A comprehensive energy scheduling method and system considering wind and light load uncertainty

Abstract

The application discloses a kind of comprehensive energy scheduling method and system considering wind and light load uncertainty, the comprehensive energy scheduling method constructs the comprehensive energy system model covering energy supply equipment, energy conversion equipment and energy storage equipment, through the coordinated scheduling of multiple energy, and introduce load demand response mechanism flexible adjustment energy supply and demand relationship;In the aspect of describing uncertain factors, the present application generates wind and light output and load demand scene covering uncertainty space based on Monte Carlo sampling of multiple probability distribution functions, and obtains typical scene by combining K-means clustering algorithm and synchronous back substitution elimination technology for scene reduction;In the aspect of processing uncertain factors, a two-stage robust optimization method is introduced, the fluctuation range of probability distribution is limited by synthesizing norm constraint distribution fuzzy set, the influence of deviation on optimization result is reduced, so as to process the uncertainty of wind and light output and load demand.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of integrated energy system scheduling, and particularly relates to an integrated energy scheduling method and system considering wind and light load uncertainty. BACKGROUND

[0002] An integrated energy system (IES) is an innovative energy supply system integrating multiple energy forms, which can realize the collaborative complementation between different energy carriers, has significant advantages in improving energy utilization efficiency, economic feasibility and reducing environmental pollution, etc. The advanced optimization scheduling strategy of the integrated energy system integrating renewable energy can effectively improve the reliability and flexibility of the supply side, and therefore this integration method is considered as one of the effective methods to solve the current resource shortage and environmental challenges in the energy field. Renewable energy has been used in many occasions, such as rural areas, communities and industrial parks, but the uncertainty of renewable energy in the IES has caused hidden troubles to the demand for stable energy supply on the user side. Similarly, the volatility of load demand also brings great challenges to the stable operation and optimal scheduling of the system. The traditional deterministic optimization method is difficult to guarantee robustness, and the stochastic optimization method relies on accurate probability distribution, actual data is scarce and the distribution is difficult to accurately describe, resulting in conservative or unstable optimization results. The robust optimization method only considers the worst case, which may sacrifice the economy, and it is difficult to balance the complex constraints of multi-energy coupling. Therefore, how to effectively manage the impact of uncertainty factors on the system and realize the minimization of system operation cost and stable optimal operation of the IES, and efficiently utilize energy, has become a problem to be solved. SUMMARY

[0003] The purpose of the application is to provide an integrated energy scheduling method and system considering wind and light load uncertainty.

[0004] In a first aspect, the application provides an integrated energy scheduling method considering wind and light load uncertainty, which comprises the following steps:

[0005] The micro-combustion engine start-stop cost and the total system operation cost are taken as the optimization objectives, a target function and constraint conditions of a comprehensive energy system model covering energy supply equipment, energy conversion equipment and energy storage equipment are constructed, a probability distribution model of photovoltaic output, wind turbine output and electric heating load is respectively constructed, and random sampling is performed according to different probability distribution models and constraint conditions to obtain scenes of different probabilities of photovoltaic units, wind turbine units, electric loads and heating loads, so as to obtain a typical scene set of wind-solar-load combined output and corresponding probabilities, a distribution robust optimization model is constructed, and the probability distribution of the typical scene is constrained to obtain an uncertain probability distribution fuzzy set, the distribution robust optimization model is decomposed into a main problem and a sub-problem, the main problem and the sub-problem are iterated through a decomposition algorithm based on columns and constraints until a termination condition is reached, and an optimal comprehensive energy scheduling scheme is obtained.

[0006] As preferred, the probability distribution model of photovoltaic output is constructed based on a beta distribution, the probability distribution model of wind turbine output is constructed based on a Weibull distribution, and the probability distribution model of electric heating load is constructed based on a normal distribution.

[0007] As preferred, the target function of the constructed comprehensive energy system model is as follows:

[0008]

[0009] In the formula, is the micro-combustion engine start-stop cost; is the micro-combustion engine start-stop cost coefficient; is the is the micro-combustion engine state variable at the moment, taking values of 0 or 1; is the total system operation cost; is the power grid purchase cost; is the gas grid purchase cost; is the energy storage cost; is the abandoned wind cost; is the abandoned light cost; is the electricity sales revenue; is the demand response cost.

[0010] As preferred, the demand response cost is introduced in the target function corresponding to the total system operation cost; the demand response cost includes compensation fees of a translatable load participating in demand response, compensation fees of a transferable load after participating in demand response and compensation fees of a cuttable load after participating in demand response.

[0011] As preferred, the specific process of obtaining the typical scene set of wind-solar-load combined output is as follows:

[0012] The K-means clustering algorithm is used to compress the sampling scenes, extract a typical scene set of photovoltaic, fan, electrical load and thermal load independent output, and connect these different kinds of scenes in Cartesian form, and further reduce the scenes by using the synchronous back substitution elimination technology to obtain a typical scene set composed of multiple groups of wind and light load joint output typical scenes.

[0013] As preferred, the scene clustering number in the K-means clustering algorithm is obtained by using the profile coefficient method.

[0014] As preferred, the probability distribution of the typical scene is constrained by using a comprehensive norm.

[0015] As preferred, the objective function of the main problem includes the operation cost, the compensation cost and the variable related to the uncertainty.

[0016] As preferred, the specific process of iteratively solving the two-stage distribution robust optimization model is as follows:

[0017] Initialize the iteration number of uncertain variables; set the convergence precision, upper limit value and lower limit value respectively; set the probability of each typical scene to the reference probability distribution value; solve the main problem to obtain the first-stage variable and the optimal solution, and update the lower limit value according to the optimal solution corresponding to the main problem; solve the sub-problem based on the first-stage variable to obtain the second-stage variable and the optimal solution, and update the upper limit value according to the optimal solution corresponding to the sub-problem; if the difference between the upper limit value and the lower limit value is greater than the convergence precision, add the second-stage variable and the constraint condition related thereto to the main problem, and solve the main problem again until the difference between the upper limit value and the lower limit value is less than or equal to the convergence precision, stop iteration, and obtain the optimal solution.

[0018] In the second aspect, the application provides a comprehensive energy scheduling system considering the uncertainty of wind and light load, which comprises energy supply equipment, energy conversion equipment, energy storage equipment and a control module; the energy supply equipment comprises photovoltaic units and wind turbine units; the energy conversion equipment comprises electric boilers and combined heat and power units; the energy storage equipment comprises storage batteries and heat storage tanks; the control module is used to execute the above-mentioned comprehensive energy scheduling method, and controls the energy supply equipment, the energy conversion equipment and the energy storage equipment based on the generated optimal comprehensive energy scheduling scheme.

[0019] The application has the beneficial effects that:

[0020] 1. The application is based on the combination of Monte Carlo sampling method of multiple probability distribution functions and K-means clustering algorithm to construct a typical scenario for describing the uncertainty of wind and light load; Monte Carlo can capture extreme scenarios through complete random sampling to test the performance of the system in the worst case, ensure the reliability of the scheduling scheme, solve the problem of insufficient scene coverage in the prior art that cannot fully capture extreme cases, enhance the robustness of the model, and reduce the high cost risk in actual operation.

[0021] 2. The application adopts a two-stage distribution robust optimization method to process the wind and light load uncertainty in the system, realizes the balance between economy and robustness in the process of optimal scheduling, and optimizes the electric and thermal load distribution by introducing the operation mechanism of demand response, further optimizes the system operation, and reduces the system operation cost, so that the application can timely adjust the unit output to find the optimal solution to optimize the cost under different scenarios.

[0022] 3. The application adopts comprehensive norm constraint typical scenario probability distribution, which is better than the single norm constraint in economy; compared with the traditional random optimization and robust optimization methods for handling uncertainty optimization problems, the robustness of the application is better than that of random optimization, and the economy is more prominent than that of robust optimization. DETAILED DESCRIPTION

[0023] In order to more clearly set forth the technical scheme of the application, the following will briefly introduce the drawings needed to be used according to the embodiments, and obviously, the description in the drawings is only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor under the premise of the drawings.

[0024] Figure 1 The overall flowchart of the application.

[0025] Figure 2 The flowchart of the combined scenario generation in the application.

[0026] Figure 3 The flowchart of solving the distribution robust optimization model in the application.

[0027] Figure 4 The wind and light combined output and load scenario in the application; wherein, (a) is a photovoltaic output schematic diagram; (b) is a wind turbine output schematic diagram; (c) is an electric load schematic diagram; (d) is a thermal load schematic diagram.

[0028] Figure 5The following are the combined wind and solar power output and load scenarios based on Latin hypercube sampling and K-means clustering algorithms: (a) is the photovoltaic power output scenario; (b) is the wind turbine power output scenario; (c) is the electrical load scenario; and (d) is the thermal load scenario.

[0029] Figure 6 Schematic diagrams of the combined demand response load optimization for different energy storage configurations; where (a) is a schematic diagram of a scenario without energy storage; (b) is a schematic diagram of a scenario combining electricity storage and thermal storage; (c) is a schematic diagram of an electricity storage scenario; and (d) is a schematic diagram of a thermal storage scenario.

[0030] Figure 7 The diagram shows the results of the electrothermal optimization. Among them, (a) is a schematic diagram of the electrical energy optimization results; and (b) is a schematic diagram of the thermal energy optimization results. Detailed Implementation

[0031] The present invention will be further described below with reference to the accompanying drawings.

[0032] like Figure 1 As shown, a comprehensive energy dispatching method considering the uncertainty of wind and solar load includes the following steps:

[0033] Step 1: Construct a comprehensive energy system model

[0034] 1-1. Objective Function

[0035] The integrated energy system supplies and stores energy through power supply equipment, energy conversion equipment, and energy storage equipment. Power supply equipment includes photovoltaic (PV) units and wind turbines (WT); energy conversion equipment includes electric boilers (EB) and combined heat and power (CHP) units; energy storage equipment includes batteries and thermal storage tanks. These devices coordinate their operation to meet the system's electrical and thermal loads under a demand response mechanism. In addition, the integrated energy system also supplies energy through grid electricity purchases and gas grid gas purchases. With the goal of minimizing the start-up and shutdown costs of the micro-turbines in the CHP unit and the total system operating cost, the objective function for constructing the integrated energy system model encompassing power supply equipment, energy conversion equipment, and energy storage equipment is as follows:

[0036]

[0037] In the formula, Cost of starting and stopping the micro gas turbine; This refers to the start-stop cost coefficient for micro gas turbines. for The state variable of the micro gas turbine at any given moment takes the value of 0 or 1; The total operating cost of the system; The cost of purchasing electricity from the power grid; Cost of purchasing gas online; For energy storage costs; Cost of wind curtailment; Cost of abandoning light; For revenue from electricity sales; Cost of responding to demand.

[0038] Electricity purchase cost The cost of natural gas required for combined heat and power (CHP) Energy storage costs Cost of wind curtailment Cost of abandoned light Electricity sales revenue and demand response costs It is expressed as follows:

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] In the formula, The price at which IES purchases electricity from the upper-level power grid; for The amount of electricity that IES purchases from the upper-level power grid at all times; IES purchases gas prices from its superior gas network; for The amount of gas IES purchases from the higher-level gas network at all times; The degradation coefficient of electrical energy storage; The degradation coefficient of thermal energy storage; for The charging power of the energy storage device at any time; for Discharge power of the energy storage device at any given time; for The charging power of the thermal storage equipment at all times; for Discharge power of the thermal storage device at all times; This is the operating cost coefficient for the wind turbine; for Forecast wind power output at any given time; for The actual wind power generated at any given moment, in kW; This is the photovoltaic operating cost coefficient; for Predicted photovoltaic power at any given time; for The actual photovoltaic power involved at any given moment; The price at which IES sells electricity to the upstream power grid; for IES sells electricity to the upper-level power grid at all times; Compensation costs for shiftable loads participating in demand response; This refers to the compensation cost for transferable loads participating in demand response; This is to reduce compensation costs after loads participate in demand response.

[0047] Compensation costs for shiftable loads participating in demand response The model is:

[0048]

[0049] In the formula, The compensation price for a unit power load shift; This refers to the set of time periods in which load shifting begins; This is the power distribution vector before it participates in scheduling; Power distribution vector From the start period Shift to the starting time period The power distribution vector, when At that time, the load did not shift. Power distribution vector The sum of electrical power consumption; Power distribution vector During a certain period of time The translation state takes a value of 0 or 1; and These are the start and end times of the shiftable time period, respectively. Duration; for .

[0050] Compensation costs for transferable loads participating in demand response The model is:

[0051]

[0052] In the formula, The compensation price for a unit power load transfer; for Load power that can be transferred during certain periods; and These are the start and end times of the transfer period, respectively; for The transition status of the time period, with a value of 0 or 1.

[0053] This can reduce the compensation costs for load participation in demand response. The model is:

[0054]

[0055] In the formula, The compensation price for a reduction in unit power load; This is a dummy variable, representing a period during which the load can be reduced. The state of reduction; After participating in scheduling Power during a given time period; Before participating in scheduling Power during a given time period; for Load reduction factor for a given time period.

[0056] 1-2. Constraints

[0057] Transferable load power The power constraints are as follows:

[0058]

[0059] In the formula, and These are the minimum and maximum values ​​of the transferable load power, respectively.

[0060] At the same time, to prevent frequent equipment start-ups and shutdowns due to excessive load dispersion, a minimum continuous operating time for transferable loads is required. To impose constraints. Let from Constantly monitor the operation of the transferred load; if Then in The operating status of the equipment where the load can be transferred at any time has not changed; if At that time, The equipment containing the transferable load begins load transfer at a specific time, starting from that time. arrive During this period, the overall operating status of the equipment should meet the following requirements:

[0061]

[0062] In the formula, This is the minimum continuous running time.

[0063] Constraints are applied to control the maximum and minimum consecutive cut times, as well as the number of cuts:

[0064]

[0065] In the formula, This represents the maximum number of cuts. This is the maximum continuous reduction time; This is the minimum continuous reduction time.

[0066] Wind turbine power constraints:

[0067]

[0068] In the formula, This represents the maximum wind power output. This represents the maximum photovoltaic power.

[0069] CHP unit operating constraints:

[0070]

[0071] In the formula, for time Output electrical power; for Gas-to-electric conversion efficiency; for Input air power; and They are respectively Input gas energy upper and lower limit power; The heat-to-electricity ratio produced by CHP; and They are respectively The upper and lower limits of the thermoelectric ratio.

[0072] Climbing constraints:

[0073]

[0074] In the formula, This is the upper limit of the ramp rate for combined heat and power units; This is the lower limit of the ramp rate for combined heat and power (CHP) units.

[0075] Energy storage device operating constraints:

[0076]

[0077] In the formula, and for Real-time energy storage device charging and discharging power, kW; and for The charging / discharging indicator is set to either 0 or 1. The upper limit for energy storage charging and discharging capacity is kW; for Real-time energy storage capacity, kW; and These are the upper and lower limits of energy storage capacity; and To improve the energy storage charging and discharging efficiency.

[0078] Electric power balance constraints:

[0079]

[0080] In the formula, for Electrical load during a given time period; for The electrical load can be shifted during different time periods; for The amount of electrical load that can be transferred during a given period.

[0081] Thermal power balance constraint:

[0082]

[0083] In the formula, The electrothermal conversion efficiency of the electric boiler; for Electric boiler power at all times; for Heat load at any given time; for The heat load can be reduced at any time.

[0084] Step 2: For the uncertainties in wind and solar power output and electricity and heat load in the integrated energy system, generate multiple representative wind and solar load scenarios and their probability distributions, incorporate the uncertainties into the optimization model, so that the optimization scheduling strategy can cope with the random changes in renewable energy output and load to a certain extent, and improve the system's ability to absorb renewable energy and its operational robustness.

[0085] Based on the historical data characteristics of photovoltaic power output, wind turbine output, and electric and heat load, Monte Carlo (MC) random sampling is used to generate multiple scenarios, and the following probability distribution model is combined to cover the uncertainty space:

[0086] Photovoltaic power output is significantly affected by weather conditions (such as irradiance and cloud cover), and its normalized power value is usually distributed in the interval [0, 1]. Therefore, the Beta distribution is used to describe the probabilistic characteristics of photovoltaic power output, and its probability density function is... for:

[0087]

[0088] In the formula, , These are two shape parameters, and the distribution pattern (such as symmetry and skewness) can be adjusted by fitting historical data. The Beta function is calculated using the following formula: Gamma function In the integer field, satisfy It supports efficient parameter calculation.

[0089] Wind turbine output is closely related to wind speed, and the probability distribution of wind speed usually follows a Weibull distribution. Therefore, the Weibull distribution is chosen to describe the probabilistic characteristics of wind turbine output, and its probability density function is... for:

[0090] In the formula, As a scale parameter, it determines the typical range of wind speed values ​​and reflects the characteristics of "average wind speed"; For shape parameters, when When the Weibull distribution degenerates into an exponential distribution; when At this time, the Weibull distribution is similar to the Rayleigh distribution, and can flexibly adapt to the wind speed characteristics of different regions.

[0091] The uncertainty of electric heating load demand is typically influenced by multiple factors such as user behavior and temperature changes. Its prediction error follows a normal distribution; therefore, the normal distribution is chosen to describe the probabilistic characteristics of load demand. Its probability density function... for:

[0092]

[0093] In the formula, It represents the expected value of the load and reflects the typical energy consumption pattern of users; Indicates the range of load fluctuations. The larger the value, the higher the uncertainty.

[0094] Step 3: Due to the massive number of scenes generated by Monte Carlo sampling, directly using them for optimization would result in extremely high computational complexity. Therefore, the K-means clustering algorithm is used to compress the sampled scenes and extract a set of typical scenes. In determining the number of scene clusters, a larger number is not always better. If the number is too high, it may cause a decrease in the computational speed of the optimization model. Therefore, it is necessary to find an appropriate number of clusters based on the data itself. This invention selects the silhouette coefficient method to determine the number of scene clusters.

[0095] like Figure 2As shown, based on a large amount of scene data generated by MC sampling, the K-means clustering algorithm is used to reduce the number of scenarios with different probabilities of individual output from photovoltaic units, wind turbines, electrical loads, and thermal loads. These four types of scenarios are then connected by a Cartesian link, and the synchronous back-substitution elimination technique is used for further scenario reduction, ultimately yielding Y groups of typical scenarios of combined wind and solar load output and their corresponding probabilities; where Y is the number of typical scenarios of combined wind and solar load output. The objective function value for each group of combined scenarios can then be calculated, and these values ​​are summed for system analysis.

[0096] Step 4: Build a multi-bar optimization model

[0097] Given the potential limitations of historical wind and solar load data, typical scenarios obtained through scene clustering may not fully reflect actual fluctuations. Therefore, a distributed fuzzy set optimization method is introduced. This method uses typical scenarios as input and defines a distributed fuzzy set through comprehensive norm constraints, thereby limiting the fluctuation range of the probability distribution, reducing the impact of bias on the optimization results, and balancing robustness and economy in the optimization process.

[0098] A two-stage bibloc bar optimization model is constructed. The first stage lays the foundation for the normal operation of the system and performs preliminary cost optimization, minimizing fixed costs. Its objective function comprehensively considers various costs, such as operating costs, compensation costs, and variables related to uncertainty. The second stage primarily focuses on adjusting decisions for uncertain scenarios. Based on the decisions made in the first stage and the values ​​of the uncertainty variables, it calculates the system's costs or benefits under different uncertainty scenarios and feeds this back to the first stage to adjust its decision variables, minimizing the expected adjustment cost in the worst-case scenario. The two-stage sub-Bruker optimization model matrix is ​​represented as follows:

[0099]

[0100]

[0101]

[0102]

[0103]

[0104] In the formula, x is the variable for the first stage; y b are variables for the second stage; m and n are cost coefficients corresponding to variables for the first and second stages, respectively. For the scene Probability of occurrence; for A set that satisfies the condition; The scenario probability distribution value that maximizes the objective function; The unit deviation penalty coefficient; For the scene The actual output deviation of wind power and photovoltaic power and the load fluctuation vector; Number of scene clusters; To optimize the matrix or vector in the model constraints that matches the variables; For all constraints related to the variables in the first stage; Equation For the relevant constraints of the variables in the second stage; Equation Japanese style This is a two-stage variable coupling constraint.

[0105] Step 5: Based on the probability distribution of typical scenarios constrained by the comprehensive norm, obtain the fuzzy set of uncertain probability distributions. The probability value of each discrete scenario can belong to any feasible region. To balance robustness and economy during the optimization process and ensure the actual probability distribution... To ensure the model fluctuates within a reasonable range and more closely approximates actual data characteristics, a comprehensive norm is introduced as a constraint to limit the fluctuation range of the probability distribution of wind and light scenarios and loads, allowing the optimization model to remain robust even under the worst-case probability distribution. Specifically, the comprehensive norm includes the 1-norm and the ∞-norm; by setting constraints based on the comprehensive norm, a fuzzy set of distributions containing multiple possible probability distributions is defined. This fuzzy set reflects the distributional uncertainty of the perceived uncertainty factors; that is, although the probability of each scenario cannot be precisely determined, the reasonable fluctuation range of the probability distribution can be limited by the comprehensive norm constraint, avoiding scenario bias caused by the limitations of historical data. During the optimization process, the model balances economy and robustness by minimizing the expected cost under the worst-case distribution, reducing the impact of extreme scenarios on system operating costs. This method allows the optimization model to consider a wider range of uncertainties, enhancing its adaptability to uncertainty. If the norm is considered alone... or This could lead to extreme or one-sided probability fluctuations. Therefore, a comprehensive norm constraint is constructed, whose distribution is a fuzzy set. The following constraints must be satisfied:

[0106]

[0107] In the formula, Typical scenario The reference probability distribution value; The constraint radius is the allowable probability deviation under the corresponding norm constraint. This represents the number of typical scenarios.

[0108] The deviation of the scenario probability satisfies the following probability guarantee:

[0109]

[0110] In the formula: This represents the probability that an event is true. The number of historical scenes.

[0111] Let the right side of the above equation be equal to... and Then it transforms into the formula:

[0112]

[0113] Where α1 and α ∞ This represents the confidence level that the probability distribution needs to satisfy.

[0114] Step 6: Iteratively solve the two-stage sub-Bruker optimization model to obtain the optimal integrated energy dispatch scheme.

[0115] like Figure 3 As shown, for the constructed The two-stage decomposition model matrix is ​​decomposed into two parts: the main problem and the subproblems. A column and constraint generation (C&CG)-based decomposition algorithm is then used to solve the subproblems. The C&CG algorithm iterates through the main problem and subproblems, gradually tightening the upper and lower bounds until an optimal solution meeting certain accuracy requirements is obtained, at which point the iteration stops.

[0116] The objective function L of the main problem can be expressed in matrix form as follows:

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123] In the formula, Given a threshold; This represents the total number of model iterations.

[0124] The subproblem has a max-min two-level structure, given the variables in the first stage. The objective function is then solved as follows:

[0125]

[0126] In the subproblem, the discrete scenario probability values ​​are uncorrelated with other variables. Therefore, the subproblem can be broken down into two steps: first, find the inner minimum value. Then solve the outer max problem U. The specific solution is shown in the following equation:

[0127]

[0128]

[0129] like Figure 3 As shown, the specific process of iteratively solving the two-stage Bruker optimization model is as follows:

[0130] (1) Initialize the number of iterations of the uncertain variables i = 1; set the convergence accuracy ε, upper bound UB = ∞, lower bound LB = 0; set the occurrence probability of each typical scenario to its reference probability distribution value.

[0131] (2) By solving the main problem, the variables of the first stage are obtained. and optimal solution According to the optimal solution Update lower bound value ;

[0132] (3) Based on the variables of the first stage Solve the subproblems to obtain the probability of the scenario occurring. and optimal solution According to the optimal solution Update upper bound value ;

[0133] (4) Perform a convergence check. If UB-LB≤ε, stop the iteration and obtain the optimal solution; otherwise, update the main problem. = and add new variables The main problem is then solved again, along with the associated constraints.

[0134] Example 1

[0135] The model was applied to a community-level integrated energy system. Time-of-use electricity prices, gas prices, and model parameters for each device are shown in Tables 1-3. Then, comparisons were made between different generation methods for different scenarios, different data volumes, different norms, and other uncertainty methods, and the overall cost of joint demand response was analyzed.

[0136] Table 1. Energy and natural gas prices from the upstream power grid

[0137]

[0138] Table 2 Energy Conversion Equipment Parameter Settings

[0139] Device name Energy conversion efficiency Capacity / kW CHP 0.92 300 EB 0.9 250

[0140] Table 3 Energy Storage Device Parameter Settings

[0141] Device name Capacity lower bound constraint Capacity upper bound constraint Capacity / kW Charging / discharging efficiency Electric storage device 0.1 0.9 400 0.95 Thermal storage device 0.1 0.9 300 0.95

[0142] (1) Comparison of different scene generation methods

[0143] 1000 initial scenarios of wind and solar load were generated using MC sampling and Latin hypercube sampling (LHS) based on different distribution functions, respectively. K-means clustering was used to reduce these scenarios to 20 scenarios with different probabilities of individual output from photovoltaic units, wind turbines, electrical loads, and thermal loads. These four scenarios were then connected by a Cartesian link, and synchronous back-substitution elimination technology was used for further scenario reduction, ultimately resulting in four typical scenarios of combined wind and solar load output, such as... Figure 4 and Figure 5 As shown.

[0144] from Figure 4 The wind, solar, and electricity / heat load scenario diagrams show that the data for each scenario is relatively dispersed across different times and scenario numbers. Taking photovoltaic output as an example, the different colored curves representing different scenarios show significant differences in values ​​at different times, covering a wide range. This dispersion indicates that the scenarios generated by Monte Carlo sampling can capture a wider range of possible situations, reflecting more uncertainty and variability. Since the changes in these parameters are diverse and random in actual operation, this method has advantages in simulating complex combined wind and solar output and load conditions. The output value ranges also differ between scenarios. Some scenarios have relatively wide output value fluctuation ranges, possibly covering a large span from low to high power, while some scenarios have relatively narrow output value fluctuation ranges, reflecting their relatively more stable output or smaller fluctuation ranges. This also reflects the representativeness of the reduced scenarios in reflecting the diversity of overall scenarios, including both situations with drastic output changes and relatively stable situations. The data curves for each scenario are highly volatile. Taking the electricity load scenario as an example, the curve fluctuates significantly, reflecting the rapid changes and uncertainties in the load at different times. This greater volatility better reflects the characteristic of real-world load changes, and can more realistically simulate the system's operating state under different load fluctuations.

[0145] exist Figure 5Among these, the typical scenario data are relatively concentrated, and the values ​​of the curves for different scenarios are quite similar at certain times. For example, in the wind turbine output scenario diagram, some curves almost overlap at certain times, indicating that the scenarios generated by LHS lack diversity and may not fully cover all possible operating conditions, failing to adequately reflect some extreme or special situations. Furthermore, the data curves are relatively smooth with low volatility. For example, in the electrical load scenario diagram, the curve changes relatively gently, which may not accurately reflect the rapid changes in electrical load during actual operation. This is problematic for studies that need to consider the impact of rapid load fluctuations on the system. Figure 5 The data may not provide sufficiently accurate information.

[0146] Five schemes were designed as shown in Table 4, keeping other factors constant. In Scheme 1, the photovoltaic output scenario was generated using the LHS method, while the wind turbine output, electrical load, and thermal load scenarios were all generated using the MC method. In Scheme 2, the photovoltaic output scenario was generated using the MC method, the wind turbine output scenario using the LHS method, and the electrical load and thermal load scenarios using the MC method. In Scheme 3, the photovoltaic output, wind turbine output, and thermal load scenarios were generated using the MC method, while the electrical load scenario was generated using the LHS method. In Scheme 4, the photovoltaic output, wind turbine output, and electrical load scenarios were generated using the MC method, while the thermal load scenario was generated using the LHS method. In Scheme 5, the photovoltaic output, wind turbine output, electrical load, and thermal load scenarios were all generated using the MC method. By calculating the sum of the cost optimization results for all typical combined scenarios in the above schemes, a scenario generation method with better optimization effect can be explored. The calculation results are shown in Table 5.

[0147] Table 4 Scenario Design

[0148] Photovoltaic output scene generation method Fan output scene generation method Electric load scene generation method Thermal load scene generation method Scheme 1 LHS MC MC MC Scheme 2 MC LHS MC MC Scheme 3 MC MC LHS MC Scheme 4 MC MC MC LHS Scheme 5 MC MC MC MC

[0149] Table 5 Cost optimization results for each scenario

[0150] Scene scheme Cost optimization result (unit: yuan) Scheme 1 11812.5782 Scheme 2 12290.1790 Scheme 3 12980.7465 Scheme 4 12113.1033 Scheme 5 11812.5713

[0151] As shown in Table 5, with other parameters remaining constant, the cost optimization result of Scheme 5 is 11812.5713 yuan, which is the lowest among the five schemes. Scheme 1 has a cost of 11812.5782 yuan, which is close to Scheme 5, but Scheme 5 has a lower cost. Scheme 2 has a cost of 12290.1790 yuan, Scheme 3 has a cost as high as 12980.7465 yuan, and Scheme 4 has a cost of 12113.1033 yuan, all of which are significantly higher than Scheme 5. This indicates that Scheme 5 has a significant advantage in cost control, can further reduce the conservatism of the obtained results, achieve the optimization goal at a lower cost, help to formulate a more accurate scheduling plan, save some unnecessary additional expenses, and is more attractive from an economic perspective. On the other hand, LHS may introduce additional sampling bias in the optimization solution process, making the convergence speed unstable, which in turn affects the optimization computation efficiency and optimization results.

[0152] (2) Comparison of data volume generated in different scenarios

[0153] To highlight the impact of the size of the N initial scene data generated using MC on the overall cost, under the premise that other factors remain unchanged, according to Figure 2 The method described in the process is based on generating 200, 500, 1000, and 1500 sets of wind and solar power output and load scenario data, respectively, and the optimization results are shown in Table 6.

[0154] Table 6 Comparison of results with different numbers of data points for different wind and solar power output and load scenarios

[0155] Number of wind-solar output and load scenes Comprehensive cost / yuan 200 11892.3943 500 11873.4966 1000 11812.5713 1500 11779.3077

[0156] As shown in Table 6, the overall cost calculated by the system shows a significant downward trend as the number of initial scenarios generated by MC gradually increases. This is because the deviation between the probability distribution constructed based on these scenarios and the actual probability distribution gradually decreases as the number of scenarios increases. In the optimization process of an integrated energy system, the accuracy of the probability distribution plays a crucial role in decision-making. When the constructed probability distribution more accurately approximates the real situation, the information relied upon for the optimization process becomes more reliable, thus effectively reducing the conservatism of the optimization results. This means that the system can allocate resources more rationally during operation, avoiding unnecessary cost increases caused by overly conservative decisions, thereby achieving a reduction in overall cost.

[0157] (3) Comparison of different norms

[0158] To illustrate the difference between comprehensive norm constraints and those considering only 1-norm or ∞-norm constraints, when and The calculation results of the single norm are shown in Tables 7 and 8, respectively, when the values ​​are 0.6 and 0.99. The comparison results show that using the comprehensive norm method can reduce the conservatism introduced by the optimization, compared to considering only the 1-norm or... - Norm constraints are more economical, which suggests that comprehensive norm constraints are less conservative than those that only consider single norm constraints.

[0159] Table 7 Comparison of results between the composite norm and the 1-norm

[0160]

[0161] Table 8. Synthetic Norm and - Norm results comparison

[0162]

[0163] (4) Comparison of other uncertainty methods

[0164] The DRO (Differentiated Robust Optimization) algorithm proposed in this invention is compared with SO (Stochastic Optimization) and RO (Robust Optimization), and the comparison results are shown in Table 9.

[0165] Table 9 System operating costs under different uncertainty optimization methods

[0166] Method Comprehensive cost / yuan Wind-solar light abandoned amount / % SO 10213.8842 0.00 RO 14742.3211 0.00 DRO 11812.5713 0.00

[0167] Table 9 shows that the overall cost obtained using the RO optimization model is the highest, while the overall cost obtained using the SO optimization model is the lowest, and the overall cost obtained using the DRO model falls in between. The stochastic optimization model, due to considering the uncertainty of the distribution parameters of the uncertain variables, has low conservatism. The robust optimization model, on the other hand, does not consider the distribution parameters, resulting in overly conservative results. However, the sub-robust optimization model considers the distribution under the worst-case scenario, balancing conservatism and economy, and achieving a more ideal result.

[0168] (5) Joint demand response comprehensive cost analysis

[0169] right Figure 4 A comprehensive cost analysis of joint demand response is conducted in typical scenario 1, such as... Figure 7 As shown, the load transfer, load shifting, and load reduction in typical scenario 1 mainly occur between 10:00-13:00, 14:00-15:00, and 17:00-21:00, mostly during peak electricity price periods, alleviating peak electricity demand and playing a role in peak shaving and valley filling. Furthermore, combined with... Figure 6It is known that the battery is always charged during the off-peak electricity price period and discharged during the peak electricity price period throughout the entire dispatch period, thus reducing the system operating cost by utilizing the difference between peak and off-peak time-of-use electricity prices; the thermal storage tank also always charges during the off-peak heat load period and releases heat during the peak heat load period, assisting the electric boiler and cogeneration unit in meeting the heat demand. To further explore the impact of energy storage equipment on the two-stage robust optimization model of joint demand response, keeping other factors constant, the experimental scenario schemes shown in Table 10 were set up:

[0170] Table 10 Scene Design

[0171] Scheme 1 Scheme 2 Scheme 3 Scheme 4 Energy storage configuration No energy storage Electric storage+thermal storage Electric storage Thermal storage

[0172] The combined demand response electrical and thermal optimization load results for four scenario solutions with different energy storage configurations are as follows: Figure 6 As shown.

[0173] from Figure 6 The curves for the four scenarios show that the impact of different energy storage configurations on final cost optimization is primarily significant during peak load periods. Peak load periods are further influenced by peak-hour electricity prices, necessitating increased electricity purchases or micro-turbine power supply to meet users' electricity and heat load demands, thus raising costs. The final optimized costs for each scenario are: RMB 12850.1467, RMB 11812.5713, RMB 12624.5927, and RMB 12012.1309, respectively. This demonstrates that configuring appropriate energy storage devices can play a positive role in reducing system costs.

[0174] from Figure 7 As shown in (a), typical scenario 1 primarily uses photovoltaic wind turbines and micro-gas turbines for power generation, and purchases large quantities of electricity during the off-peak hours (05:00-08:00). Under the premise of optimizing its own power load dispatch, the system can sell surplus electricity generated by the photovoltaic wind turbines to external users based on real-time power demand; the energy storage equipment mainly charges during off-peak hours and discharges during peak load periods and peak-hour pricing. Figure 7 As shown in (b), in typical scenario 1, CHP (converted natural gas) and electric boilers are mainly used for heating. The electric boiler converts electrical energy into heat energy to meet the heat load demand. The thermal energy storage equipment stores heat from 06:00 to 08:00 and releases heat when the output of the heating equipment is low or the heat load is relatively high. The addition of the thermal storage equipment effectively reduces the processing capacity of the heat recovery system, thus further reducing the system operating cost during the output periods of the thermal storage equipment. During the periods of 17:00-18:00, 20:00-21:00, and 22:00-23:00, the output of the thermal storage tank and electric boiler equipment reduces the output of the micro-turbine, allowing the system to absorb more wind power output during these periods.

[0175] This invention employs a combination of Monte Carlo sampling based on multiple probability distribution functions and K-means clustering algorithm to construct typical scenarios to describe the uncertainty of wind and solar load. Monte Carlo sampling, through completely random sampling, can capture extreme scenarios, such as extremely high or low wind and solar output and load demand, to test the system's performance under worst-case conditions, ensuring the reliability of the scheduling scheme, enhancing model robustness, and reducing the high-cost risks in actual operation. A two-stage bibliometric robust optimization model is used to handle the uncertainty of wind and solar load in the system, achieving a balance between economy and robustness during the optimization scheduling process, and flexibly adjusting the energy supply and demand relationship by introducing a load demand response mechanism. The first stage of the model determines a relatively conservative but feasible decision scheme for the normal operation of the system to cope with possible uncertainties; the second stage is mainly responsible for calculating the system's costs or benefits under different uncertainty scenarios based on the decision results of the first stage and the values ​​of uncertainty variables, and feeding this back to the first stage to adjust the decision variables of the first stage. Simultaneously, to ensure the actual probability distribution... Within a reasonable range and closer to the characteristics of actual data, the 1-norm and ∞-norm are introduced as constraints to limit the probability distribution values ​​of each scenario. This allows the optimization model to consider a wider range of uncertainties and enhances the model's adaptability to uncertainty.

[0176] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A comprehensive energy dispatching method considering the uncertainty of wind and solar loads, characterized in that: Includes the following steps: Using the start-up and shutdown costs of micro-gas turbines and the total operating cost of the system as optimization objectives, an objective function and constraints are constructed for a comprehensive energy system model encompassing energy supply equipment, energy conversion equipment, and energy storage equipment. Probability distribution models for photovoltaic (PV) output, wind turbine output, and electrical and thermal loads are constructed separately. Random sampling is performed based on different probability distribution models and constraints to obtain scenarios where PV, wind turbine, electrical load, and thermal load output individually have different probabilities, thus acquiring a typical set of scenarios for combined wind and solar load output and their corresponding probabilities. A sub-Bruker optimization model is constructed, and the probability distribution of typical scenarios is constrained to obtain a fuzzy set of uncertain probability distributions. The sub-Bruker optimization model is decomposed into a main problem and sub-problems. The main problem and sub-problems are iterated continuously using a column- and constraint-based decomposition algorithm until the termination condition is met, resulting in the optimal comprehensive energy dispatch scheme. The objective function of the constructed integrated energy system model is as follows: ； In the formula, Cost of starting and stopping the micro gas turbine; This refers to the start-stop cost coefficient for micro gas turbines. for The state variable of the micro gas turbine at any given moment takes the value of 0 or 1; The total operating cost of the system; The cost of purchasing electricity from the power grid; Cost of purchasing gas online; For energy storage costs; Cost of wind curtailment; Cost of abandoning light; For revenue from electricity sales; Cost of responding to demand; The specific process for obtaining a typical scenario set of combined wind and solar power output is as follows: The K-means clustering algorithm is used to compress the sampling scenarios, extract typical scenario sets of photovoltaic, wind turbine, electric load and heat load output individually, and connect these different types of scenarios with Cartesian connection. The synchronous back-substitution elimination technique is used to further reduce the scenarios, resulting in a typical scenario set composed of multiple groups of typical scenarios of wind and solar load combined output. The specific process of iteratively solving the two-stage Bruker optimization model is as follows: Initialize the number of iterations for the uncertain variables; set the convergence precision, upper bound, and lower bound respectively; set the probability of occurrence of each typical scenario to its reference probability distribution value; Solve the main problem to obtain the first-stage variables and the optimal solution, and update the lower bound value according to the optimal solution corresponding to the main problem; Solve the subproblem based on the first-stage variables to obtain the second-stage variables and the optimal solution. Update the upper bound value according to the optimal solution corresponding to the subproblem. If the difference between the upper bound value and the lower bound value is greater than the convergence accuracy, add the second-stage variables and related constraints to the main problem and solve the main problem again until the difference between the upper bound value and the lower bound value is less than or equal to the convergence accuracy. Stop the iteration and obtain the optimal solution.

2. The integrated energy dispatching method considering the uncertainty of wind and solar loads according to claim 1, characterized in that: The probability distribution model for photovoltaic power output is constructed based on the beta distribution; the probability distribution model for wind turbine power output is constructed based on the Weiber distribution; and the probability distribution model for electric heating load is constructed based on the normal distribution.

3. The integrated energy dispatching method considering the uncertainty of wind and solar loads according to claim 1, characterized in that: The objective function corresponding to the total system operating cost incorporates demand response cost; demand response cost includes compensation costs for shiftable loads participating in demand response, compensation costs for transferable loads participating in demand response, and compensation costs for loads that can be reduced participating in demand response.

4. The integrated energy dispatching method considering the uncertainty of wind and solar loads according to claim 1, characterized in that: The number of scene clusters in the K-means clustering algorithm is obtained using the silhouette coefficient method.

5. The integrated energy dispatching method considering the uncertainty of wind and solar loads according to claim 1, characterized in that: The probability distribution of the typical scenarios described is constrained by a comprehensive norm.

6. The integrated energy dispatching method considering the uncertainty of wind and solar loads according to claim 1, characterized in that: The objective function of the main problem includes operating costs, compensation costs, and variables related to uncertainty.

7. A comprehensive energy dispatching system considering the uncertainty of wind and solar loads, comprising energy supply equipment, energy conversion equipment, energy storage equipment, and a control module; the energy supply equipment includes photovoltaic units and wind turbine units; the energy conversion equipment includes electric boilers and combined heat and power units; the energy storage equipment includes batteries and thermal storage tanks; characterized in that: The control module is used to execute the integrated energy dispatching method of claim 1, and to control the energy supply equipment, energy conversion equipment and energy storage equipment based on the generated optimal integrated energy dispatching scheme.

Images