Integrated energy scheduling method and system considering demand response uncertainty

Abstract

The present application relates to the technical field of multi-energy scheduling, in particular to a comprehensive energy scheduling method and system considering demand response uncertainty, which provides for obtaining electric, gas and heat load data and renewable energy prediction output data in a comprehensive energy system, based on user response triggering boundary effect, cross-period load rebound effect and heterogeneous energy physical coupling constraints, a comprehensive demand response uncertainty model considering space-time coupling characteristics is constructed; based on the model, a double-layer coordinated control framework of source-side controller advanced decision and load-side execution unit tracking response is constructed; according to historical data, load tracking deviation is estimated and instruction correction amount is calculated, and the optimal control instruction sequence is obtained through iterative updating, and is converted into source-side unit output instruction and load-side load control instruction to execute scheduling. The present application realizes the collaborative optimization operation of source and load under the demand response uncertainty environment, improves the renewable energy consumption rate, suppresses the load peak-valley difference and guarantees the user comfort.

Description

Technical Field

[0001] This invention relates to the field of multi-energy dispatching technology, specifically to a comprehensive energy dispatching method and system that takes into account the uncertainty of demand response. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Integrated Energy Systems (IES) provide an effective way to enhance the absorption of renewable energy by coupling electricity, gas, and heat networks. Integrated Demand Response (IDR) allows users to flexibly adjust their multi-energy usage behavior based on dynamic incentive signals, and is a key means of tapping the potential for demand-side regulation.

[0004] In actual operation, Integrated Demand Response (IDR) exhibits strong uncertainty; For example, users exhibit response-triggered boundary effects to excitation signals, resulting in nonlinear and discontinuous changes in their response behavior, which are difficult to accurately characterize using traditional elastic coefficients. For example, in order to maintain energy comfort, users will experience load rebound across time periods, that is, load reduction in the previous stage often triggers compensatory growth in the subsequent stage. For example, heterogeneous energy sources such as electricity, gas, and heat are physically coupled and constrained on the demand side. Adjustments made by users to one type of energy source will indirectly affect the available adjustment space for other energy sources.

[0005] These uncertainties combine to make it difficult for IES to accurately predict the true response capabilities of load aggregators (LAs).

[0006] Meanwhile, IES typically aim to minimize the supply-demand imbalance on both sides, while LA focuses on minimizing its energy costs; thus, the goals of IES and LA differ. Traditional centralized dispatching, on the other hand, assumes that both sides have the same goal, which can easily lead to strategy deviations in an environment of IDR uncertainty. That is, if the control commands formulated by IES deviate from the actual response boundary of LA, they may not only fail to smooth out load fluctuations but also exacerbate the supply-demand imbalance. Summary of the Invention

[0007] This invention provides a comprehensive energy dispatching method and system that considers the uncertainty of demand response. The IES and LA are modeled as a two-level coordinated control model (mathematically equivalent to a Stackelberg game). The IES acts as the source-side controller to generate dynamic control commands, and the LA acts as the load-side execution unit to track the response. The convergence state of the control commands is solved iteratively by LSTM-PID, and a spatiotemporal dual-dimensional IDR uncertainty quantification model is introduced.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of the present invention provides a comprehensive energy dispatching method that takes into account demand response uncertainty, comprising the following steps: Obtain load data for electricity, gas, and heat in the integrated energy system, as well as predicted output data for renewable energy. Based on the user response to the excitation signal triggering boundary effect, cross-period load rebound effect, and physical coupling constraints between heterogeneous energy sources, construct an integrated demand response uncertainty model. The integrated demand response uncertainty model includes a spatial multi-energy coupling sub-model for quantifying the substitution elasticity of heterogeneous energy sources, and a time dynamic response sub-model for characterizing the user's cross-period load compensation behavior. Based on the comprehensive demand response uncertainty model, the integrated energy system is used as the source-side controller and the load aggregator is used as the load-side execution unit. A two-layer coordinated control framework is constructed, in which the source-side controller makes advance decisions and the load-side execution unit tracks the response. The source-side controller generates a dynamic control command sequence to maximize the system's supply and demand regulation efficiency, and the load-side execution unit tracks the response plan by minimizing the local power tracking deviation. The load tracking deviation is estimated based on historical user response data, and the control command correction amount is calculated based on the estimated deviation. The control command sequence is iteratively updated until the control command convergence state is reached, and the optimal control command sequence is obtained. The optimal control command sequence is transformed into source-side unit output commands and load-side load control commands, driving physical equipment to perform scheduling actions and achieving coordinated and optimized operation of both the source and load sides.

[0009] Furthermore, the spatial multi-energy coupling sub-model is constructed by decomposing the coupling effect between heterogeneous energy sources into two dimensions: response willingness and response capability. Response willingness represents the user's cognitive sensitivity to the relative differences in excitation signals between heterogeneous energy sources, while response capability represents the physical adjustment boundary constrained by real-time energy consumption status.

[0010] Furthermore, the responsiveness is as follows: ; In the formula, and ( ) represent the upper and lower limits of the baseline load for user k energy sources during the scheduling period, respectively. This represents the baseline load of k energy sources at time t. and These represent the upward and downward adjustment capabilities of k types of load at time t, respectively. This represents the normalized adjustment capability of k energy sources at time t.

[0011] Furthermore, the willingness to respond is shown in the following formula: ; In the formula, This indicates the user's willingness to switch from electricity to gas at time t. , These represent the maximum amplitude of the electrical excitation and pneumatic excitation signals within a scheduling cycle, respectively.

[0012] Furthermore, the time-dynamic response sub-model is shown in the following equation: ; In the formula, ( ) represents a traditional user comfort model for k types of energy. This represents the user's overall comfort level with respect to the k types of energy at time t. The sensitivity coefficient of the excitation signal. and These represent the real-time excitation signal amplitude and the average reference value of the excitation signal within the scheduling period, respectively. These are the status parameters of IDR.

[0013] Furthermore, the time-dynamic response sub-model also includes a trigger response mechanism based on the accumulation of comfort deviations, specifically: setting a trigger boundary for the user's comfort response to the k-th energy source. When the user's overall comfort level falls below the trigger threshold at time t due to over-responding to the stimulus signal, the user actively adjusts the dynamic response at time t+1 by reducing the intensity of the response to the stimulus signal. This will bring overall comfort back to a feasible level.

[0014] Furthermore, the objective function of the source-side controller is shown in the following equation: ; ; ; ; ; In the formula, The total amount of energy output by the IES to the load aggregator is regulated; This represents the flexible load dispatch adjustment amount issued by IES to the load aggregator; This represents the discounted value of the amount of energy injected into the IES from external energy suppliers; This represents the physical loss calculation of the internal devices of the IES; T is the scheduling period. ( () represent the amplitudes of the control commands provided by IES to LA for the k types of energy at time t. For the corresponding energy output power, , , , These represent the unit cost of electricity, unit cost of gas, amount of electricity obtained by IES from external energy suppliers at time t, and amount of gas obtained, respectively. / / / These represent the conversion factors for the operating status of various types of coupled units. / / / These represent the input power of the corresponding unit at time t.

[0015] Furthermore, the objective function of the load-side execution unit is shown in the following equation: ; ; ; In the formula, This refers to the energy loss caused by the physical characteristics of the energy storage system during charging and discharging. The equivalent energy deviation caused by differences in conversion efficiency when users switch energy types; and All models employ quadratic functions to characterize the saturation properties of physical regulation capabilities; , , and These represent the conversion factors for physical losses during the charging and discharging of electrical energy storage. / These represent the energy storage charge / discharge power at time t; and These represent the conversion efficiency deviation factor for user substitution.

[0016] Furthermore, the control command convergence state is specifically defined as a state in which neither the source-side controller nor the load-side execution unit can obtain a better regulation effect by unilaterally changing the control strategy, which is mathematically equivalent to Nash equilibrium. During the iterative update process, the incremental regulation efficiency caused by the adjustment of the control signal is taken as the convergence target to approach zero.

[0017] A second aspect of the present invention provides an integrated energy dispatching system that takes into account demand response uncertainty, comprising: The model building module is configured to: acquire load data of electricity, gas and heat in the integrated energy system and forecast output data of renewable energy; construct an integrated demand response uncertainty model based on the user's response to the incentive signal triggering boundary effect, cross-period load rebound effect and physical coupling constraints between heterogeneous energy sources; the integrated demand response uncertainty model includes a spatial multi-energy coupling sub-model for quantifying the substitution elasticity of heterogeneous energy sources, and a time dynamic response sub-model for characterizing the user's cross-period load compensation behavior; The source-load dual-layer coordinated control module is configured as follows: based on the comprehensive demand response uncertainty model, the integrated energy system is used as the source-side controller and the load aggregator is used as the load-side execution unit to construct a dual-layer coordinated control framework in which the source-side controller makes advance decisions and the load-side execution unit tracks the response. Among them, the source-side controller generates a dynamic control command sequence to maximize the system's supply and demand regulation efficiency, and the load-side execution unit tracks the response plan by minimizing the local power tracking deviation. The solution module is configured to: estimate the load tracking deviation based on historical user response data, calculate the control command correction amount based on the estimated deviation, and iteratively update the control command sequence until the control command convergence state is reached to obtain the optimal control command sequence; The scheduling output module is configured to convert the optimal control command sequence into source-side unit output commands and load-side load control commands, drive physical equipment to perform scheduling actions, and achieve coordinated and optimized operation of both the source and load sides.

[0018] Compared with existing technologies, one or more of the above technical solutions have the following beneficial effects: By constructing a comprehensive demand response uncertainty model, complex behaviors such as user response trigger boundaries, cross-period load rebounds, and heterogeneous energy physical coupling are transformed into quantifiable mathematical constraints, overcoming the technical shortcomings of traditional fixed elasticity coefficient models in characterizing user response inaccurately. By establishing a two-layer coordinated control framework with source-side controllers making advance decisions and load-side execution units tracking responses, the conflicting adjustment demands between the integrated energy system and load aggregators are incorporated into a closed-loop iterative mechanism. This avoids the strategy failure caused by centralized optimization assuming consistent goals between the two sides. By combining data-driven approaches with closed-loop correction, the convergence state of control commands is approximated, improving the robustness of the scheduling strategy under uncertain demand response environments. Experimental results demonstrate that this scheme increases wind and solar power absorption and reduces the peak-valley difference between electricity and gas loads. Attached Figure Description

[0019] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0020] Figure 1This is a schematic diagram of the IES operating framework provided in one or more embodiments of the present invention; Figure 2 This is a schematic diagram of the IES-LA two-layer coordinated control model provided in one or more embodiments of the present invention; Figure 3 This is a schematic diagram of the game-solving process provided by one or more embodiments of the present invention; Figure 4 This is a schematic diagram of basic load data and wind and solar power forecasting provided by one or more embodiments of the present invention; Figure 5 This is a schematic diagram of multi-scenario load comparison provided by one or more embodiments of the present invention, wherein (a) is a schematic diagram of multi-scenario electrical load comparison, (b) is a schematic diagram of multi-scenario electrical load comparison, and (c) is a schematic diagram of multi-scenario electrical load comparison. Figure 6 This is a scatter diagram of electrical loads before and after coupling provided in one or more embodiments of the present invention; Figure 7 This is a schematic diagram of changes in user energy consumption experience provided by one or more embodiments of the present invention, wherein (a) is a schematic diagram of changes in user electricity consumption experience, (b) is a schematic diagram of changes in user natural gas consumption experience, and (c) is a schematic diagram of changes in user thermal energy consumption experience; Figure 8 This is a schematic diagram of the electrical load device output in "Scenario 5" provided by one or more embodiments of the present invention; Figure 9 This is a schematic diagram of the gas load device output for "Scenario 5" provided in one or more embodiments of the present invention; Figure 10 This is a schematic diagram of the heat load equipment output for "Scenario 5" provided in one or more embodiments of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] Terminology Explanation: An IES (Integrated Energy System) is a regional energy system that integrates multiple energy forms such as electricity, gas, and heat. It includes renewable energy units (wind power, solar power) and multi-energy coupling and conversion equipment such as CHP, P2G, GB, and EB. It is the main energy supplier and game leader in this scheme, responsible for coordinating the production, conversion, and storage of various energy sources within the system, and providing electricity, gas, and heat to the load aggregator (LA).

[0024] LA (Load Aggregator) is an intermediary entity between IES and end users, responsible for aggregating end user load resources and participating in system response.

[0025] IDR (Integrated Demand Response) refers to the flexible adjustment of users' use of various energy sources such as electricity, gas, and heat based on incentive signals, including load reduction, load shifting, and substitution between different energy sources (such as "replacing gas with electricity").

[0026] Two-layer coordinated control is a source-load hierarchical interaction framework. The source-side controller makes decisions first, and the load-side execution unit optimizes its tracking response accordingly. The two sides eventually reach a convergence state of control commands.

[0027] The control command convergence state refers to the state in which neither the source nor the load can achieve a better regulation effect by unilaterally changing the control strategy (mathematically known as Nash equilibrium).

[0028] Examples of uncertainty in Integrated Demand Response (IDR) are as follows: The response trigger boundary effect refers to changes in the user's psychological threshold. For example, if the amplitude of the excitation signal increases by 5%, the user will not respond, but if it increases by 20%, the load will be significantly reduced. The response amount does not change smoothly with the signal.

[0029] The load rebound effect across time periods can occur when users reduce their electricity consumption during periods of high demand (e.g., by raising the air conditioner's set temperature), feel uncomfortable after the change in room temperature, and then resume or even exceed their original electricity consumption levels in the next period, resulting in a load rebound.

[0030] Physical coupling between heterogeneous energy sources: For example, if users reduce electric heating, they must use more gas heating to maintain room temperature. The reduction in electricity actually forces the gas load to increase. In other words, the user's adjustment of electricity will indirectly affect the available adjustment space of gas.

[0031] Excitation signals refer to control commands issued by an integrated energy system to load aggregators to guide them in adjusting their energy consumption behavior. At the physical level, excitation signals are technical parameters used to characterize the real-time scarcity of energy or the system's supply-demand balance. Specific implementation forms include, but are not limited to, the following types: (1) Price signal: The monetary price of a unit of energy is used as the incentive carrier, and the price level reflects the real-time supply and demand tension of energy. It should be noted that when the price signal is used as the physical realization of the incentive signal, its essence is to use price as an economic lever to guide users to adjust their energy consumption behavior, rather than a profit distribution tool in a commercial sense.

[0032] (2) Carbon emission intensity signal: The carbon emission factor per unit of energy is used as the excitation carrier. The signal amplitude reflects the cleanliness of the energy and guides users to give priority to the use of low-carbon energy.

[0033] (3) System frequency deviation signal: The deviation between the real-time frequency of the power grid and the rated frequency is used as the excitation carrier. The signal amplitude directly reflects the power imbalance of the system and guides users to participate in frequency regulation.

[0034] (4) Renewable energy consumption pressure signal: The wind curtailment rate, solar curtailment rate or the deviation between the real-time output and the predicted output of renewable energy is used as the excitation carrier, and the signal amplitude reflects the urgency of the system to flexibly adjust the demand side.

[0035] (5) Comprehensive control command: The composite control quantity formed by weighting and fusing one or more of the above signals is used as a unified command to be issued.

[0036] For scheduling problems with uncertain demand response, existing methods often use fixed elasticity coefficients or linear models to characterize user response and assume that the Integrated Optimization System (IES) and the Load Balancing (LA) objectives are aligned for centralized optimization. This approach ignores the nonlinearity of user response and the cross-period compensation effect, and cannot handle conflicts in the adjustment demands of both sides. In practice, the control commands generated by the IES often deviate from the true response boundary of the LA, leading to a worsening of the peak-valley load difference, a decrease in wind and solar power absorption rates, and even exacerbating the supply-demand imbalance.

[0037] To address this, this scheme explicitly models the conflict between the regulation demands of IES and LA as a two-layer coordinated control model, introduces a spatiotemporal dual-dimensional IDR uncertainty model (spatial multi-energy coupling matrix + temporal comfort compensation mechanism), and adopts an LSTM-PID dynamic command iterative solution strategy to gradually approach the convergence state of the control command in the closed-loop feedback, thereby achieving coordinated optimal scheduling of the source and load sides.

[0038] Specifically, this scheme transforms complex uncertainties into quantifiable mathematical constraints, including a spatial multi-energy coupling sub-model (to solve the problem of heterogeneous energy coupling) and a time dynamic response sub-model (to solve the problem of load rebound). It models IES and LA as a Stackelberg game (leader-follower relationship) and uses an LSTM-PID iterative solution strategy to approximate the convergence state, thereby transforming uncertainty into "compensable terms".

[0039] Stackelberg game, often translated as Stackelberg game or Stackelberg game, is a leader-follower game.

[0040] A comprehensive energy dispatching approach that considers demand response uncertainty includes the following steps: Step S1: Construct a comprehensive demand response uncertainty model that considers spatiotemporal coupling characteristics; Among them, the comprehensive demand response uncertainty model includes a spatial multi-energy coupling sub-model and a time dynamic response sub-model; The spatial multi-energy coupling sub-model quantifies the substitution elasticity and physical constraints among heterogeneous energy sources such as electricity, gas, and heat by constructing a multi-energy coupling matrix. The time-based dynamic response sub-model characterizes users' load rebound behavior across time periods by constructing a trigger response model based on the accumulation of comfort deviations. Step S2: Establish a source-load two-layer coordinated control model; specifically: By using the integrated energy system as the source-side controller and the load aggregator as the load-side execution unit, a two-layer coordinated control framework is constructed, in which the source-side controller makes decisions in advance and the load-side execution unit tracks and responds. The goal of the source-side controller is to generate a dynamic control command sequence with the aim of maximizing the system's supply and demand regulation efficiency. The objective of the load-side execution unit is to minimize the local power tracking deviation and to provide feedback on the tracking response plan. Step S3: Solve the two-level coordinated control model; specifically: use LSTM to estimate the load tracking deviation based on historical user response data, use the PID controller to calculate the control command correction amount, and iteratively update the control command sequence until the control command convergence state is reached. Step S4: Execute scheduling using the control command sequence obtained from the solution; specifically: convert the control command sequence into source-side unit output commands and load-side load control commands, drive physical equipment to perform scheduling actions, and achieve coordinated and optimized operation of both source and load sides.

[0041] The detailed process is described below with reference to the accompanying diagram.

[0042] 1. IDR uncertainty model considering spatiotemporal coupling characteristics.

[0043] The essence of user participation in IDR (Induced Responsive Response) is the behavioral decision-making process they make in response to external stimulus signals. Based on stimulus-response theory in experimental psychology, a user's sensitivity to stimulus signals is not constant but exhibits complex coupling characteristics and a psychological threshold effect. Traditional elasticity coefficients are insufficient to capture users' random response behavior when faced with dynamic stimulus signals. Therefore, this solution, considering various uncertainties in IDR, further considers multi-energy coupling effects and dynamic response characteristics, deeply analyzing the user's dynamic response characteristics from both spatiotemporal dimensions: In the spatial dimension, coupling matrices are used to quantify the substitution elasticity between heterogeneous energy sources, revealing the lateral interaction relationship between multiple energy flows such as electricity, gas, and heat at the same moment; In the time dimension, a dynamic response model based on comfort compensation is constructed to accurately characterize users' cross-period load adjustment behavior in order to maintain an overall energy consumption experience. This method achieves a refined characterization of users' complex energy consumption behavior, providing a reliable basis for subsequent optimized scheduling.

[0044] 1.1 Spatial multi-energy coupling effect.

[0045] There is a deep coupling effect among the heterogeneous energy responses in an integrated energy system. This coupling effect is not a simple linear superposition, but essentially manifests as a relationship of mutual constraint and substitution between heterogeneous energy sources in meeting end-user demand. At the micro level, although electricity, gas, and heat are heterogeneous energy sources with different carrier forms, they all follow the principle of energy conservation when meeting end-user energy needs.

[0046] Taking the heating scenario as an example, there is a heat substitution effect between electric heating and gas heating: when the incentive signal guides users to significantly reduce their electricity consumption, the heat gap caused by the reduction in electricity consumption must be filled by natural gas due to the conservation of heat energy and users' rigid demand for thermal comfort.

[0047] This means that users' peak-shaving behavior on electrical load at specific times will directly compress their space for reducing gas load, and may even force them to compensate by increasing gas energy consumption. This creates a multi-energy response feasible domain on the demand side that dynamically changes with the operating status. If this micro-coupling mechanism based on physical conversion characteristics is ignored, and the independent response optimization of a single energy source is simply pursued, the scheduling strategy will deviate from the actual physical boundary, making it impossible to achieve true multi-energy synergy and mutual assistance.

[0048] To achieve a refined quantification of this coupling effect, this scheme deconstructs it into two dimensions: willingness to respond and response capability. Willingness to respond characterizes a user's cognitive sensitivity to the relative differences in excitation signals between heterogeneous energy sources. It determines the type and tendency of energy coupling substitution (e.g., "electricity instead of gas" or "gas instead of electricity"). Response capability characterizes the physical adjustment boundary constrained by real-time energy consumption status, specifically reflecting the user's maximum responsive margin at a given moment. The significance of the coupling effect is positively correlated with both the user's willingness to respond and their response capability.

[0049] Based on this, this solution constructs the following response capability quantification model: (1) In the formula, and ( ) represent the baseline load upper and lower limits of user k energy sources during the scheduling period, respectively. This represents the baseline load of k energy sources at time t. and These represent the upward and downward adjustment capabilities of k types of load at time t, respectively. This represents the normalized adjustment capability of k energy sources at time t. When the load is high, it indicates that the user has a strong ability to absorb the load; conversely, when the load is low, it indicates that the user has a strong ability to reduce the load.

[0050] Taking electrical coupling as an example, the response intention model is shown in the following equation: (2) In the formula, This indicates the user's willingness to switch from electricity to gas at time t. , These represent the maximum values ​​of electrical excitation and pneumatic excitation within a scheduling cycle, respectively. When... When the excitation is at a certain time, it means that compared to pneumatic excitation, electrical excitation puts greater energy pressure on users, and users tend to use pneumatic energy instead of electrical energy; conversely, users tend to use electrical energy instead of pneumatic energy.

[0051] To accurately describe a user's actual regulation potential in multi-energy coupling, a physical constraint model based on the product of heterogeneous energy regulation capabilities is first constructed. Specifically, a user's ability to substitute gas for electricity depends on the combined effect of the ability to reduce electrical load and the ability to increase gas load, as expressed below: (3) However, physical conditions alone are insufficient to trigger coupled behavior; the driving force of differences in excitation signals must also be considered. Therefore, the response intention is utilized. As an activation switch, a coupling coefficient is constructed as shown in equation (4). This ensures that the coupling behavior is triggered only when the user's willingness to respond aligns with their responsiveness. (4) In the formula, This represents the electro-gas coupling coefficient at time t. The formula can be explained as follows: when the user's desired response is to replace electricity with gas, i.e. hour, It will automatically become 0 if the physical conditions are not met, because to achieve gas-to-electricity replacement, there needs to be a significant reduction in electrical load and a significant capacity to absorb gas load. ,at this time and At least one of them is 0. Equation (4) simplifies to And the result is positive; otherwise, the same applies.

[0052] Based on the derived electro-gas-thermal coupling coefficient and the symmetry characteristics of energy interaction, a demand-side multi-energy coupling matrix is ​​constructed to capture the multi-energy coupling effect of users. Simultaneously, to endow the model with adaptive capability when both reduction-type and absorption-type IDRs coexist, diagonal elements of the state parameter correction matrix are introduced. Finally, the comprehensive response model considering multi-energy coupling effects is established as follows: (5) (6) In the formula, and ( The numbers and represent the nonlinear response quantities of different energy sources participating in SLDR (Signaled Demand Response) and ABDR (Alternative Demand Response) at time t, respectively. This represents the incentive-based demand response (IBDR) adjustment amount for k types of energy at time t, taking into account the dynamic response characteristics of users. and Let represent the IDR response of k energy sources at time t before and after considering the multi-energy coupling effect. and The values ​​are the same, and the other coupling coefficients are similar. Let be the state parameter of IDR for k energy sources at time t. If it is a reduction type IDR, its corresponding value is 1; if it is an absorption type IDR, its corresponding value is -1.

[0053] 1.2 Time-based dynamic response characteristics.

[0054] In real life, user participation in IDR is essentially a continuous decision-making process with temporal correlation characteristics, rather than an isolated, memoryless event. To correct the traditional perspective's neglect of this dynamic characteristic, this solution takes a comfort compensation perspective to deeply analyze the dynamic response characteristics of users across time periods.

[0055] This dynamic response characteristic stems from the cumulative and memory effects of users' comfort perception. Specifically, users' energy consumption decisions are driven by both real-time stimulus signals and historical energy consumption patterns. When a user reduces load (e.g., lowers the air conditioning temperature setting) in response to a high-amplitude stimulus signal during a certain period, the loss of comfort does not dissipate instantly but accumulates over time at both the psychological and physiological levels, creating a strong demand for comfort recovery. Once the user's psychological tolerance threshold is reached, the user will adjust the load additionally to compensate for the previous comfort deficit. This mechanism is physically manifested as the influence of human comfort tolerance and mathematically as a nonlinear shift of the load curve on the time axis. By constructing a comfort compensation model that incorporates memory characteristics and quantifying this cross-period coupling behavior, the risk of system power imbalance caused by scheduling strategies neglecting load rebound can be effectively avoided.

[0056] 1.2.1 Improved comfort model.

[0057] Accurately quantifying user comfort is a prerequisite for evaluating the effectiveness of IDR implementation. Given that user participation in demand response is essentially a coordination process seeking a balance between maintaining physical energy consumption experience and the degree of responsiveness to external incentive signals, a simple load-based evaluation cannot fully encompass this complex psychological decision-making mechanism. Therefore, this solution constructs a two-dimensional evaluation system that includes satisfaction with energy consumption methods and perceived deviation from incentive signals, in order to achieve a comprehensive description of overall user comfort.

[0058] Energy usage satisfaction aims to quantify the degree to which load curve reconstruction deviates from users' established lifestyles, characterizing the user's tolerance boundary for changes in physical energy usage behavior. In the basic scenario without IDR (Indexed Energy Management), users plan their energy usage entirely based on personal preferences and daily routines; at this point, their energy usage perfectly matches their psychological expectations, and physical satisfaction reaches its peak. After IDR implementation, users adjust their energy usage based on the amplitude of incentive signals to reduce perceived incentive bias. However, this change in established energy usage habits inevitably leads to a decrease in physical energy usage experience. Traditional assessment models often limit themselves to this physical dimension, neglecting the corrective effect of incentive signals on users' willingness to respond, resulting in biased assessments of response potential.

[0059] A user's sensitivity to the stimulus signal depends not only on the absolute amplitude of the signal, but also on the degree to which the real-time signal deviates from their psychologically expected reference value (i.e., the periodically averaged signal). Therefore, this solution introduces stimulus deviation perception and constructs the following improved comprehensive comfort assessment model: (7) In the formula, ( ) represents a traditional user comfort model for k types of energy. This represents the user's overall comfort level with respect to the k types of energy at time t. This is the excitation signal sensitivity coefficient, used to quantify the user's response strength to deviations in the excitation signal. and These represent the real-time excitation signal amplitude and the average reference value of the excitation signal within the scheduling period, respectively.

[0060] This model introduces state parameters. The product of the excitation signal deviation term and the excitation signal deviation term forms an evaluation mechanism that matches operational reliability indicators with adjustment requirements. When the system is in a reduced IDR state ( And the real-time excitation signal is higher than the reference mean, or the system is in an absorbent IDR state. Furthermore, when the real-time excitation signal is lower than the reference average, the perceived excitation deviation is always positive. This indicates that the excitation signal is aligned with the adjustment demand, and the deviation pressure felt by the user matches the system requirements.

[0061] This consistent perception of signal-level deviation can effectively offset the physical discomfort caused by changes in energy consumption, thereby improving overall comfort. Conversely, when the excitation signal is opposite to the adjustment demand, the perceived excitation deviation is negative, which will exacerbate the user's overall discomfort and reduce overall comfort.

[0062] 1.2.2 Comfort compensation mechanism based on psychological threshold.

[0063] Users' energy consumption decisions exhibit memory effects and cross-time period coupling characteristics. To accurately characterize this dynamic process, this paper uses comprehensive comfort as the core hub connecting decisions in adjacent time periods. Based on the trigger response theory of accumulated comfort deviations, a trigger boundary for the user's comfort response to the k-th energy type is defined. If, at time t, the user's overall comfort level falls below the trigger threshold due to over-responding to the stimulus signal, at time t+1, the user will actively adjust the dynamic response quantity at the cost of reducing the intensity of the response to the stimulus signal. The aim is to bring overall comfort back to an acceptable and feasible range. This dynamic adjustment process based on state feedback can be modeled as a conditional optimization problem with the objective of minimizing comfort deviation.

[0064] (8) In the formula, Let k be the dynamic energy response quantities triggered by the comfort compensation mechanism at time t+1. Given that this adjustment behavior is a user-initiated response based on their own comfort needs, rather than being forced by external commands, this scheme classifies it as an adaptive adjustment component within incentive-based demand response.

[0065] (9) In the formula, This represents the nonlinear response of k energy sources at time t in the IBDR.

[0066] 2IES Framework and Operating Mode.

[0067] This solution establishes a collaborative scheduling framework that includes external energy suppliers, IES (Environmental Engineering Systems), and LA (Local Energy Providers). For example... Figure 1 As shown, the IES serves as the main energy supplier of the system, integrating renewable energy units (PV, WT) and multi-energy coupling equipment (CHP, GB, EB, P2G) to achieve the conversion and complementarity of heterogeneous energy sources. The demand-side LA integrates diverse flexible loads of electricity, gas, and heat, and is equipped with EES and GES, improving the LA's regulation elasticity and response flexibility.

[0068] In renewable energy units, PV stands for photovoltaic power generation, and WT stands for wind power generation; In multi-energy coupling equipment, CHP is a combined heat and power unit, GB is a gas boiler, EB is an electric boiler, and P2G is an electric-to-gas conversion device. In demand-side energy storage (LA), EES stands for electric energy storage systems, such as lithium batteries or pumped hydro storage; GES stands for gas energy storage systems, such as gas storage tanks, which store gas during off-peak hours and release gas during peak hours.

[0069] Both IES and LA are regulatory entities with independent decision-making power, and their regulatory demands and objectives naturally conflict in the supply-demand interaction. Furthermore, given the hierarchical decision-making sequence and non-cooperative game nature of the source-load interaction process, this scheme models it as follows: Figure 2The illustrated IES-LA two-layer coordinated control model is as follows: the IES acts as the source-side controller, and the LA acts as the load-side execution unit. The IES issues control commands to the LA, and the LA feeds back load adjustment amounts to the IES. Mathematically, this model is equivalent to a Stackelberg game model where the IES is the leader and the LA is the follower. The decision variables for the Integrated Energy System (IES) can be: the amplitude of the electricity control command, the amplitude of the gas control command, and the energy dispatch strategy, with the objective of "maximizing the supply and demand regulation efficiency of the IES." The decision variables for the LA can be: the electricity, gas, and heat loads that can be reduced, the electricity and gas loads that can be substituted, and the energy storage output, with the objective of minimizing the local power tracking deviation.

[0070] This model can accurately depict the dynamic constraints between the control command issuer (source-side control command issuer) and the responder (load-side command tracking party). Its two-level decision-making mechanism is highly consistent with the actual power system operation logic, providing the most suitable mathematical paradigm for solving resource optimization allocation problems involving multi-objective competition and cascaded decision-making.

[0071] The specific system framework and game interaction process are as follows: Figure 1 As shown: The Integrated Energy System (IES) plays a dominant role and has the right to make prior decisions. It is committed to maximizing the absorption rate of renewable energy sources such as wind and solar power by optimizing the internal multi-energy flow scheduling, while aiming to maximize the system's supply and demand regulation efficiency, and generating a dynamic control command sequence for the LA. The LA (Local Energy Controller) is in a subordinate position. Upon receiving control commands, it needs to fully consider the multiple uncertainties of IDR (Independent Power Distribution) and its own flexibility and resource adjustability, while ensuring that the energy comfort of end users is within an acceptable range. It aims to minimize local power point tracking deviation and then feeds back the energy consumption strategy. Subsequently, the IES (Environmental Engineering System) uses LSTM (Low-Speed ​​Module) and a PID controller to dynamically update the control command sequence and feeds it back to the LA. Through this closed-loop information exchange, both parties continuously correct their respective strategies until the supply and demand sides reach a convergence state of control commands (mathematical Nash equilibrium). This state means that neither the source nor the load can achieve a better adjustment effect by unilaterally changing the control strategy. At this point, coordinated control reaches a steady state (mathematical game convergence), and the system achieves optimal overall operational reliability and renewable energy absorption rate.

[0072] 3. Source-load interaction model and solution method based on two-layer coordinated control.

[0073] 3.1 IES operating model considering IDR uncertainty.

[0074] This scheme designs a quantitative two-level optimization mathematical model. The upper and lower levels of the model are coupled with control command variables and power variables, which together constitute the following mathematical expression.

[0075] 3.1.1 IES Objective Function.

[0076] As the core manager of the system, IES aims to maximize the efficiency of system supply and demand regulation by optimizing internal unit output and external control strategies. (10) In the formula, The total amount of energy output from IES to LA is regulated; This indicates the amount of flexible load dispatch adjustment issued by IES to LA; This represents the discounted value of the amount of energy injected into the IES from external energy suppliers; This represents the physical loss of internal devices within the IES.

[0077] (11) (12) (13) (14) In the formula, T is the scheduling period. ( () represent the amplitudes of the control commands provided by IES to LA for the k types of energy at time t. For the corresponding energy output power, , , , These represent the unit cost (electricity), unit cost (gas), amount of electricity, and amount of gas that IES obtains from an external energy supplier at time t, respectively. / / / These represent the conversion factors for the operating status of various types of coupled units. / / / These represent the input power of the corresponding unit at time t.

[0078] 3.1.2 IES constraints.

[0079] While formulating its own optimization strategy, the IES needs to ensure the balance of energy transmission and distribution within its management area and maintain the safe and stable operation of the coupled equipment. The specific constraints are described below.

[0080] 3.1.2.1 Energy balance constraint: (15) 3.1.2.2 CHP Unit Constraints: (16) (17) (18) In the formula, and These represent the power generation and heating capacity of a combined heat and power (CHP) unit, respectively. and Each corresponds to its upper and lower limits. and These represent the power generation efficiency and heating efficiency of CHP, respectively.

[0081] 3.1.2.3 Constraints on P2G (electric to gas), gas-fired boilers (GB), and electric boilers (EB): (19) (20) In the formula, w represents P2G, GB, and EB, respectively. and It represents the output and input power of w at time t. It is the energy conversion efficiency of w. and These correspond to their upper and lower limits, respectively.

[0082] 3.1.2.4 Output constraints of renewable energy units: (twenty one) (twenty two) In the formula, and Let WT and PV represent the output power at time t, respectively. / Each corresponds to its upper limit.

[0083] 3.1.2.5 Control command amplitude constraint.

[0084] To ensure that control commands remain within the physically executable range on the load side, upper and lower limit constraints must be imposed on the amplitude of electrical / pneumatic control commands issued by the IES: (twenty three) In the formula, and ( ) represent the upper and lower limits of the amplitude of the IES electrical / pneumatic control command at time t, respectively.

[0085] In addition, the unit cost and quantity of energy obtained by IES from the upper-level power grid and upper-level gas grid should meet the following constraints.

[0086] (twenty four) (25) In the formula, and Let represent the upper and lower limits of the unit cost for IES to obtain energy from an external energy supplier at time t, respectively. and This corresponds to the upper and lower limits of the quantity of energy that can be purchased.

[0087] The load reduction command correction factor constraints are as follows: (26) 3.1.3 LA objective function.

[0088] As the responding entity, LA aims to maintain local supply and demand balance by optimizing energy storage charging and discharging strategies and energy substitution behaviors, that is, minimizing the local supply and demand power deviation. (27) In the formula, This represents the energy loss caused by physical characteristics such as internal resistance and conversion efficiency during the charging and discharging process of an energy storage system. This represents the equivalent energy deviation caused by differences in conversion efficiency when a user switches energy types. Both are modeled using quadratic functions to characterize the saturation characteristics of physical regulation capabilities.

[0089] (28) (29) In the formula, , , and These represent the conversion factors for physical losses during the charging and discharging of electrical energy storage. / These represent the energy storage charge / discharge power at time t; and These represent the conversion efficiency deviation factor for user substitution.

[0090] 3.1.4 LA constraint.

[0091] 3.1.4.1 Supply and demand balance constraints: Substitute the comprehensive energy response obtained from the multi-energy coupling matrix calculation in formula (6) into the local demand calculation to constrain the actual physical balance of LA. The relevant constraints are as follows: (30) 3.1.4.2 Energy storage constraints: (31) In the formula, and They represent the upper limits of energy storage charging and discharging power at time t, respectively. and Let represent the charge and discharge states of the stored energy at time t, respectively. This represents the stored energy at time t. This represents the self-loss coefficient of energy storage. and These represent the upper and lower limits of energy storage capacity, respectively.

[0092] 3.2 Model Solving.

[0093] In the aforementioned game theory model, due to the highly nonlinear and random nature of user responses, traditional analytical methods struggle to directly obtain user response information to control commands. Therefore, this proposal suggests a two-layer solution strategy combining LSTM and a PID controller, such as... Figure 3 As shown, after setting the basic parameters, the IES dynamic control command generation strategy, LA energy optimization strategy and IES overall optimization scheduling are performed in sequence. When the convergence condition is met, the control command sequence and energy strategy are output. If the convergence condition is not met, the PID of the control command sequence is dynamically updated by predicting load adjustment through LSTM, and the updated command sequence is fed back to the IES dynamic control command generation strategy until the convergence condition is met, and the required control command sequence and corresponding energy consumption strategy are obtained.

[0094] The LSTM predicts the user's next response behavior and load tracking deviation based on historical user response data, and transmits this information to the PID controller to calculate the adjustment efficiency increment under the current control command sequence. The convergence objective is to bring the adjustment efficiency increment caused by the control signal adjustment close to zero. The PID controller dynamically calculates the optimal control command correction for the next iteration based on the prediction deviation fed back by the LSTM, updating the control signal and transmitting the new control signal to the LA. This process repeats until the control command convergence condition (mathematical game-theoretic convergence) is met.

[0095] During the iterative process, the LSTM network acts as a data-driven behavioral simulator, capturing the temporal relevance and psychological inertia of user decisions to generate forward-looking estimated load response states for tentatively issued control commands. Based on these estimated states, the system calculates the "load fluctuation gradient," which is the deviation between the expected response and the estimated response. This fluctuation gradient serves as an error input signal to drive the downstream PID controller.

[0096] To overcome the operational risks posed by the inherent prediction bias of the LSTM data model, the PID controller utilizes a closed-loop feedback mechanism for correction. In parameter settings, the system does not employ fixed constants but strictly implements a "dual-mode adaptive mechanism" to adapt to varying load conditions: during periods of severe load fluctuation or in the early stages of iteration, the system uses "aggressive coefficients" (larger adjustment parameters) to rapidly explore the solution space of control commands; when the system enters a stable convergence phase, it dynamically switches to "conservative coefficients" (smaller adjustment parameters) for fine-tuning of the control commands.

[0097] Furthermore, a state-aware performance adjustment mechanism is embedded in the PID control loop. By constructing a comprehensive adaptive factor, the marginal adjustment performance on both the source and load sides is continuously evaluated. When the dynamic command update drives the system to approach the convergence equilibrium point, the iteration step size of the control command is constrained by this adaptive factor and subjected to mathematical damping, causing the update step size to smoothly decay to zero. This ensures that the closed-loop iteration process naturally and stably reaches the final control command convergence state.

[0098] 4. Scheduling.

[0099] Once the source-side controller (IES) and the load-side execution unit (LA) reach the convergence state of the control commands (mathematically equivalent to Nash equilibrium), the model outputs two key results: first, the optimal control command sequence for electricity and gas energy issued by the IES to the LA (i.e., the amplitude of the excitation signal at each moment); and second, the optimal tracking response plan fed back by the LA under the control commands (including the IDR adjustment of electricity, gas, and heat loads, the charging and discharging sequence of energy storage, and the heterogeneous energy substitution operation scheme).

[0100] In actual scheduling execution, IES first sends the optimal control instruction sequence obtained from the solution to LA as scheduling instructions.

[0101] After receiving the control command, the LA performs the following physical operations according to the optimal tracking response plan obtained from the model solution: During periods of high amplitude excitation signal, electricity demand is reduced by cutting interruptible loads or activating energy storage discharge. During periods of low amplitude of the incentive signal, excess renewable energy can be absorbed by increasing the electricity load or starting energy storage charging. When the difference in amplitude between electric and gas excitation signals exceeds the user response trigger boundary, an energy switching operation of "electricity instead of gas" or "gas instead of electricity" is performed to adjust the energy type of the terminal energy-consuming equipment.

[0102] At the same time, based on the tracking response plan fed back by LA, IES adjusts the output of coupled units such as CHP, P2G, GB, and EB to ensure that the output on the source side is balanced with the demand on the load side in real time.

[0103] Through the above process, the control commands and tracking response strategies obtained from the mathematical model are transformed into source-side unit output commands and load-side load control commands, driving physical equipment to perform actual scheduling actions, thereby realizing source-load coordinated optimization operation between IES and LA.

[0104] The essence of DR uncertainty lies in the fact that users, as individuals with autonomous decision-making capabilities, cannot have their behavior accurately predicted or be subject to forced commands. Traditional methods attempt to "overcome" this uncertainty through more accurate predictive models, but no matter how sophisticated the model is, the randomness of user behavior always exists.

[0105] This approach does not aim to accurately predict user behavior. Instead, it incorporates uncertainty into the closed-loop regulation process through a two-layer coordinated control framework. The source-side controller does not need to know the exact response function of LA; it only needs to approximate the convergent state through iterative trial and error. This design philosophy of "closed-loop feedback + iterative approximation" makes the system adaptive rather than fragile in the face of uncertainty.

[0106] From a cybernetics perspective, this scheme essentially reconstructs the IES scheduling problem into a two-stage tracking control system with uncertainty: the source-side controller outputs a setpoint (control command), and the load-side execution unit tracks this setpoint. The deviation between the two is gradually eliminated through LSTM prediction and PID correction. In this framework, IDR uncertainty is no longer a passively accepted "disturbance term" but a "compensable term" actively incorporated into the closed-loop regulation.

[0107] Case verification.

[0108] To verify the effectiveness of the proposed model, a typical day's energy cycle (IES) was selected as the object of analysis. The baseline load of the system's basic operating data (electricity / gas / heat multi-energy) and the predicted output of renewable energy are as follows: Figure 4 As shown. The operating parameters of CHP units, P2G and other coupled units, EES and GES are set according to existing literature. The time-sharing electrical control commands, gas control commands and multi-energy control commands are only examples.

[0109] To explore the specific contribution of the proposed model to the system's operational reliability, renewable energy absorption rate, and load regulation capability, the model was subjected to the following five scenario settings, as shown in Table 1.

[0110] Table 1 Scene Setting

[0111] The scheduling results of IES and LA in each scenario are shown in Table 2.

[0112] Table 2 Optimization results under different scenarios

[0113] (1) Comparative analysis of incentive and guidance mechanisms.

[0114] This solution compares the load curves of scenario 1 and scenario 2, such as... Figure 5 As shown in (a), compared to the initial load, the peak-to-valley difference of the electrical load curve in scenario 1 is reduced, demonstrating the peak-shaving and valley-filling effect of IDR. However, the peak-to-valley difference of the gas load is worsened, as shown in (a). Figure 5 As shown in (b), the ideal scheduling model fails; while the heat load changes in a more consistent manner. Figure 5 As shown in (c) in the figure.

[0115] Compared to Scenario 1, Scenario 2's load curve exhibits superior peak shaving and valley filling effects, with a smoother curve. Specifically, the peak-to-valley difference between electrical and gas loads decreased from 624.2kW ​​and 572.9kW to 461.8kW and 417.9kW, respectively, representing reductions of 26.02% and 27.06%. This is because, to ensure effective system response during peak and valley periods and mitigate the risk of insufficient response, the IES, as the leader, must implement a wider range of control signal amplitude adjustments and control strategies. This conservative strategy, based on an uncertainty model, forces the LA to allocate more flexible resources, thus achieving more thorough peak shaving and valley filling in Scenario 2 than in Scenario 1. This improves the reliability and stability of IES scheduling while significantly enhancing the overall operational reliability of the system.

[0116] (2) Comparative analysis of multi-energy coupling effect.

[0117] Figure 6 The distribution characteristics of electrical and gas load points under independent and coupled scheduling are shown. A comparison reveals that the operating domain area in Scenario 3 shows a convergence trend compared to Scenario 2. In Scenario 2, due to the separation of the electrical and gas systems, users blindly participate in IDR, resulting in many extreme points where energy demand is met but at a high cost. In Scenario 3, the coupling matrix provides the user-side with the flexibility of multi-energy coupling, enabling the execution of complex physical coupling strategies.

[0118] Specifically: At 7:00, the system implemented a precise "electricity-for-gas" substitution, increasing electrical load to meet heat demand and achieving a deep reduction in natural gas consumption, effectively smoothing out peak gas load. During the peak electrical load period from 11:00 to 13:00, the system switched to a "gas-for-electricity" strategy, using natural gas to replace peak electricity, significantly reducing the power supply pressure on the grid. Furthermore, during the nighttime off-peak period at 1:00, the system activated heat load flexibility, enabling the cross-time shift of heat energy. Through these multi-energy coupling physical mechanisms, the optimization algorithm adjusted the system to a more economical and efficient operating range. This not only reduced the load fluctuation range but also significantly enhanced the system's physical adjustment capabilities and energy supply reliability in response to disturbances on both the supply and demand sides.

[0119] (3) Comparative analysis of user dynamic response characteristics Figure 7 This demonstrates how dynamic response characteristics improve the user's energy experience. Comfort thresholds for electricity, gas, and heat were set at 0.8, 0.83, and 0.93, respectively. The conventional single-objective optimization scheduling in Scenario 3 resulted in a significant decrease in comfort, for example... Figure 7 In scenario (a), the electrical load comfort level drops to 0.63 at 5:00 AM, far below the user's psychological threshold, severely impacting the user's energy experience. Scenario 4 demonstrates the cross-time effect of comfort compensation. Figure 7 Taking the electrical load shown in (a) as an example, in scenario 3, where comfort levels are significantly below the threshold during the 2:00-7:00 period, scenario 4 effectively demonstrates a comfort compensation mechanism, ensuring that energy comfort meets the minimum standard during the 3:00-7:00 period. The same behavior is consistently observed in gas loads (e.g., 4:00-7:00) and heat loads, such as... Figure 7 As shown in (b) and (c) in the figure. When the comfort level is met, the user will continue to maintain the optimal economic response in the next period, thereby maximizing the system's operational reliability / renewable energy consumption rate while ensuring user satisfaction.

[0120] (4) Analysis of the advantages of the overall optimization model Figures 8-10 The output strategy of the internally coupled devices in IES is shown in "Scenario 5". Figure 8 This is a schematic diagram of the output of an electrical load device. Figure 9 This is a schematic diagram of the output of a gas-loaded device. Figure 10 This is a schematic diagram of the output of the heat load equipment.

[0121] It can be seen that at night, wind power (WT) is generated in large quantities, but the electrical load is at a low point, resulting in a light power and heavy heat phase, and the system faces a severe challenge of wind curtailment. At this time, the game equilibrium strategy prompts the IES to guide LA to increase base electricity demand by utilizing price elasticity and start EES (electric energy storage system) for charging. On the other hand, it activates the system-side P2G (electricity-to-gas) equipment to convert surplus wind power into natural gas, which not only meets the real-time gas load, but is also stored by the LA-side GES (gas energy storage system) for peak use, realizing a dual transfer of energy in terms of time and form.

[0122] Meanwhile, to make room for wind power (WT) to be fed into the grid, IES proactively reduced the power generation of the CHP (Combined Heat and Power) unit to its minimum stable operating state. Constrained by the physical principle that CHP's power output is determined by heat demand, the reduction in electricity output inevitably leads to a significant decrease in heat production, failing to meet the higher nighttime heat load demands. At this point, the system flexibly utilizes GB (Gas-fired Boilers) for supplemental heating. This collaborative operation mechanism—where P2G (Power-to-Gas) absorbs abandoned wind power, CHP (Combined Heat and Power) deeply regulates peak loads, and GB (Gas-fired Boilers) flexibly supplements heating—effectively decouples heat and power, significantly improving the system's wind power absorption capacity.

[0123] During peak daytime electricity load and fluctuating photovoltaic output, a period characterized by heavy electricity consumption and light heat consumption, the system operation strategy shifts to optimal supply-demand matching. On one hand, this period coincides with the grid load ramp-up phase, leading to increased system generation costs and electricity demand. The P2G (electricity-to-gas) unit proactively reduces its power output, prioritizing electricity originally intended for gas production to meet electricity demand, effectively reducing the overall peak-shaving pressure on the system's source side. On the other hand, because the P2G unit reduces gas production, the system coordinates the output of the GB (gas-fired boiler) unit with external gas purchase strategies to maintain a gas load supply-demand balance while ensuring that the CHP unit utilizes gas for power generation and heat production. This price-signal-based operation strategy not only avoids losses caused by frequent P2G unit start-ups and shutdowns but also achieves optimal time-scale coupling and scheduling of electricity, gas, and heat by limiting electricity-to-gas conversion during specific periods.

[0124] The performance comparison for different scenarios is shown in Table 3.

[0125] Table 3 Performance Comparison in Different Scenarios

[0126] Table 3 shows that Scenario 5 achieves a comprehensive improvement in system performance compared to Scenario 1. On the one hand, by guiding the flexible response on the load side and coordinating with the source-side equipment, the wind and solar power absorption capacity increases by 3.46%, effectively alleviating the problem of wind and solar power curtailment. On the other hand, users' energy comfort in the three dimensions of electricity, gas, and heat is improved. The results prove that the model proposed in this solution successfully balances low-carbon and environmental protection goals with user energy satisfaction while tapping the potential for demand-side regulation, achieving a win-win optimization effect for both source and load.

[0127] In summary, Scenario 5, through the two-sided game interaction between multiple coupled units on the source side and energy storage on the load side, constructed the optimal energy flow path during different load characteristic periods, realized the cascade utilization of energy, and further promoted the consumption of new energy sources. This verifies the advantages of the proposed model in improving the system's operational flexibility and the efficiency of multi-energy coordinated regulation.

[0128] Correspondingly, integrated energy dispatching systems that consider demand response uncertainty include: Integrated energy dispatching systems that consider demand response uncertainty include: The model building module is configured to: acquire load data of electricity, gas and heat in the integrated energy system and forecast output data of renewable energy; construct an integrated demand response uncertainty model based on the user's response to the incentive signal triggering boundary effect, cross-period load rebound effect and physical coupling constraints between heterogeneous energy sources; the integrated demand response uncertainty model includes a spatial multi-energy coupling sub-model for quantifying the substitution elasticity of heterogeneous energy sources, and a time dynamic response sub-model for characterizing the user's cross-period load compensation behavior; The source-load dual-layer coordinated control module is configured as follows: based on the comprehensive demand response uncertainty model, the integrated energy system is used as the source-side controller and the load aggregator is used as the load-side execution unit to construct a dual-layer coordinated control framework in which the source-side controller makes advance decisions and the load-side execution unit tracks the response. Among them, the source-side controller generates a dynamic control command sequence to maximize the system's supply and demand regulation efficiency, and the load-side execution unit tracks the response plan by minimizing the local power tracking deviation. The solution module is configured to: estimate the load tracking deviation based on historical user response data, calculate the control command correction amount based on the estimated deviation, and iteratively update the control command sequence until the control command convergence state is reached to obtain the optimal control command sequence; The scheduling output module is configured to convert the optimal control command sequence into source-side unit output commands and load-side load control commands, drive physical equipment to perform scheduling actions, and achieve coordinated and optimized operation of both the source and load sides.

[0129] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A comprehensive energy dispatching method considering demand response uncertainty, characterized in that, Includes the following steps: Obtain load data for electricity, gas, and heat in the integrated energy system, as well as predicted output data for renewable energy; construct an integrated demand response uncertainty model based on the user response to incentive signals triggering boundary effects, cross-period load rebound effects, and physical coupling constraints between heterogeneous energy sources. The integrated demand response uncertainty model includes a spatial multi-energy coupling sub-model for quantifying the elasticity of heterogeneous energy substitution, and a time dynamic response sub-model for characterizing users' cross-period load compensation behavior. Based on the comprehensive demand response uncertainty model, the integrated energy system is used as the source-side controller and the load aggregator is used as the load-side execution unit. A two-layer coordinated control framework is constructed, in which the source-side controller makes advance decisions and the load-side execution unit tracks the response. The source-side controller generates a dynamic control command sequence to maximize the system's supply and demand regulation efficiency, and the load-side execution unit tracks the response plan by minimizing the local power tracking deviation. The load tracking deviation is estimated based on historical user response data, and the control command correction amount is calculated based on the estimated deviation. The control command sequence is iteratively updated until the control command convergence state is reached, and the optimal control command sequence is obtained. The optimal control command sequence is transformed into source-side unit output commands and load-side load control commands, driving physical equipment to perform scheduling actions and achieving coordinated and optimized operation of both the source and load sides.

2. The integrated energy dispatching method considering demand response uncertainty as described in claim 1, characterized in that, The spatial multi-energy coupling sub-model is constructed by decomposing the coupling effect between heterogeneous energy sources into two dimensions: response willingness and response capability. Response willingness represents the user's cognitive sensitivity to the relative differences in excitation signals between heterogeneous energy sources, while response capability represents the physical adjustment boundary constrained by real-time energy consumption status.

3. The integrated energy dispatching method considering demand response uncertainty as described in claim 2, characterized in that, The response capability is shown in the following formula: ； In the formula, and These represent the upper and lower limits of the baseline load for user k energy sources during the scheduling period, respectively. This represents the baseline load of k energy sources at time t. and These represent the upward and downward adjustment capabilities of k types of load at time t, respectively. This represents the normalized adjustment capability of k energy sources at time t.

4. The integrated energy dispatching method considering demand response uncertainty as described in claim 2, characterized in that, The desired response is expressed as follows: ； In the formula, This indicates the user's willingness to switch from electricity to gas at time t. , These represent the maximum amplitude of the electrical excitation and pneumatic excitation signals within a scheduling cycle, respectively.

5. The integrated energy dispatching method considering demand response uncertainty as described in claim 1, characterized in that, The time-dynamic response sub-model is shown in the following equation: ； In the formula, Traditional user comfort models representing k types of energy. This represents the user's overall comfort level with respect to the k types of energy at time t. The sensitivity coefficient of the excitation signal. and These represent the real-time excitation signal amplitude and the average reference value of the excitation signal within the scheduling period, respectively. These are the status parameters of IDR.

6. The integrated energy dispatching method considering demand response uncertainty as described in claim 5, characterized in that, The time-dynamic response sub-model also includes a trigger response mechanism based on the accumulation of comfort deviations, specifically: setting a trigger boundary for the user's comfort response to the k-th energy source. When the user's overall comfort level falls below the trigger threshold at time t due to over-responding to the stimulus signal, the user actively adjusts the dynamic response at time t+1 by reducing the intensity of the response to the stimulus signal. This will bring overall comfort back to a feasible level.

7. The integrated energy dispatching method considering demand response uncertainty as described in claim 1, characterized in that, The objective function of the source-side controller is shown in the following equation: ； ； ； ； ； In the formula, The total amount of energy output by the IES to the load aggregator is regulated; This represents the flexible load dispatch adjustment amount issued by IES to the load aggregator; This represents the discounted value of the amount of energy injected into the IES from external energy suppliers; This represents the physical loss calculation of the internal devices of the IES; T is the scheduling period. These represent the control command amplitudes of the k energy sources provided by IES to the load aggregator at time t. For the corresponding energy output power, , , , These represent the unit cost of electricity, unit cost of gas, amount of electricity obtained by IES from external energy suppliers at time t, and amount of gas obtained, respectively. / / / These represent the conversion factors for the operating status of various types of coupled units. / / / These represent the input power of the corresponding unit at time t.

8. The integrated energy dispatching method considering demand response uncertainty as described in claim 1, characterized in that, The objective function of the load-side execution unit is shown in the following equation: ； ； ； In the formula, This refers to the energy loss caused by the physical characteristics of the energy storage system during charging and discharging. The equivalent energy deviation caused by differences in conversion efficiency when users switch energy types; and All models employ quadratic functions to characterize the saturation properties of physical regulation capabilities; , , and These represent the conversion factors for physical losses during the charging and discharging of electrical energy storage. / These represent the energy storage charge / discharge power at time t; and These represent the conversion efficiency deviation factor for user substitution.

9. The integrated energy dispatching method considering demand response uncertainty as described in claim 1, characterized in that, The control command convergence state is specifically defined as a state in which neither the source-side controller nor the load-side execution unit can obtain a better regulation effect by unilaterally changing the control strategy, which is mathematically equivalent to Nash equilibrium. During the iterative update process, the incremental regulation efficiency caused by the adjustment of the control signal is taken as the convergence target to approach zero.

10. A system for implementing the integrated energy dispatching method considering demand response uncertainty as described in any one of claims 1-9, characterized in that, include: The model building module is configured to: acquire load data of electricity, gas and heat energy and renewable energy predicted output data in the integrated energy system, and build an integrated demand response uncertainty model; The integrated demand response uncertainty model includes a spatial multi-energy coupling sub-model for quantifying the elasticity of heterogeneous energy substitution, and a time dynamic response sub-model for characterizing users' cross-period load compensation behavior. The source-load dual-layer coordinated control module is configured as follows: based on the comprehensive demand response uncertainty model, the integrated energy system is used as the source-side controller and the load aggregator is used as the load-side execution unit to construct a dual-layer coordinated control framework in which the source-side controller makes advance decisions and the load-side execution unit tracks the response. The solution module is configured to: estimate the load tracking deviation based on historical user response data, calculate the control command correction amount based on the estimated deviation, and iteratively update the control command sequence until the control command convergence state is reached to obtain the optimal control command sequence; The scheduling output module is configured to convert the optimal control command sequence into source-side unit output commands and load-side load control commands, drive physical equipment to perform scheduling actions, and achieve coordinated and optimized operation of both the source and load sides.

Images