the park's integrated energy system, establish the electrical power balance equation, the thermal power balance equation, and the operating constraints of each piece of equipment. Based on historical wind and solar power output data, an empirical distribution and Wasserstein distance fuzzy set are constructed. The power balance equation, which includes wind and solar prediction errors, is transformed into a split-Bruker chance constraint. It is then equivalently transformed into a deterministic linear constraint through dual theory. The power balance equation, thermal power balance equation, and the operating constraints of each device are combined to form a scheduling safe and feasible region. With minimizing the total system operating cost as the optimization objective, a day-ahead scheduling mixed integer linear programming model is constructed based on the scheduling safe and feasible region. By inputting the day-ahead predicted operating data, the day-ahead start-stop status and day-ahead output plan of each device are obtained by solving the model. Using the day-ahead start-stop state as a hard constraint and the day-ahead output plan as a benchmark, an initial population is generated. Intraday rolling multi-objective particle swarm optimization is performed within the scheduling safety and feasibility domain to obtain the Pareto optimal scheduling solution set. Based on real-time collected measured operation data, the system uses the state-space model of the park's integrated energy system as the prediction model, executes model predictive control within the safe and feasible scheduling domain, and makes real-time corrections with reference to the intraday scheduling scheme selected by the Pareto optimal scheduling solution set, outputting real-time control commands for each device.

[0029] The scheduling method provided in this embodiment will be further explained below: For example, such as Figure 2 As shown in the figure, the specific steps of the integrated energy dispatching method for a park provided in this embodiment are as follows: Phase 1: Constructing a comprehensive energy system model for the park. This involves establishing a "source-grid-load-storage" system model that includes photovoltaic, wind power, gas turbines, electric boilers, electric-to-gas conversion equipment, battery energy storage, and thermal storage tanks; determining the operational constraints for each device, including upper and lower output limits, ramp rate constraints, mutual exclusion constraints for energy storage charging and discharging, and tie-line power constraints; and establishing electrical power balance equations and thermal power balance equations.

[0030] The second stage involves establishing a partial Bruker chance-constrained optimization model and implementing an equivalent transformation. The objective function is to minimize the total system operating cost, encompassing electricity purchase cost, gas purchase cost, equipment operation and maintenance cost, carbon trading cost, and electricity sales revenue. A fuzzy set of wind and solar power output probability distributions is constructed based on Wasserstein distance, transforming the power balance constraint containing uncertain variables into a partial Bruker chance constraint. Through mathematical derivation, the partial Bruker chance constraint is equivalently transformed into a set of deterministic linear constraints, resulting in a efficiently solvable mixed-integer linear programming model. This transformation eliminates probabilistic constraints, providing a decision space that fully satisfies safety constraints for subsequent multi-objective particle swarm optimization algorithms.

[0031] Phase 3: Precise solution for the day-ahead phase. Using a 24-hour scheduling cycle and a 1-hour resolution, based on the next day's wind and solar power output forecast data, a mathematical programming solver is invoked to precisely solve the mixed-integer linear programming model obtained in Step 2, determining the start-up and shutdown status and basic power output plan for each unit.

[0032] Phase 4: Intraday Rolling Optimization. Using a 15-minute cycle and a 4-hour rolling window, wind and solar power output information is updated based on ultra-short-term forecast data. Using the deterministic constraints obtained in Step 2 as the safety boundary, an improved multi-objective particle swarm optimization algorithm is employed for rapid optimization. Based on the daily plan, random perturbations are added to generate the initial population; adaptive inertia weights and Gaussian mutation operators are introduced, and a correlation mechanism between mutation probability and Wasserstein radius is established; under Pareto dominance, an external archive is maintained, and a Pareto optimal solution set is output for decision-makers to choose from.

[0033] Phase 5: Real-time Phase Correction. Using a 5-minute cycle, model predictive control is employed to fine-tune the daily scheduling plan based on measured operational data, compensating for prediction errors and ensuring stable system operation.

[0034] As a preferred embodiment, the scheduling method is further described below: like Figure 1 As shown, this embodiment takes a typical industrial park's integrated energy system as the implementation object. The following parameters are only preferred implementation schemes, and those skilled in the art can adjust them according to the actual scale.

[0035] This system adopts an integrated "source-grid-load-storage" architecture, encompassing three major subsystems: power supply, heating, and gas supply. The system configuration includes: a 2MW (peak) photovoltaic power generation array, a 3MW (rated) wind turbine generator, a 5MW gas turbine, a 2MW electric boiler, a 1MW power-to-gas conversion device, a 4MWh battery energy storage system, a 6MWh thermal storage tank, and a maximum grid interconnection power of 8MW. The peak electrical load of the industrial park is approximately 8MW, and the peak thermal load is approximately 5MW. The specific implementation of this invention is described in detail below according to the implementation sequence, with the specific steps as follows: Step 1: Determine the system equipment composition and parameters. Based on the actual energy demand and equipment configuration of the park, the system comprises seven main types of equipment: photovoltaic power generation array, wind turbine generator set, gas turbine, electric boiler, electric-to-gas conversion equipment, battery energy storage system, and thermal storage tank, as well as the external power grid interface. Record the rated capacity, efficiency parameters, and operating limitations of each device. The gas turbine has a power generation efficiency of 0.38, a heat-to-power ratio of 1.35, an upper limit of ramp rate of 1.5 MW / h, and a lower limit of ramp rate of 1.0 MW / h. The battery charge / discharge efficiency is 0.92, the allowable state of charge range is 0.2 to 0.9, and the maximum interconnection power of the grid interconnection line is 8 MW.

[0036] Step 2: Establish the power balance equation. During any scheduling period t, the sum of the output of all generating equipment in the system equals the sum of the power consumption of all consuming equipment:

[0037] in, For the power generation of gas turbines, and The actual output of photovoltaic and wind power, respectively. and These are the electricity purchased online and the electricity sold online, respectively. and These are the battery discharge and charging power, respectively. For electrical load, For the power consumption of the electric boiler, This refers to the electrical power consumption of the power-to-gas conversion equipment. The actual output of photovoltaic and wind power is expressed as the sum of the day-ahead forecast and the forecast error:

[0038]

[0039] in, and The day-ahead forecasts for solar and wind power are respectively provided. and This represents the prediction error.

[0040] Step 3: Next, establish the heat power balance equation. The heat supply of the heat source equipment equals the sum of the heat load and the heat charge of the storage tank:

[0041] in The thermoelectric ratio of the gas turbine is 1.35 in this embodiment; and These are the heat release and heat charging power of the thermal storage tank, respectively. This refers to the heat load.

[0042] Step 4: Determine the operating constraints for each piece of equipment. Each piece of equipment is subject to upper and lower output limits:

[0043] in, The start / stop state variable is 0-1. MW, MW.

[0044] The gas turbine also needs to meet the ramp rate constraint:

[0045] in, MW / h, MW / h.

[0046] The battery must satisfy the state of charge update equation, the state of charge range constraint, and the charge-discharge mutual exclusion constraint:

[0047]

[0048]

[0049] in, E bat =4MWh, which is the rated capacity of the battery, and Δt is the length of the dispatch period. The state of charge needs to be maintained at... Within the range. The mutual exclusion constraint between charge and discharge states is: The scheduling cycle must meet the following conditions at the end. This ensures a balanced cycle.

[0050] The tie line constraint limits the power purchase and sale to no more than 8MW, and power purchase and sale cannot be carried out simultaneously.

[0051] Step 5: Establish the objective function, with minimizing the total system operating cost as the objective function:

[0052] Electricity purchase cost Time-of-use pricing is used. Gas purchase cost. The price of natural gas is approximately 3.25 yuan / m³. Maintenance costs... The carbon trading cost is calculated as follows: gas turbine: 0.032 yuan / kWh; electric boiler: 0.008 yuan / kWh; electricity-to-gas conversion equipment: 0.045 yuan / kWh. carbon price The cost is approximately 80 yuan / tCO2, and the carbon emissions consist of three parts:

[0053] in, tCO2 / MWh, tCO2 / MWh, Free carbon credits.

[0054] Step 6: Construct the distributed bar chance constraint. Construct an empirical distribution based on 365 days of historical wind and solar power output data. Take the confidence level. Calculate the Wasserstein distance radius:

[0055] Construct a fuzzy set of probability distributions centered on the empirical distribution and bounded by this radius:

[0056] The power balance equation, which includes wind and solar forecasting errors, is rewritten in a distributed bar chance constraint form, taking the probability of violation. :

[0057] The constraint means that, under all possible probability distributions in the fuzzy set, the probability of satisfying the power balance is not less than 95%.

[0058] Step 7: Equivalent transformation to deterministic linear constraints. Using the properties of Wasserstein distance and duality theory, the split-Bruker chance constraint constructed in the previous step is equivalently transformed into a linear inequality:

[0059] in, For the Wasserstein radius probability of violation and historical data sample size The jointly determined safety margin constant can be directly calculated from the empirical distribution quantiles of historical data. Through this transformation, the original optimization model with probabilistic constraints becomes a standard mixed-integer linear programming model, which no longer contains any probability terms. This transformation step is the core of the entire method, providing a decision space that fully satisfies the safety constraints for the subsequent multi-objective particle swarm optimization algorithm.

[0060] Step 8: Implement day-ahead scheduling. The scheduling cycle is 24 hours, with a time resolution of 1 hour, for a total of 24 time periods. Input the following 24-hour forecast curves for electricity load, heat load, photovoltaic output, wind power output, time-of-use electricity price, and natural gas price. Input the transformed mixed-integer linear programming model into the Gurobi mathematical programming solver, setting the MIPGap tolerance to 1% and the maximum solution time to 300 seconds. The solver outputs the start-up and shutdown status and output plan for each unit for 24 time periods. The unit start-up and shutdown status is passed as a hard constraint to the intraday stage and remains unchanged during the day.

[0061] Step 9: Implement intraday rolling optimization. The intraday time step is 15 minutes, and the rolling window length is 4 hours, i.e., looking forward 16 time periods. A complete optimization calculation is performed once per rolling cycle, but only the scheduling instructions for the first time period are executed. The next cycle window is moved forward 15 minutes before recalculation. Algorithm parameter settings: Population size. Maximum number of iterations The external file capacity is 100. Based on the current plan, a random uniform perturbation within ±15% is applied to the position vector of each particle to generate the initial population. All particles must fall within the safe and feasible region defined by the deterministic constraints obtained in step 7; violators are absorbed by the boundary.

[0062] Step 10: Particle Fitness Evaluation and Ranking. Calculate two fitness values ​​for each particle: total operating cost. and total carbon emissions All values ​​are calculated based on cumulative values ​​over 16 rolling periods. Pareto dominance relations are used for sorting; solution A dominates solution B if and only if A is not inferior to B on both objectives and is strictly superior to B on at least one objective. All solutions not dominated by any other solution constitute the current Pareto front and are stored in an external archive. The external archive's size is controlled by an adaptive mesh method; when the archive is full, solutions in the densest mesh regions are deleted to maintain a uniform distribution of solutions.

[0063] Step 11, Particle Velocity and Position Update. The velocity and position of each particle are updated using the following formula:

[0064]

[0065] in, and As a learning factor, and A random number between [0,1] This represents the historical best position of particle i. A global leader selected from external archives. Adaptive inertia weights. The value decreases linearly from an initial value of 0.9 to a final value of 0.4 with the number of iterations.

[0066] After the update, boundary absorption processing is applied to particles that exceed the constraint boundaries.

[0067] Step 12: Apply adaptive mutation operation. (Using probability...) Apply Gaussian mutations to each dimension of the updated particle position:

[0068] Mutation probability With Wasserstein radius Related:

[0069] in Based on the probability of mutation, This is the sensitivity coefficient. The reference radius. When Increasing the power output indicates a greater uncertainty in wind and solar power output, which in turn increases the probability of mutation and enhances global exploration capabilities; when... Decreasing the value indicates that when uncertainty is reduced, the probability of mutation decreases accordingly, thus accelerating local convergence.

[0070] Step 13: Convergence Check and Output. If the number of iterations reaches... Or the relative change in the hypervolume index of the Pareto front is less than 30 generations. If the condition is not met, the iteration terminates. All Pareto optimal solutions stored in the external archive are output, allowing decision-makers to select an economical, balanced, or low-carbon scheduling scheme based on current carbon trading prices and actual needs. By converting opportunity constraints into deterministic linear constraints, the multi-objective particle swarm optimization algorithm can efficiently search within the safe and feasible region, effectively reducing the time required for a single solution and meeting the real-time requirements of intraday rolling optimization.

[0071] Step 14: Implement real-time correction. During the real-time operation phase, a 5-minute control cycle is used. A model predictive control method is employed, using the state-space model of the park's integrated energy system as the predictive model and the current actual output state of each device as the initial state. The optimization objective is to minimize the weighted sum of squares of the deviations between the output of each device and the intraday reference plan. The constraints are consistent with the deterministic constraints in Step 7, but the control time domain is shortened to 15 minutes. Each control cycle collects the current measured wind and solar power output and load data, updates the initial state, solves the constraint optimization problem to obtain the control commands for the current 5 minutes, and executes only the commands for the first control cycle. The process is repeated in the next cycle.

[0072] For example, such as Figure 4 As shown, this embodiment also provides a comprehensive energy dispatching system for a park, including: A module is established to create electrical power balance equations, thermal power balance equations, and operating constraints for each device based on the equipment composition and operating parameters of the park's integrated energy system. The constraint transformation module is used to construct an empirical distribution and Wasserstein distance fuzzy set based on historical wind and solar power output data. It transforms the power balance equation, which includes wind and solar prediction errors, into a partial Bruker chance constraint, and then transforms it into a deterministic linear constraint through dual theory. The module combines the power balance equation, the thermal power balance equation, and the operating constraints of each device to form a scheduling safe and feasible domain. The solution module is used to construct a day-ahead scheduling mixed integer linear programming model based on the scheduling safety feasible region with the optimization objective of minimizing the total system operating cost. It takes the day-ahead predicted operating data as input and solves to obtain the day-ahead start-up and shutdown status and day-ahead output plan of each device. The optimization module is used to generate an initial population based on the day-ahead start-stop status as a hard constraint and the day-ahead output plan as a benchmark, and to perform intraday rolling multi-objective particle swarm optimization within the scheduling safety and feasible domain to obtain the Pareto optimal scheduling solution set. The correction module is used to perform model predictive control within the safe and feasible domain of scheduling, based on the real-time collected measured operating data and the state space model of the park's integrated energy system as the prediction model. It uses the intraday scheduling scheme selected by the Pareto optimal scheduling solution set as a reference for real-time correction and outputs real-time control commands for each device.

[0073] The present invention also provides a park integrated energy dispatching device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the park integrated energy dispatching method.

[0074] The present invention also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the integrated energy dispatching method for the park.

[0075] When the processor executes the computer program, it implements the above-mentioned steps of integrated energy scheduling in the park, such as: establishing electrical power balance equations, thermal power balance equations, and operating constraints of each device based on the equipment composition and operating parameters of the integrated energy system in the park. Based on historical wind and solar power output data, an empirical distribution and Wasserstein distance fuzzy set are constructed. The power balance equation, which includes wind and solar prediction errors, is transformed into a split-Bruker chance constraint. It is then equivalently transformed into a deterministic linear constraint through dual theory. The power balance equation, thermal power balance equation, and the operating constraints of each device are combined to form a scheduling safe and feasible region. With minimizing the total system operating cost as the optimization objective, a day-ahead scheduling mixed integer linear programming model is constructed based on the scheduling safe and feasible region. By inputting the day-ahead predicted operating data, the day-ahead start-stop status and day-ahead output plan of each device are obtained by solving the model. Using the day-ahead start-stop state as a hard constraint and the day-ahead output plan as a benchmark, an initial population is generated. Intraday rolling multi-objective particle swarm optimization is performed within the scheduling safety and feasibility domain to obtain the Pareto optimal scheduling solution set. Based on real-time collected measured operation data, the system uses the state-space model of the park's integrated energy system as the prediction model, executes model predictive control within the safe and feasible scheduling domain, and makes real-time corrections with reference to the intraday scheduling scheme selected by the Pareto optimal scheduling solution set, outputting real-time control commands for each device.

[0076] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing preset functions, the instruction segments describing the execution process of the computer program in the integrated energy dispatching equipment of the park. For example, the computer program can be divided into a setup module, a constraint transformation module, a solution module, an optimization module, and a correction module; the specific functions are as follows: The setup module is used to establish the power balance equation, the thermal power balance equation, and the operating constraints of each device based on the equipment composition and operating parameters of the integrated energy system of the park; the constraint transformation module is used to construct an empirical distribution and a Wasserstein distance fuzzy set based on historical wind and solar power output data, transforming the power balance equation, including wind and solar prediction errors, into a partial Bruker chance constraint, and equivalently transforming it into a deterministic linear constraint through dual theory, combining the power balance equation, the thermal power balance equation, and the operating constraints of each device to form a dispatching safe and feasible region; the solution module is used to... With minimizing operating costs as the optimization objective, a day-ahead scheduling mixed-integer linear programming model is constructed based on the scheduling safety feasible region. The model is solved by inputting day-ahead predicted operating data to obtain the day-ahead start-stop status and day-ahead output plan for each device. An optimization module generates an initial population based on the day-ahead start-stop status as a hard constraint and the day-ahead output plan as a baseline. It then performs intraday rolling multi-objective particle swarm optimization within the scheduling safety feasible region to obtain the Pareto optimal scheduling solution set. A correction module, based on real-time collected measured operating data and using the state-space model of the park's integrated energy system as a prediction model, performs model predictive control within the scheduling safety feasible region. It uses the intraday scheduling scheme selected from the Pareto optimal scheduling solution set as a reference for real-time correction and outputs real-time control commands for each device.

[0077] The integrated energy dispatching equipment for the park can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The integrated energy dispatching equipment may include, but is not limited to, processors and memory. Those skilled in the art will understand that the above are examples of integrated energy dispatching equipment for the park and do not constitute a limitation on it. It may include more components than described above, or combine certain components, or use different components. For example, the integrated energy dispatching equipment may also include input / output devices, network access devices, buses, etc.

[0078] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or any conventional processor. This processor serves as the control center for the integrated energy dispatching system in the park, connecting various parts of the integrated energy dispatching equipment throughout the park via various interfaces and lines.

[0079] The memory can be used to store the computer program and / or modules. The processor realizes various functions of the park's integrated energy dispatching equipment by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory.

[0080] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function (such as sound playback, image playback, etc.). The data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital cards (SD cards), flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0081] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the aforementioned integrated energy dispatching method for a park.

[0082] If the modules / units integrated into the park's comprehensive energy dispatch system are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0083] Based on this understanding, the present invention can implement all or part of the processes in the above-mentioned integrated energy dispatching method for industrial parks, or it can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above-mentioned integrated energy dispatching method for industrial parks. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or a preset intermediate form, etc.

[0084] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0085] It should be noted that the content contained in the computer-readable storage medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0086] Compared with existing scheduling measures, this invention provides a comprehensive energy scheduling method and system for industrial parks, which has the following advantages: This invention transforms the chance constraints of the split-blob bar into deterministic linear constraints, constructing a two-layer solution framework: "the split-blob bar determines the safety boundary, and the improved multi-objective particle swarm optimization algorithm searches for Pareto optimality within the safety domain." Furthermore, an adaptive correlation mechanism between mutation probability and Wasserstein radius is introduced in the intraday rolling optimization, enabling the system to coordinate and optimize at three time scales: 1 hour before the day, 15 minutes intraday, and 5 minutes in real-time. Compared to existing technologies where split-blob bar optimization relies on mathematical programming methods such as column and constraint generation, resulting in excessively long solution times that fail to meet the real-time requirements of intraday rolling scheduling, or where multi-objective particle swarm optimization directly handles probabilistic constraints leading to potential violations of safety constraints, this method overcomes the technical obstacle of effectively integrating split-blob bar optimization and heuristic algorithms due to complex probabilistic constraints. While ensuring that the safety constraints of the scheduling scheme are fully satisfied, it compresses the single solution time from tens of seconds to sub-seconds, while outputting a Pareto optimal solution set that balances economic efficiency and low carbon emissions, achieving a significant reduction in wind and solar curtailment rates and an effective improvement in overall energy utilization.

[0087] The above embodiments are merely one of the implementation methods for achieving the technical solution of the present invention. The scope of protection claimed by the present invention is not limited to this embodiment, but also includes any variations, substitutions and other implementation methods that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for integrated energy dispatching in a park, characterized in that, include: Based on the equipment composition and operating parameters of the park's integrated energy system, electrical power balance equations, thermal power balance equations, and operating constraints of each piece of equipment are established. Based on historical wind and solar power output data, an empirical distribution and Wasserstein distance fuzzy set are constructed. The power balance equation, which includes wind and solar prediction errors, is transformed into a split-Bruker chance constraint. It is then equivalently transformed into a deterministic linear constraint through dual theory. The power balance equation, thermal power balance equation, and the operating constraints of each device are combined to form a scheduling safe and feasible region. With minimizing the total system operating cost as the optimization objective, a day-ahead scheduling mixed integer linear programming model is constructed based on the scheduling safe and feasible region. By inputting the day-ahead predicted operating data, the day-ahead start-stop status and day-ahead output plan of each device are obtained by solving the model. Using the day-ahead start-stop state as a hard constraint and the day-ahead output plan as a benchmark, an initial population is generated. Intraday rolling multi-objective particle swarm optimization is performed within the scheduling safety and feasibility domain to obtain the Pareto optimal scheduling solution set. Based on real-time collected measured operation data, the system uses the state-space model of the park's integrated energy system as the prediction model, executes model predictive control within the safe and feasible scheduling domain, and makes real-time corrections with reference to the intraday scheduling scheme selected by the Pareto optimal scheduling solution set, outputting real-time control commands for each device.

2. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, The integrated energy system of the park adopts an integrated source-grid-load-storage architecture, including power supply, heating and gas supply subsystems. The equipment consists of photovoltaic power generation arrays, wind turbine generators, gas turbines, electric boilers, electric-to-gas conversion equipment, battery energy storage systems, thermal storage tanks and grid interconnection lines. The operating constraints of each piece of equipment include upper and lower limits of equipment output, gas turbine ramp rate constraints, state of charge constraints and charge / discharge mutual exclusion constraints of battery energy storage systems, and power interaction constraints of grid interconnection lines. The power balance equation is: During any scheduling period t, the sum of the output of all generating equipment in the system equals the sum of the consumption of all consuming equipment. in, For the power generation of gas turbines; and These refer to the actual output of photovoltaic and wind power, respectively. and These refer to the power consumption for online electricity purchases and the power consumption for retail sales, respectively. and These are the battery discharge and charging power, respectively. For electrical load; This refers to the electrical power consumption of the electric boiler. The power consumption of the electro-gas conversion equipment; The actual output of photovoltaic and wind power is expressed as the sum of the day-ahead forecast and the forecast error: in, and The day-ahead forecasts for solar and wind power are respectively provided. and This represents the prediction error; The heat power balance equation is: the heat supply of the source equipment equals the sum of the heat load and the heat charge of the storage tank. in, The heat-to-electric ratio of a gas turbine; and These are the heat release and heat charging power of the thermal storage tank, respectively. This refers to the heat load.

3. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, The process of constructing the bibliophilic chance constraint is as follows: An empirical distribution is constructed based on historical wind and solar power output data for a preset number of days. The Wasserstein distance radius is calculated according to a preset confidence level. A fuzzy set of probability distribution is constructed with the empirical distribution as the center and the Wasserstein distance radius as the boundary. A preset violation probability is set to rewrite the power balance equation, which includes wind and solar prediction errors, into a split-Bruker chance constraint form. Then, it is equivalently transformed into a deterministic linear constraint through dual theory.

4. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, The day-ahead forecast operating data includes electricity load forecast curve, heat load forecast curve, photovoltaic power output forecast curve, wind power output forecast curve, time-of-use electricity price, and natural gas price; A mathematical programming solver is used to solve the day-ahead scheduling mixed integer linear programming model, and the day-ahead start and stop states obtained from the solution are passed as hard constraints to the intraday optimization stage.

5. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, The velocity and position of particles are updated using adaptive inertial weights, and an adaptive Gaussian mutation operation is applied to the updated particle positions. The mutation probability is positively correlated with the Wasserstein distance radius, and the mutation probability increases with the greater uncertainty of wind and solar power output. All particles satisfy the constraint boundary of the scheduling safe and feasible region, and boundary absorption processing is performed on particles that go out of bounds.

6. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, When performing intraday rolling multi-objective particle swarm optimization, the total operating cost and total carbon emissions are calculated for each particle. All particles are sorted using the Pareto dominance relation. The Pareto optimal solution is stored in an external archive and the size of the external archive is controlled by an adaptive mesh method.

7. The integrated energy dispatching method for industrial parks according to claim 1, characterized in that, In the process of real-time correction based on the intraday scheduling scheme selected by the Pareto optimal scheduling solution set, the optimization objective is to minimize the weighted sum of squares of the deviation between the output of each device and the intraday reference plan. The initial state of the system is updated in each control cycle, and the constrained optimization problem is solved within the scheduling safety feasible region. Only the control instructions of the first control cycle are executed.

8. A comprehensive energy dispatching system for industrial parks, characterized in that, include: A module is established to create electrical power balance equations, thermal power balance equations, and operating constraints for each device based on the equipment composition and operating parameters of the park's integrated energy system. The constraint transformation module is used to construct an empirical distribution and Wasserstein distance fuzzy set based on historical wind and solar power output data. It transforms the power balance equation, which includes wind and solar prediction errors, into a partial Bruker chance constraint, and then transforms it into a deterministic linear constraint through dual theory. The module combines the power balance equation, the thermal power balance equation, and the operating constraints of each device to form a scheduling safe and feasible domain. The solution module is used to construct a day-ahead scheduling mixed integer linear programming model based on the scheduling safety feasible region with the optimization objective of minimizing the total system operating cost. It takes the day-ahead predicted operating data as input and solves to obtain the day-ahead start-up and shutdown status and day-ahead output plan of each device. The optimization module is used to generate an initial population based on the day-ahead start-stop status as a hard constraint and the day-ahead output plan as a benchmark, and to perform intraday rolling multi-objective particle swarm optimization within the scheduling safety and feasible domain to obtain the Pareto optimal scheduling solution set. The correction module is used to perform model predictive control within the safe and feasible domain of scheduling, based on the real-time collected measured operating data and the state space model of the park's integrated energy system as the prediction model. It uses the intraday scheduling scheme selected by the Pareto optimal scheduling solution set as a reference for real-time correction and outputs real-time control commands for each device.

9. A comprehensive energy dispatching device for industrial parks, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the integrated energy dispatching method for the park as described in any one of claims 1-7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it is used to implement the integrated energy dispatching method for the park as described in any one of claims 1-7.