Multi-stage elastic scheduling method for gas-electricity coupled energy system considering source-load flexibility coordination

Abstract

The application discloses a multi-stage elastic scheduling method of a gas-electricity coupling energy system considering source-load flexibility coordination, relates to the technical field of energy system optimal scheduling, and comprises the following steps: a multi-energy flow physical coupling model containing a gas turbine, a ground source heat pump and natural gas pipe network dynamic pipe storage is established; equivalent thermal parameters of buildings are identified through a machine learning method fusing physical mechanism constraints, indoor temperature elastic boundaries are determined based on a thermal comfort model of user sensory fuzziness, and load end demand response reserve capacity is calculated; a collaborative reserve pool composed of an electric side energy storage, a load end demand response and a natural gas pipe network dynamic pipe storage is constructed based on the multi-energy flow coupling model and the reserve capacity, and multi-stage elastic scheduling is executed accordingly: in a day-ahead stage, source-load-grid collaborative reserve capacity is evaluated and optimal scheduling is performed; in a post-disaster stage, load hierarchical recovery is performed with the aim of prolonging system survival time and guaranteeing important load power supply.

Description

Technical Field

[0001] This invention relates to the field of energy system optimization and scheduling technology, and more specifically, to a multi-stage flexible scheduling method for gas-electric coupled energy systems that takes into account source-load flexible coordination. Background Technology

[0002] Gas-electric coupling systems improve the reliability and low-carbon level of multi-energy supply by physically integrating gas turbines and electric drive equipment. However, in the event of power or gas outages caused by extreme weather such as extreme cold, freezing temperatures, or typhoons, most existing flexible dispatch strategies still treat electric energy storage as a single passive backup resource.

[0003] This traditional model has two main problems: First, in order to cope with extreme failures that have a low probability of occurrence but high destructive power, the system must be equipped with large-scale energy storage, which increases the demand for backup resources and puts great scheduling pressure on the system operation.

[0004] Meanwhile, traditional methods typically treat user-side load as rigid demand, failing to fully consider building thermal inertia and users' proactive response and load reduction capabilities under extreme conditions. On the other hand, current scheduling models, when dealing with uncertainty, often require pre-assuming a probability distribution of failures or imposing strict rigid supply constraints on critical loads. However, in actual extreme disasters, the patterns of failure occurrence are difficult to predict accurately. Once the actual scenario deviates from the assumed model, it can easily lead to reserve depletion and scheduling strategy failure.

[0005] Therefore, it is necessary to explore an elastic scheduling method that can fully tap the flexible adjustment potential at both ends of the source and load and adapt to multiple uncertainties under extreme operating conditions. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a multi-stage elastic scheduling method for gas-electric coupled energy systems that takes into account source-load flexible coordination, in order to solve the problem that traditional scheduling methods lack a source-load coordinated elastic scheduling mechanism and are difficult to maintain system operation and ensure power supply to important loads under extreme power outage and gas outage scenarios.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A multi-stage elastic dispatch method for gas-electric coupled energy systems that considers source-load flexible coordination includes the following steps: A multi-energy flow physical coupling model is established, incorporating gas turbines, ground source heat pumps, and dynamic storage in the natural gas pipeline network. Equivalent building thermal parameters are identified using a machine learning method that integrates physical mechanism constraints. Indoor temperature elastic boundaries are determined based on a thermal comfort model with user perception fuzziness, and load-side demand response reserve capacity is calculated. Based on the multi-energy flow coupling model and reserve capacity, a collaborative reserve pool consisting of electricity-side energy storage, load-side demand response, and dynamic storage in the natural gas pipeline network is constructed. Multi-stage elastic scheduling is then implemented: in the day-ahead phase, the source-load-grid collaborative reserve capacity is assessed and optimized; in the post-disaster phase, load tiered recovery is implemented with the goal of extending system survival time and ensuring power supply to critical loads.

[0008] In a preferred embodiment, the multi-energy flow physical coupling model includes: the gas-electric-thermal energy conversion equation of the gas turbine, the electric-thermal conversion equation of the ground source heat pump, and the dynamic pipeline storage equation of the natural gas network; wherein the dynamic pipeline storage equation quantifies the transient natural gas stored in the pipeline as a dispatchable reserve capacity based on the node pressure and pipeline airflow continuity.

[0009] In a preferred embodiment, the machine learning method for identifying building equivalent thermal parameters by incorporating physical mechanism constraints employs a physical information neural network.

[0010] In a preferred embodiment, the calculation of load-side demand response reserve capacity includes: identifying equivalent building thermal parameters using a physical information neural network, the loss function of which includes the residuals of the building thermodynamic differential equation; establishing a trapezoidal membership function for user thermal comfort perception based on fuzzy theory to determine the minimum allowable indoor limit temperature trajectory during a fault; calculating the minimum thermal power required to maintain a baseline comfort level based on the currently identified parameters and the limit temperature, and mapping it to the load-side demand response reserve capacity in the electrical power space through the heating performance coefficient of the ground source heat pump.

[0011] In a preferred embodiment, the optimized scheduling includes: performing optimized scheduling based on the collaborative backup pool during the day-ahead phase to meet the minimum disaster relief backup capacity requirement; constructing a Wasserstein fuzzy set based on historical meteorological and fault data, and introducing a pluralistic chance constraint to ensure that the power supply satisfaction rate of important loads is not lower than a preset confidence level under the worst probability distribution within the fuzzy set; and solving for the day-ahead scheduling plan with the goal of minimizing energy purchase cost and carbon emission penalty cost.

[0012] In a preferred embodiment, the post-disaster phase involves tiered load recovery: an emergency recovery optimization model is established with the goal of maximizing system survival time and load recovery amount; recovery weight coefficients are set according to load importance, and power supply is restored to important loads first; dynamic pipeline storage resources of the natural gas network are prioritized to support gas turbine operation to ensure power supply to important loads; when pipeline storage support capacity is insufficient, the discharge of electric energy storage is scheduled; when the energy storage charge state approaches the safe lower limit, the integrated demand response is further invoked to increase the reserve capacity.

[0013] In a preferred embodiment, the dynamic pipeline storage of the natural gas network is converted into equivalent electrical power reserve capacity through a gas turbine and participates in optimized scheduling.

[0014] The technical effects and advantages of this invention regarding the multi-stage elastic scheduling method for gas-electric coupled energy systems with flexible source-load coordination are as follows: This invention establishes a multi-energy flow physical coupling model including gas turbines, ground source heat pumps, and dynamic natural gas pipeline storage. Based on a machine learning method incorporating physical mechanism constraints, it identifies equivalent building thermal parameters and quantifies load-side demand response reserve capacity using a user-perceived fuzzy thermal comfort model. This constructs a collaborative reserve pool composed of electricity-side energy storage, demand response, and dynamic natural gas pipeline storage, enabling coordinated scheduling of multi-energy resources across the source, load, and grid. Furthermore, by conducting joint reserve capacity assessment and optimized scheduling during the day-ahead phase and implementing a load tiered recovery strategy during the post-disaster phase, it improves the system's resilience and reliability under extreme disaster conditions. By introducing a distributed bar chance constraint method to characterize the uncertainty of extreme failures, it helps improve the robustness and risk resistance of scheduling decisions, effectively solving the problem that traditional scheduling methods lack a source-load coordinated elastic scheduling mechanism, making it difficult to maintain system operation and ensure power supply to critical loads under extreme power and gas outage scenarios. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the multi-stage flexible scheduling method for a gas-electric coupled energy system that takes into account the flexible coordination of source and load, provided in an embodiment of the present invention.

[0016] Figure 2 This is a diagram showing the change in indoor temperature during an extreme failure.

[0017] Figure 3 This chart compares the energy consumption and survival time of the main grid power and gas outage system under extreme disasters. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, Figure 1 This invention presents a multi-stage elastic scheduling method for gas-electric coupled energy systems that considers source-load flexible coordination, comprising the following steps: S1. Establish a multi-energy flow physical coupling model that includes gas turbines, ground source heat pumps and dynamic storage of natural gas pipelines; S2 identifies the building's equivalent thermal parameters by integrating physical mechanism constraints through machine learning methods, and determines the indoor temperature elastic boundary based on the thermal comfort model of user perception ambiguity, and calculates the load-side demand response reserve capacity. S3, Based on the aforementioned multi-energy flow coupling model and reserve capacity, a collaborative reserve pool is constructed, consisting of electricity-side energy storage, load-side demand response, and dynamic pipeline storage in the natural gas network, and multi-stage elastic scheduling is performed accordingly: The current phase involves assessing the source-load-grid collaborative backup capacity and optimizing scheduling. In the post-disaster phase, the goal is to extend the system's survival time and ensure power supply to critical loads, and load tiered restoration is implemented.

[0020] This invention breaks away from the traditional static standby mode that relies solely on electricity-side energy storage, innovatively incorporating the dynamic pipeline storage effect and comprehensive demand response of the natural gas pipeline network into the standby resource pool. By coordinating the use of implicit pipeline energy storage and load-cutting potential in both temporal and spatial dimensions, it significantly extends the system's survival time under extreme disasters involving both gas and electricity outages without increasing investment in redundant physical energy storage. Addressing the issue of missing or noisy measurement data due to communication disruptions during disasters, this invention introduces a physical information neural network regularized by physical laws to identify equivalent building thermal parameters and incorporates the perceptual ambiguity of user thermal comfort. This mechanism overcomes the overfitting risk of purely data-driven algorithms while ensuring that the extracted load-cutting potential strictly meets thermodynamic physical boundaries and the baseline for human survival, effectively preventing a dangerous overestimation of adjustable elastic standby capacity. In the day-ahead scheduling phase, it innovatively introduces a split-bar chance constraint based on the Wasserstein metric. Compared to traditional deterministic absolute power supply constraints or conventional chance constraints, this method does not require assumptions about the precise probability distribution of extreme disasters. It can still strictly guarantee the power supply satisfaction rate of primary critical loads even under the worst probability distribution scenarios, and successfully achieves the optimal mathematical balance between the daily low-carbon optimized operation of the integrated energy system and the disaster resilience under extreme faults.

[0021] S1. Establish a multi-energy flow physical coupling model that includes gas turbines, ground source heat pumps, and dynamic storage of natural gas pipelines.

[0022] In the event of extreme disasters, the external power grid or main gas source may face sudden interruptions. To accurately assess the survivability and support capabilities of the system, this step establishes a multi-energy flow physical coupling model that takes into account the gas turbine, ground source heat pump, thermal storage device, and dynamic storage of the natural gas pipeline network.

[0023] 1. Modeling of core gas-electric and electrothermal coupling equipment The gas-electric coupling system realizes the mutual conversion of gas, electricity and heat three-phase energy flow through gas turbine (GT) and ground source heat pump (GSHP).

[0024] A. Dynamic conversion model for gas turbines: The gas turbine, serving as the system's core backup power source, consumes natural gas while simultaneously outputting electricity and heat. Its output electrical power... With heat production power The physical association is represented as: (1) (2) In the formula, Let be the volumetric flow rate of natural gas consumed by the gas turbine at time t; It is the higher calorific value of natural gas; and These represent the power generation efficiency and heat recovery efficiency of the gas turbine, respectively.

[0025] B. Ground source heat pump conversion model: Ground source heat pumps utilize shallow geothermal energy to efficiently generate heat by consuming electricity. Their heat production capacity... The model is as follows: (3) In the formula, The electrical power consumed by the ground source heat pump; Its heating performance coefficient.

[0026] 2. Dynamic Energy Boundary Modeling of Thermal Storage Devices To mitigate heat load fluctuations and provide passive heating support during faults, a thermal energy storage (TES) device is introduced. Its state evolution equations and heat charge / discharge constraint modeling are as follows: (4) (5) In the formula, Let be the heat stored at time t; This is the self-loss coefficient; and These are the heat charging efficiency and the heat dissipation efficiency, respectively. and These are the power for heat charging and heat dissipation, respectively. The scheduling time step can be set to 15 minutes or 1 hour in actual implementation, depending on the system scheduling accuracy.

[0027] 3. Dynamic pipeline reserve modeling for natural gas pipeline networks Traditional steady-state natural gas flow models neglect the pipeline's own natural gas storage capacity. In the face of sudden failures, the transient linepack within the pipeline can serve as a hidden, flexible backup resource, providing critical short-term fuel support for the gas turbine.

[0028] For a natural gas pipeline connecting node m and node n, its internal pipe storage... The relationship between the pressure at both ends of the pipeline is approximately as follows: (6) In the formula, This is a physical constant of the pipeline, which is related to the pipeline length, cross-sectional area, and gas temperature; and These are the instantaneous pressures at the two end nodes, respectively. It should be noted that, The physical expression is: ,in For the volume of the pipe, It is the gas compressibility factor. The gas constant is... This refers to the gas temperature. If the temperature and compressibility factor change, It should be a variable, not an absolute constant.

[0029] After introducing dynamic pipe storage, the airflow continuity equation of the pipeline is modified to: (7) (8) In the formula, and These represent the inflow and outflow of natural gas into the pipeline, respectively. This dynamic equation extends the natural gas pipeline network from a simple transmission medium to an energy storage carrier with energy buffering capabilities, participating in subsequent joint minimum standby capacity assessments.

[0030] S2 identifies the building's equivalent thermal parameters by incorporating machine learning methods that integrate physical mechanism constraints, determines the indoor temperature elastic boundary based on a thermal comfort model with user perception ambiguity, and calculates the load-side demand response reserve capacity.

[0031] In the event of an external gas and electricity supply disruption due to extreme failures, fully utilizing the integrated demand response (IDR) capabilities on the user side can significantly reduce the system's reliance on expanding expensive energy storage equipment. This step identifies building thermal parameters using a deep learning algorithm regularized by physical laws, and combines this with the user's fuzzy perception of temperature to accurately quantify the dynamic adjustability margin of the IDR (demand response reserve capacity).

[0032] 1. Identification of building thermal parameters under extreme conditions based on PINN The heating or cooling load on the user side has significant thermal inertia, and its actual response potential is highly dependent on the building's equivalent thermal resistance. With equivalent heat capacity In extreme disasters, communication networks are often damaged, leading to significant data loss or severe noise in smart meter readings. Traditional pure data-driven algorithms are prone to overfitting in such scenarios with small sample sizes or harsh data conditions.

[0033] To address this, a Physical Information Neural Network (PINN) is introduced, embedding the laws of building thermodynamics as a hard penalty term into the loss function of the neural network. (Building indoor temperature) The first-order differential dynamic equation is: (9) In the formula, The total heat power provided to the building by the integrated energy system is supplied by a ground source heat pump. With gas turbine heat recovery Together they constitute, and their expression is: ,in To provide heating power for the ground source heat pump, To provide heating power for waste heat recovery from gas turbines, This refers to the outdoor ambient temperature.

[0034] Construct a joint loss function that includes data loss and physical equation residuals. : (10) In the formula, For sparse actual measured temperature; Predict temperature for the network; N and M are the number of measurement data points and physical matching points, respectively; and These are adaptive weight coefficients. Through iterative optimization of this loss function, even in the event of missing data during the failure period, the identification of [the target data] can be ensured. and Strictly adhere to thermodynamic boundaries to prevent overestimating the potential for subsequent load reduction.

[0035] The physical information neural network is trained using the gradient descent optimization algorithm. The network parameters can be iteratively updated using either the Adam algorithm or stochastic gradient descent (SGD) to optimize the joint loss function. It converges to the minimum value.

[0036] 2. User perception ambiguity and minimum indoor temperature trajectory modeling When the system is running normally, the indoor temperature is usually maintained at the user's optimal set value. However, under extreme failure conditions, the user's tolerance for temperature deviations is ambiguous, allowing the temperature to deviate from the optimal value within certain limits without triggering a serious survival crisis. For example... Figure 2 As shown in the figure, the indoor temperature changes during an extreme failure. Due to the building's thermal inertia, the indoor temperature slowly decreases from the optimal set temperature of 22°C, and remains below the user's acceptable lower limit of 16°C, thus significantly reducing the power consumption of the ground source heat pump.

[0037] Establish a membership function to describe the fuzziness of user thermal comfort perception. Taking winter heating as an example, the lowest acceptable indoor temperature limit is defined as... Its trapezoidal fuzzy membership function is expressed as: (11) when When the load approaches zero, it indicates that users have reached their tolerance limit, and the load no longer has the elasticity to be reduced further. This is addressed by setting a minimum permissible fuzzy satisfaction threshold during system failures. This allows us to deduce the trajectory of the lowest physically permissible indoor temperature under the current conditions. .

[0038] 3. Quantification of Adjustable IDR Backup Margin under Source-Load Coordination The determined lowest indoor temperature trajectory Substituting these values ​​into the identified thermodynamic differential equations, the minimum rigid heat power required to maintain this limiting temperature can be calculated. .

[0039] Furthermore, considering the conversion efficiency of the ground source heat pump in step S1... This maps the elasticity of thermodynamic space to the reduction potential of electrical power space. It defines the load-side integrated demand response at time t as the upward adjustment of reserve capacity (i.e., the maximum reducible electrical load). : (12) In the formula, This refers to the baseline operating power of the ground source heat pump determined during the recent scheduling phase.

[0040] Similarly, for hot water load and conventional, interruptible, non-critical electrical load, the corresponding power reduction can be calculated and compared with the above. The combined system will form a comprehensive demand response reserve capacity.

[0041] S3, Based on the aforementioned multi-energy flow coupling model and reserve capacity, a collaborative reserve pool is constructed, consisting of electricity-side energy storage, load-side demand response, and dynamic pipeline storage in the natural gas network, and multi-stage elastic scheduling is performed accordingly: The current phase involves assessing the combined backup capacity of the source-load network and optimizing its scheduling. In the post-disaster phase, the goal is to extend the system's survival time and ensure power supply to critical loads, and load tiered restoration is implemented.

[0042] Traditional deterministic scheduling often proves vulnerable to extreme failures (such as main grid power outages caused by typhoons or gas supply interruptions caused by freezing), while conventional chance constraints rely excessively on precise assumptions about the probability distribution of failures. To balance the low-carbon optimization of daily system operation with high resilience under extreme failures, this paper proposes a three-stage collaborative scheduling architecture that includes dynamic reserve assessment, day-ahead optimized scheduling, and post-disaster tiered recovery.

[0043] Phase 1: Three-dimensional collaborative minimum reserve capacity assessment of source-load-grid system During the day-ahead dispatch phase, the system must reserve sufficient backup capacity to cope with potential risks of sudden power and gas outages. Unlike the traditional static backup relying solely on electric energy storage (ESE), this method constructs a three-dimensional combined backup pool that includes energy storage, integrated demand response (IDR), and dynamic pipeline storage of natural gas.

[0044] Define the minimum disaster recovery backup capacity required by the system at time t as follows: The total increased reserve capacity actually provided by the system Must meet: (13) In the formula, The maximum reserve capacity that electric energy storage can provide based on the current state of charge (SOC) can be expressed as: ,in To store the maximum discharge power, For energy storage, the current state of charge, This represents the lower limit of energy storage safety. The reserve capacity for the integrated demand response at the load end calculated in step S2; The equivalent electrical power that can be provided by the dynamic storage of the natural gas pipeline network in step S1 through gas turbine conversion is reserved.

[0045] The aforementioned collaborative mechanism significantly reduces the capacity configuration requirements of a single energy storage device.

[0046] 2. Phase Two: Day-ahead Scheduling Based on Wasserstein Partial Bruker Chance Constraints Under the premise of meeting the Phase 1 reserve constraints, the system performs day-ahead low-carbon optimized scheduling. The probability distribution of power and gas outages due to extreme disasters is also considered. Precise data is often difficult to obtain. This invention utilizes historical meteorological and power grid fault data to construct an empirical distribution. And introduce the Wasserstein metric to construct a fuzzy set of the true probability distribution. : (14) In the formula, The set of all possible probability distributions; Wasserstein distance metric; The radius of the fuzzy set is used to reflect the degree of conservatism of decision-makers regarding extreme disaster risks.

[0047] For critical loads within the system (such as hospitals and core data centers in industrial parks), the traditional absolutely rigid power supply constraint is changed by introducing Distributed Broken Chance Constraint (DRCC). This requires that, under the worst-case probability distribution within the fuzzy set, the power supply satisfaction rate of critical loads should still not be lower than the confidence level. : (15) In the formula, This refers to the actual power that the system can provide to critical loads under extreme operating conditions; This is for critical load demand.

[0048] The objective function of the current scheduling aims to minimize the system's daily operating costs (electricity and gas purchase costs) and carbon emission penalty costs. (16) To solve this model, the DRCC with infinite-dimensional probability distribution is transformed into a deterministic convex optimization constraint or mixed integer linear constraint (MILP) that is easy to solve using strong duality theory. The solution is then given to a commercial solver (such as CPLEX or Gurobi) for global optimization, which outputs the day-ahead optimal equipment output trajectory.

[0049] 3. Phase Three: Load Tiered Recovery Strategy Based on Fault Severity Ranking When an extreme disaster actually occurs and causes an external energy outage, the system enters the post-disaster emergency recovery phase. At this time, the scheduling objective shifts from optimizing for low carbon emissions in the daytime phase to maximizing survival time and load recovery.

[0050] Based on the degree of system damage and the depletion of the source-load grid's reserve, a recovery optimization model taking into account load weights is established: (17) In the formula, The time when the fault occurred; This is the expected end point of the fault. The set of all load nodes in the system; The weighting coefficient for the i-th type of load (the weight of primary critical loads is much greater than that of interruptible flexible loads); The actual power restored to node i.

[0051] During emergency dispatching, dynamic pipeline storage is prioritized to support the gas turbine supply to the core load. Subsequently, the electric energy storage is activated for discharge. When the energy storage SOC approaches its lower limit, the load-side IDR is reduced based on the lowest indoor temperature trajectory in step S2, until the external power grid or gas network returns to normal. This dispatching sequence ensures priority utilization of the low-cost resource of dynamic pipeline storage, followed by energy storage resources. Demand response is then invoked only when energy storage approaches its safe lower limit, thereby maximizing system lifetime.

[0052] 4. Verification of Extreme Fault Recovery Effectiveness and Survival Time Assessment This step verifies the effectiveness of the elastic scheduling strategy based on three-dimensional backup coordination and perception fuzziness in extending system lifetime and ensuring power supply to critical loads by setting up a scenario of power and gas outages caused by extreme cold disasters.

[0053] 1) Setting extreme disaster scenarios The integrated energy system is simulated to experience an extreme cold and freezing disaster on a typical winter day. The fault scenario parameters are as follows: The outage will occur from 18:00 to 22:00, a continuous 4-hour evening rush hour.

[0054] Fault phenomenon: An N-1 cascade fault in the external distribution network causes a complete power supply interruption. Meanwhile, the extreme cold caused the upstream natural gas pressurization station to shut down, resulting in an interruption of the external gas supply.

[0055] System status: At the time of the fault, the initial state of charge (SOC) of the electric energy storage (ESE) was 80%; the initial storage pressure of the natural gas pipeline was sufficient; the outdoor temperature was -5°C, and the optimal indoor temperature setting was [not specified]. The maximum tolerable fuzzy temperature is 22°C. The temperature is 16°C.

[0056] 2) Comparison strategy between control group and experimental group Two scheduling strategies are compared: Strategy A (Traditional Static Backup Dispatch): Relying solely on electric energy storage (ESE) to discharge during power outages to support all loads, without considering dynamic natural gas pipeline storage, and strictly maintaining indoor temperature at 22°C (rigid heat load).

[0057] Strategy B (the source-load-grid coordinated elastic scheduling proposed in this invention): In the instant of power and gas outage, energy storage discharge and natural gas pipeline storage are coordinated to support the small output of gas turbines. Based on the user perception fuzzy model, the indoor temperature limit is relaxed to 16°C, releasing the Integrated Demand Response (IDR) reserve capacity to prioritize the protection of first-level critical loads such as hospitals.

[0058] 3) Survival time and multidimensional resource response analysis Under Strategy B, during the first two hours of the initial fault period, the system utilizes dynamic storage to maintain the gas turbine at its minimum steady-state operation, replacing some of the energy storage output. Simultaneously, combined with the equivalent thermal parameters identified in step S2, the building's indoor temperature slowly decreases following thermodynamic inertia, resulting in a significant reduction in the power consumption of the ground source heat pump (GSHP).

[0059] When the storage is depleted and the indoor temperature reaches the lower limit of 16°C, the system switches to extreme survival mode, and the stored energy releases the remaining power. Compared with Strategy A, which depletes the stored energy within 2.5 hours after a fault, leading to a system-wide power outage, Strategy B successfully extends the system survival time to more than 4 hours, fully covering the preset extreme fault period and achieving zero power outage for the primary critical load.

[0060] Figure 3 The simulation results depict energy storage and temperature changes. Under extreme cold conditions with power and gas outages, Strategy A, rigidly adhering to the requirement of maintaining normal system operation, resulted in its electrical energy storage discharging at an extremely high rate. The red dashed line indicates that approximately 2.5 hours after the fault (around 20:30), the energy storage SOC dropped to the safe lower limit of 10%, leading to a complete system collapse and inability to support critical loads in subsequent periods. Conversely, when using Strategy B (the blue solid line), the system SOC decline exhibited a clear two-stage nonlinear characteristic. 1.5 hours before the fault, the system significantly slowed the energy storage degradation rate by utilizing residual gas pipeline storage to support a small output from the gas turbine. Ultimately, Strategy B successfully extended the energy storage support time to over 4 hours, ensuring power continuity throughout the evening peak period.

[0061] The extended survival time of Strategy B also benefits from the flexible reduction of demand-side heat load. The green solid line illustrates the indoor temperature evolution trajectory based on PINN thermal parameters and perceptual ambiguity. During a fault, the system stops full-power heating, allowing the indoor temperature to slowly decrease from 22°C due to building thermal inertia. After reaching the extreme tolerance boundary of 16°C around 20:30, the ground source heat pump outputs only a minimal basic power to maintain the lower limit temperature. This elastic margin of transition from thermodynamic space to electrical power space effectively avoids premature depletion of energy storage. Table 1 compares the key performance indicators of the traditional static standby scheduling strategy and the source-load-grid coordinated elastic scheduling strategy proposed in this invention. It can be seen that this invention has significant advantages in system survival time, power supply continuity for important loads, and energy storage utilization efficiency.

[0062] Table 1

[0063] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0064] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0065] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0066] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0067] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0068] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-stage elastic scheduling method for a gas-electric coupled energy system considering flexible source-load coordination, characterized in that, include: Establish a multi-energy flow physical coupling model that includes gas turbines, ground source heat pumps, and dynamic storage of natural gas pipelines; The equivalent thermal parameters of the building are identified by machine learning methods that integrate physical mechanism constraints, and the indoor temperature elastic boundary is determined based on the thermal comfort model of user perception fuzziness, and the load-side demand response reserve capacity is calculated. Based on the aforementioned multi-energy flow coupling model and reserve capacity, a collaborative reserve pool is constructed, consisting of electricity-side energy storage, load-side demand response, and dynamic pipeline storage in the natural gas network, and multi-stage elastic scheduling is performed accordingly. The current phase involves assessing the source-load-grid collaborative backup capacity and optimizing scheduling. In the post-disaster phase, the goal is to extend the system's survival time and ensure power supply to critical loads, and load tiered restoration is implemented.

2. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination as described in claim 1, characterized in that, The multi-energy flow physical coupling model includes: the gas-electric-thermal energy conversion equation of the gas turbine, the electric-thermal conversion equation of the ground source heat pump, and the dynamic pipeline storage equation of the natural gas pipeline network. The dynamic pipeline storage equation, based on node pressure and pipeline gas flow continuity, quantifies the transient natural gas stored in the pipeline as a reserve capacity available for dispatch.

3. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination as described in claim 1, characterized in that, The machine learning method that integrates physical mechanism constraints employs a physical information neural network.

4. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination as described in claim 3, characterized in that, The calculated load-side demand response reserve capacity includes: The physical information neural network is used to identify the equivalent thermal parameters of a building, and its loss function includes the residuals of the building thermodynamic differential equation. Based on fuzzy theory, a trapezoidal membership function for user thermal comfort perception is established to determine the minimum allowable indoor temperature trajectory during a fault. Based on the current identification parameters and extreme temperatures, the minimum thermal power required to maintain the baseline human comfort is calculated, and the heating performance coefficient of the ground source heat pump is mapped to the load-side demand response reserve capacity of the electrical power space.

5. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination as described in claim 1, characterized in that, The optimized scheduling includes: In the current phase, optimized scheduling is performed based on the aforementioned collaborative backup pool to meet the minimum disaster recovery backup capacity requirements; Based on historical meteorological and fault data, a Wasserstein fuzzy set is constructed, and a sub-Bruker chance constraint is introduced to ensure that the power supply satisfaction rate of important loads is not lower than the preset confidence level under the worst probability distribution within the fuzzy set. The day-ahead dispatch plan is obtained by minimizing the energy purchase cost and carbon emission penalty cost.

6. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination as described in claim 1, characterized in that, The post-disaster phase will implement load tiered recovery: An emergency recovery optimization model is established with the goal of maximizing system survival time and load recovery. A restoration weighting coefficient is set based on the importance of the load, and power supply is restored to important loads first; Prioritize the use of dynamic pipeline storage resources in the natural gas network to support gas turbine operation in order to ensure power supply to critical loads; When the storage capacity is insufficient, the discharge of the electric energy storage on the pipeline side is scheduled. When the energy storage's state of charge approaches the safe lower limit, the integrated demand response is further activated to increase the reserve capacity.

7. The multi-stage elastic scheduling method for gas-electric coupled energy systems considering source-load flexible coordination according to claim 2, characterized in that, The dynamic pipeline storage of the natural gas network is converted into equivalent electrical power reserve capacity through gas turbines and participates in optimized scheduling.

Images