A multi-energy coupling-based virtual power plant optimal operation method

Abstract

This invention relates to the field of power optimization technology, specifically disclosing a virtual power plant optimization operation method based on multi-energy coupling. The method includes: extracting the thermal damping characteristics of each target node in a building complex and calculating the energy repayment demand; simulating the power superposition effect of simultaneous startup of multiple nodes, verifying safety constraints by combining grid transformer capacity limitations and ramp rate, and extracting the rebound risk time window; analyzing the user's thermal comfort tolerance limit and energy repayment demand, and dividing the heterogeneous recovery echelons; performing time-series optimization on each echelon to determine the optimal start-stop time group; analyzing latent cooling potential energy in real time and performing conflict analysis with remaining sensible heat demand, and predictively intercepting and dynamically reducing oversaturated redundant nodes. This invention helps to smooth out the rebound peak during the load recovery period and utilizes building thermal inertia, ensuring both grid safety and user comfort.

Description

Technical Field

[0001] This invention relates to the field of power optimization technology, and more specifically to a method for optimizing the operation of a virtual power plant based on multi-energy coupling. Background Technology

[0002] In modern power systems, virtual power plants participate in the grid's peak shaving and valley filling response strategies. When the grid faces peak load pressure, virtual power plants typically issue load reduction or interruption commands to temperature-controlled loads within their jurisdiction to alleviate grid supply and demand tensions. However, this command-based load control mechanism has potential vulnerabilities in the recovery phase after the response ends.

[0003] Because many temperature control devices are shut down or operating at reduced power during the response period, a significant heat (or cooling) deficit accumulates inside the building, causing the indoor temperature to deviate from the comfort range. Upon release of the response, driven by the thermostat's hysteresis logic, numerous heterogeneous devices within the building often start up simultaneously at full load to fill the heat deficit. This disordered synchronous recovery behavior can create huge instantaneous power spikes within a very short time, leading to transformer overload, voltage drops, and even relay protection tripping in the distribution substation, affecting the safe and stable operation of the power distribution network.

[0004] Therefore, this invention provides a method for optimizing the operation of a virtual power plant based on multi-energy coupling. Summary of the Invention

[0005] The purpose of this invention is to provide a method for optimizing the operation of a virtual power plant based on multi-energy coupling, so as to solve the above-mentioned background problems.

[0006] The objective of this invention can be achieved through the following technical solutions: A method for optimizing the operation of a virtual power plant based on multi-energy coupling includes the following steps: The thermal attenuation parameters of each target node in the building complex are collected during the response period. The thermal damping characteristics of the thermal attenuation parameters are extracted to obtain a thermal inertia feature set. Based on the thermal inertia feature set, the equivalent conversion of heat deficit for end users is performed to obtain the energy repayment demand value of the target node. Based on the repayment demand value, the power superposition effect under the synchronous start-up of multiple nodes is simulated, a rebound peak curve is generated, and the safety constraint verification of transformer capacity and curve ramp rate is performed respectively to establish the time window of rebound risk. Based on the time window of rebound risk, a demand assessment is conducted on the user's thermal comfort tolerance limit and energy repayment requirement value to establish a recovery urgency matrix; based on the recovery urgency matrix, the delayed response potential of different target nodes is identified, and a recovery echelon list with different priorities is established. By combining the maximum allowable ramp rate of the power grid with the total energy repayment demand, power planning is performed to obtain the target power evolution trajectory; time-series following optimization based on evolution trajectory approximation is performed on the heterogeneous recovery echelon to determine the optimal start-stop time group of the target node.

[0007] As a further technical solution of the present invention, the equivalent conversion is performed as follows: Obtain the temperature difference gap between the actual indoor temperature and the critical comfort threshold at the end of the response; Extract the equivalent heat capacity and equivalent thermal resistance from the thermal inertia feature set, and multiply the equivalent heat capacity with the temperature difference gap value to obtain the required total static sensible heat. Predict the intrusive heat flow within the recovery time window based on equivalent thermal resistance; The dynamic heat loss is obtained by integrating the intrusive heat flow over the time axis. The total static sensible heat and the dynamic heat loss are summed to obtain the total heat energy deficit. Obtain the real-time comprehensive energy efficiency ratio of the air conditioning equipment under the current operating conditions, and establish an electrical-thermal equivalent conversion logic based on the total heat energy deficit to obtain the energy repayment demand value.

[0008] As a further technical solution of the present invention, the process of performing the security constraint verification is as follows: Numerical traversal analysis was performed on the theoretical rebound peak curve, and capacity over-limit detection and rate over-limit detection were performed respectively to obtain the overload risk time set and the impact risk time set. By performing a union operation on the overload risk time set and the impact risk time set, a discrete time point set containing all potential dangerous moments is obtained; Extract the earliest time from the set of discrete time points as the starting point of the window, and extract the latest time as the ending point of the window; Establish a time interval based on the start and end points of the window, which is the time window for rebound risk.

[0009] As a further technical solution of the present invention: the method for performing the rate limit exceedance detection is as follows: Perform a first-order differential operation on the theoretical rebound peak curve to obtain the instantaneous ramp rate curve; Establish velocity collision inequalities, select the time points of velocity collision inequalities, and form a time set of impact risk.

[0010] As a further technical solution of the present invention, the method for performing the requirement analysis is as follows: Obtain the real-time indoor temperature values ​​of each target node within the time window of rebound risk, and establish the critical comfort boundary of the target node; The thermal tolerance is obtained by calculating the residual physical temperature difference between the real-time indoor temperature value and the critical boundary of comfort. Collect the energy repayment demand value of each target node, perform reverse normalization mapping on the heat tolerance index to obtain the normalized urgency value, and perform forward normalization mapping on the energy repayment demand value to obtain the normalized demand value. A two-dimensional feature vector of heat and electricity is constructed using normalized demand value as the horizontal axis feature and normalized urgency value as the vertical axis feature. Arrange the feature vectors of each target node in rows to construct a dataset representing the global recovery situation, which is the recovery urgency matrix.

[0011] As a further technical solution of the present invention: the process of performing the timing follow-up optimization is as follows: Extract the rigid recovery group, elastic adjustment group, and deep delay group from the recovery echelon list, and use the discrete time-shift traversal matching strategy to obtain the start and stop times of the rigid recovery group and the optimal delay amount of the elastic adjustment group and the deep delay group. Based on the start and stop times of the rigid recovery group, the optimal delay amounts of the elastic adjustment group and the deep delay group are used to dynamically plan the start and stop times, thereby obtaining the optimal start and stop time groups for different target nodes.

[0012] As a further technical solution of the present invention: the process of performing the discrete time-shift traversal matching is as follows: The start-up time of all target nodes in the rigid recovery group is forcibly locked at the zero time of disordered recovery. The transient power response function constructed during the power superposition effect analysis is called to calculate the aggregated power curve after superposition of all nodes in the rigid recovery group, which serves as the basic superposition layer. Establish temporary test curves for the elastic adjustment group and optimize the delay to obtain the optimal delay amount for the elastic adjustment group, and update the base overlay layer; Trajectory deviation analysis is performed on the deep delay group based on the updated base overlay to obtain the cumulative absolute deviation value between the total power curve and the target power evolution trajectory. Based on the cumulative absolute deviation value, the optimal delay amount of the deep delay group is obtained.

[0013] As a further technical solution of the present invention: The system executes the optimal start-stop timing group and collects the latent cooling potential energy and remaining sensible heat demand in the put-in electrical energy in real time. It performs conflict analysis on the latent cooling potential energy and remaining sensible heat demand, identifies the target nodes in the optimal start-stop timing group that have not executed actions, and performs predictive interception.

[0014] As a further technical solution of the present invention: the method for obtaining the latent cooling potential energy is as follows: Within the execution cycle of the optimal start-stop time group, a two-way monitoring channel is established through the smart meter and temperature sensor network to collect the cumulative power consumption data of the started nodes and the actual indoor temperature drop data in real time. An algorithm combining time integration and difference operations is used to calculate the latent cooling potential energy.

[0015] As a further technical solution of the present invention, the method for performing conflict analysis is as follows: The preset heat saturation confidence coefficient is used to multiply the remaining sensible heat demand with the heat saturation confidence coefficient to generate a corrected demand judgment threshold. The latent cooling potential energy is compared with the demand judgment threshold. When the value of the latent cooling potential energy is greater than or equal to the corrected demand judgment threshold, the current operating state is marked as a false scarcity state. Based on the trigger signal of the false scarcity state, the system immediately scans the instruction queue of the optimal start and stop time group to identify the remaining target nodes that are arranged after the current time and have not yet issued a start instruction.

[0016] The beneficial effects of this invention are as follows: 1. By extracting the thermal damping characteristics of thermal attenuation parameters and performing dynamic balance analysis of thermal resistance and capacitance, the thermal inertia and insulation capacity of the building envelope within the complex are quantified. Combining real-time energy efficiency ratio and cold start correction factor, the thermodynamic temperature gap value is transformed into an electrical energy repayment demand value. By simulating the power superposition effect under multi-node synchronous startup, a theoretical rebound peak curve is generated, which is beneficial for reflecting the grid impact patterns that may occur under disordered recovery conditions. Through safety constraint verification, the system can lock in time windows containing capacity and rate over-limit risks in advance. This facilitates moving potential overload hazards from the real-time operation stage to the planning stage, enabling the dispatching system to identify continuous time intervals that the system cannot withstand before physical actions occur, thereby reducing the probability of triggering relay protection actions or causing voltage drops during batch load recovery.

[0017] 2. By constructing a recovery urgency matrix, target nodes are divided into three heterogeneous tiers: rigid recovery, elastic adjustment, and deep latency, identifying the differentiated characteristics of different users in terms of thermal comfort and response time. This facilitates ensuring priority access for highly sensitive users while tapping into the latency response potential of less sensitive users.

[0018] 3. By combining the safety evolution slope, a piecewise linear target power evolution trajectory is constructed to set an ideal path that conforms to power grid safety specifications for the load restoration process. A discrete time-shift ergonomic matching strategy is adopted to dynamically plan the start-up time of the heterogeneous restoration echelon, and the power increments of the elastic adjustment group and the deep delay group are used to fill the residual space between the target trajectory and the basic superposition layer. Based on the trajectory approximation time-series following optimization method, the originally disordered power surges are transformed into ordered controlled ramps, which helps to improve the frequency stability and voltage quality of the distribution network during load restoration.

[0019] 4. By calculating the latent cooling potential energy in real time and performing saturation collision analysis with the remaining sensible heat demand, the system can identify a false shortage state where energy input has already met future temperature drop requirements, even when temperature sensor readings have not yet reached the target. Implementing time window freezing and predictive interception on oversaturated redundant groups can suppress control overshoot caused by feedback lag, reducing unnecessary equipment start-ups and shutdowns and energy consumption. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart of a virtual power plant optimization operation method based on multi-energy coupling according to the present invention; Figure 2 This is a flowchart illustrating whether the system is in a false scarcity state in this invention; Figure 3 This is a functional block diagram of a virtual power plant optimization operation method based on multi-energy coupling in this invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] Example 1

[0024] like Figure 1 As shown, a method for optimizing the operation of a virtual power plant based on multi-energy coupling includes the following steps: S10. Collect the thermal attenuation parameters of each target node in the building complex during the response period, extract the thermal damping characteristics of the thermal attenuation parameters, and obtain the thermal inertia feature set; based on the thermal inertia feature set, perform equivalent conversion of heat deficit for end users to obtain the energy repayment demand value of the target node. The method for obtaining the thermal inertia feature set by collecting the thermal attenuation parameters of each target node in the building complex during the response period, extracting the thermal damping characteristics of the thermal attenuation parameters, and obtaining the feature set is as follows: In some preferred embodiments, the building complex within the virtual power plant area is decomposed into granular components. The smallest thermodynamic control unit with an independent physical space enclosure structure, independent temperature control execution equipment, and independent sensing and monitoring capabilities is defined as the target node, and a control set containing all target nodes is constructed. By using the temperature control monitoring device at the target node and the external meteorological data interface, the indoor and outdoor temperatures during the response period of the target node are collected synchronously, and indoor and outdoor temperature sequences are established. By aligning the indoor temperature series and the outdoor ambient temperature series on the time axis and calculating the difference, the driving force series of the indoor and outdoor temperature difference is obtained. Discrete difference operation is performed on the indoor temperature sequence to calculate the temperature drop rate per unit time and construct a temperature gradient sequence characterizing the building's natural cooling rate. The driving force sequence and temperature gradient sequence are used as thermal decay parameters; Based on the equivalent circuit principle of building heat transfer, a first-order thermal resistance-capacitance dynamic equilibrium model is constructed. The temperature gradient sequence is input as the observation vector, and the driving force sequence is input as the perturbation term to the dynamic equilibrium model. The output is the equivalent thermal resistance and equivalent heat capacity. It should be noted that the dynamic equilibrium model defines the rate of change of indoor temperature as being jointly determined by the external heat flow intruding through thermal resistance and the internal inertial heat flow released by heat capacity. Preferably, the dynamic equilibrium model is constructed by constructing a linear identification equation with the temperature gradient sequence as the observation vector and the indoor and outdoor temperature difference driving force sequence as the input vector. It should be noted that the linear identification equation is expressed in the following form: ; in, Indicates the discrete sampling time. This represents the temperature gradient sequence value at that moment. This represents the driving force sequence value at that moment. and All are regression coefficients to be identified; The coefficients to be estimated in the linear identification equation are iteratively converged online, and the converged coefficients are... The analysis is performed as physical parameters; The preferred analytical approach is to establish a set of inverse mapping equations between the regression coefficients and physical parameters based on the principle of thermodynamic energy conservation: , ; in, For equivalent thermal resistance, For equivalent heat capacity, The pre-set basic thermal power for the building interior; It should be noted that the base thermal power is determined as follows: the historical average load of the target node during the non-response period is retrieved, and the historical average load is equivalently determined as the base thermal power based on the principle of electrothermal equivalence; for example, the base thermal power is set to 500W. Based on the inverse mapping equations, the equivalent heat capacity characterizing the building's internal heat storage capacity is calculated: ; Next, calculate the equivalent thermal resistance, which characterizes the insulation capacity of the building envelope: ; This outputs the equivalent thermal resistance and equivalent heat capacity, and combines the equivalent thermal resistance and equivalent heat capacity to establish a thermal inertia characteristic set. The method for determining the energy repayment requirement of nodes by converting the heat deficit of end users into an equivalent value based on the thermal inertia feature set is as follows: Obtain the user-defined comfort threshold and calculate the temperature difference gap between the actual indoor temperature and the comfort threshold at the end of the response. By multiplying the equivalent heat capacity of the thermal inertia feature set with the temperature difference gap value, the total static sensible heat required to pull the building's internal entities (walls, air, furniture) back to the comfort threshold from the current temperature is calculated. Based on the equivalent thermal resistance, the heat flow that intrudes through the building envelope due to the continuous indoor-outdoor temperature difference within the recovery time window is predicted, and the intrusion heat flow is obtained. The dynamic heat loss is obtained by integrating the intrusive heat flow over the time axis. The total static sensible heat and the dynamic heat loss are summed to obtain the total heat energy deficit. Obtain the comprehensive energy efficiency ratio of the air conditioning equipment under the current operating conditions, and establish the electrical-thermal equivalent conversion logic in combination with the total heat energy deficit to obtain the energy repayment demand value. Preferably, the method for constructing the electro-thermal equivalent conversion logic is as follows: The total thermal energy deficit is divided by the overall energy efficiency ratio to obtain the total electricity demand. Then, the total electricity demand is divided by the duration of the preset recovery time window to calculate the average electrical power required to fill the thermal energy deficit, which is the energy repayment demand value of the target node. It should be noted that the preset recovery time window is defined as follows: by retrieving the standard control cycle issued by the virtual power plant master station, the standard control cycle is determined as the recovery period length; for example, the recovery period length is set to 15 minutes.

[0025] S20. Based on the repayment demand value, simulate the power superposition effect under the synchronous start-up of multiple nodes, generate the rebound peak curve, and perform safety constraint verification of transformer capacity and curve ramp rate respectively, and establish the time window of rebound risk. Among them, the method for generating the rebound peak curve based on the power superposition effect under the synchronous startup of multiple nodes, simulating the repayment demand value, is as follows: In some preferred embodiments, all target nodes within the virtual power plant's jurisdiction are configured to receive the recovery command at the same time, and the corresponding time is marked as the zero time of unordered recovery; Extract the energy repayment demand value of each target node, and retrieve the start-up response characteristic parameters of the associated equipment of each target node to establish the transient power response function of the individual equipment; Preferably, the transient power response function of a single device is established as follows: based on the physical inertial characteristics of the device during cold start-up, a first-order exponential growth model is used to describe the power increase trajectory over time; a transient power response function of the following form is constructed: ; in, This represents the relative time taken from the start time. For the first The target nodes are The instantaneous power value at a given moment; For the first Energy repayment requirement value (i.e., steady-state target power) of each target node. For the first The startup time constant of the devices associated with each target node; It should be noted that the startup time constant It is a physical parameter reflecting the response speed of the equipment. The value is defined as the length of time required for the equipment power to rise from zero to 63.2% of the steady-state power. The parameter is obtained by retrieving the data from the equipment's factory nameplate. Based on the zero-time of disordered recovery, the energy repayment demand value is expanded in the time domain using the transient power response function to obtain the independent power evolution trajectory of each target node. The independent power evolution trajectories of all target nodes are linearly superimposed on the time axis to obtain the theoretical rebound peak curve characterizing the dynamic change of the total regional load; It should be noted that the power superposition equation used to perform the linear superposition calculation is as follows: ; in, Indicates the time elapsed since time zero. for Theoretical value of total regional load at time . The total number of target nodes. For the first Energy repayment requirement value for each target node; For the first Normalized startup response coefficients of each target node; Analyze the normalized start-up response coefficients in the power superposition equation; The preferred analytical approach is as follows: Based on the physical inertial characteristics of equipment cold starts, establish the following exponential functional relationship: ; in, For the first The startup time constant of each target node device represents the characteristic time required for the device power to rise from zero to steady state. Calculate the theoretical value of the total regional load at each discrete moment within the recovery period, and connect the calculation results in chronological order to generate the theoretical rebound peak curve; The method for establishing a time window for rebound risk, which involves separately performing safety constraint checks on transformer capacity and curve ramp rate, is as follows: Obtain the rated capacity parameters of the transformer in the distribution area to which the target node belongs and the safety margin coefficient set by the relay protection device, and calculate the maximum allowable power threshold of the transformer. Obtain the maximum allowable ramp rate threshold for the load defined for the distribution area in the power grid dispatching specifications; Numerical traversal analysis was performed on the theoretical rebound peak curve, and capacity over-limit detection and rate over-limit detection were performed respectively to obtain the overload risk time set and the impact risk time set. Preferably, the method for performing capacity over-limit detection is to traverse every power point on the theoretical rebound peak curve; The capacity collision factor is obtained by multiplying the safety margin factor, the rated capacity of the transformer, and the power factor. The moment points on all rebound peak curves where the power point is higher than the capacity collision factor are selected to form the overload risk time set. Preferably, the method for detecting rate exceeding the limit is as follows: perform a first-order differential operation on the theoretical rebound peak curve to obtain the instantaneous ramp rate curve; Establish the velocity collision inequality: ; in, This is the maximum permissible ramp rate threshold; The moment points of the rate collision inequality are selected to form the impact risk time set; By performing a union operation on the overload risk time set and the impact risk time set, a discrete time point set containing all potential dangerous moments is obtained; Extract the earliest time from the set of discrete time points as the starting point of the window, and extract the latest time as the ending point of the window; The time interval that the system cannot withstand is established based on the start and end points of the window, which is the time window for rebound risk.

[0026] Example 2

[0027] Please see Figure 1 As shown, a method for optimizing the operation of a virtual power plant based on multi-energy coupling includes the following steps: S30. Based on the time window of rebound risk, conduct a demand assessment on the user's thermal comfort tolerance limit and energy repayment demand value, and establish a recovery urgency matrix; based on the recovery urgency matrix, identify the delayed response potential of different target nodes, and establish a recovery echelon list with different priorities. Among them, based on the time window of rebound risk, the method of assessing the user's thermal comfort tolerance limit and energy repayment demand value, and establishing the recovery urgency matrix is ​​as follows: Preferably, the real-time indoor temperature value of each target node within the time window of rebound risk is obtained; Obtain the preset temperature control operating parameters of the target node, and extract the user's target set temperature and the maximum allowable temperature drift threshold from the temperature control operating parameters; Based on the target set temperature and the maximum allowable temperature drift threshold, the critical boundary of comfort for the target node is established. It should be noted that the critical boundary of comfort is defined as follows: when in cooling mode, the boundary is the sum of the target set temperature and the maximum allowable temperature drift threshold; when in heating mode, the boundary is the target set temperature minus the maximum allowable temperature drift threshold. The thermal tolerance is obtained by calculating the residual physical temperature difference between the real-time indoor temperature value and the critical boundary of comfort. For example, the method for calculating the remaining physical temperature difference between the real-time indoor temperature value and the critical comfort boundary is as follows: set the target set temperature to... Temperature drift threshold set to Through the formula: Calculate the heat tolerance index ; in, This represents the current temperature drift that has been consumed. This indicates the remaining available temperature buffer space. This indicates that the current temperature is still within a comfortable range; the higher the value, the greater the tolerance potential. This indicates that the current temperature has exceeded or reached the comfort threshold, at which point... A forced correction to 0 or a very small positive number indicates that the tolerance potential has been exhausted; Collect the energy repayment demand value of each target node, perform reverse normalization mapping on the heat tolerance index, and perform forward normalization mapping on the energy repayment demand value to eliminate differences in data units. Preferably, the method for performing forward normalization mapping is as follows: using the Min-Max normalization method, the energy repayment demand value is mapped to... The interval is used to obtain the normalized demand value; The heat tolerance index is processed by reverse mapping normalization to obtain a normalized urgency value; It should be noted that the reverse mapping is to unify the logic that the larger the value, the more urgent the need; Preferably, the method for constructing the recovery urgency matrix is ​​as follows: using the normalized demand value as the horizontal axis feature and the normalized urgency value as the vertical axis feature, a two-dimensional feature vector of thermal and electrical components is constructed. Arrange the feature vectors of each target node in rows to construct a dataset representing the global recovery situation, which is the recovery urgency matrix; The method for identifying the latency response potential of different target nodes based on the recovery urgency matrix and dividing them into heterogeneous recovery tiers with different priorities is as follows: The recovery urgency matrix is ​​imported into a preset clustering algorithm. Based on the convergence principle of minimizing the weighted sum of squared errors within the cluster, the data points in the recovery urgency matrix are automatically clustered, and the membership degree of each target node relative to the cluster center is output. Those skilled in the art will understand that clustering algorithms set the number of clusters. (For example Based on the convergence principle of minimizing the sum of squared errors within a class, the data points in the recovery urgency matrix are automatically clustered, and the results for each target node relative to the sum of squared errors within the class are output. The membership degree values ​​of each cluster center; Based on the membership values ​​of the cluster centers, the positional distribution of the recovery urgency matrix data points on the thermal-electric two-dimensional feature plane is determined, and the physical semantics of different clusters are defined. Preferably, the physical semantics of different clusters are defined as follows: the set of nodes whose cluster centers are located in high-demand (normalized demand value close to 1) and high-urgency (normalized urgency value close to 1) regions is defined as the first tier, given the highest priority, and used as the rigid recovery group. The set of nodes whose cluster centers are located in areas of medium demand and medium urgency is defined as the second tier, given medium priority, and used as a flexible adjustment group; The set of nodes whose cluster centers are located in low-demand, low-urgency areas is defined as the third tier and assigned the lowest priority, serving as the deep delay group. Based on the principle of maximum membership, the echelon to which each target node belongs is determined, and a list of recovery echelons containing priority labels is generated.

[0028] S40. Combining the maximum allowable ramp rate of the power grid with the total energy repayment demand, power planning is performed to obtain the target power evolution trajectory; time-series following optimization based on evolution trajectory approximation is performed on the heterogeneous recovery echelon to determine the optimal start-stop time group of the target node; The method for obtaining the target power evolution trajectory by combining the maximum allowable grid ramp rate and the total energy repayment demand is as follows: In some preferred embodiments, the energy repayment demand values ​​of all target nodes are numerically accumulated to obtain the total global recovery power within the virtual power plant's jurisdiction; Extract the maximum allowable ramp rate threshold of the load and introduce a smoothing safety factor less than 1 (e.g., 0.8), and obtain the safety evolution slope through multiplication. Based on the global total recovery power and the safety evolution slope, a piecewise linear target power evolution trajectory is constructed; Preferably, the method for constructing the target power evolution trajectory is as follows: S401. Set the recovery start time as zero point, and construct a linear upward line on the time axis with a safe evolution slope and a fixed growth rate as the climbing guide segment; S402. Monitor the value of the linear rising line in real time. When the value of the line increases to equal the total global recovery power, keep the value constant and form a flat-top saturation segment. S403. The curve formed by the connection of the beginning and end of the climbing guidance section and the flat-top saturation section is used as the standard reference, which is the target power evolution trajectory. It should be noted that the physical meaning of the target power evolution trajectory is that the target power evolution trajectory is the optimal load recovery form expected by the power grid dispatching side, that is, to complete the power recovery at the fastest safe speed without triggering the relay protection action. Among them, the method for performing time-series following optimization based on evolutionary trajectory approximation on the heterogeneous recovery echelon to determine the optimal start and stop time vector of the target node is as follows: Extract the rigid recovery group, elastic adjustment group, and deep delay group from the recovery echelon list, and use the discrete time-shift traversal matching strategy to obtain the start and stop times of the rigid recovery group and the optimal delay amount of the elastic adjustment group and the deep delay group. Preferably, the process of performing discrete time-shift traversal matching is as follows: S411. Force the start-up time of all target nodes in the rigid recovery group to be locked at the zero time of disordered recovery, call the transient power response function constructed during the power superposition effect analysis, calculate the aggregate power curve after superposition of all nodes in the rigid recovery group, and use it as the basic superposition layer. S412. Establish a temporary test curve for the elastic adjustment group and perform delay optimization to obtain the optimal delay amount for the elastic adjustment group, and update the base overlay layer. Preferably, the method for establishing a temporary test curve for the elastic adjustment group, optimizing the delay, obtaining the optimal delay amount for the elastic adjustment group, and updating the base overlay layer is as follows: Set a time-shifted scan window (e.g., from 0 seconds to 300 seconds) and set a discrete scan step size (e.g., every 5 seconds); Within the time-shifted scanning window, the overall start time of the elastic adjustment group is shifted backward by one scanning step. At each shifted position, the aggregate power of the elastic adjustment group and the sum of the base superposition layer are calculated to generate a temporary test curve. Calculate the cumulative absolute deviation between each temporary test curve and the target power evolution trajectory over the entire time axis; By comparing the cumulative absolute deviation values ​​corresponding to all translation positions, the translation time corresponding to the smallest deviation value is selected as the optimal delay amount for the elastic adjustment group. Update the base overlay layer based on the optimal delay of the elastic adjustment group; S413. Based on the updated base superposition layer, perform trajectory deviation analysis on the deep delay group to obtain the cumulative absolute deviation value between the total power curve and the target power evolution trajectory. Based on the cumulative absolute deviation value, filter to obtain the optimal delay amount of the deep delay group. The preferred method for performing trajectory deviation analysis is as follows: On top of the updated base overlay, using the same traversal optimization logic, a sliding test is performed on the start time of the deep delay group within the time-shifted scan window. Calculate the cumulative absolute deviation between the total power curve after the deep delay group is superimposed and the target power evolution trajectory; It should be noted that the essence of trajectory deviation analysis is to use the power increment of the third echelon to fill the remaining space (i.e., residual) between the target trajectory and the base superposition layer. The translation duration corresponding to the smallest deviation value is selected as the optimal delay amount for the deep delay group; The method for locking the optimal start and stop time vector is as follows: Based on the start and stop times of the rigid recovery group, the optimal delay amounts of the elastic adjustment group and the deep delay group are dynamically planned to obtain the optimal start and stop time groups for different target nodes. The preferred method for dynamic programming of start and stop times is: Establish an instruction sequence table containing all target nodes; For target nodes belonging to the rigid recovery group, their start and stop times are set to the zero time of unordered recovery. For target nodes belonging to the flexible adjustment group, the start and stop times are set to the zero time of disordered recovery plus the optimal delay of the flexible adjustment group. For target nodes belonging to the deep delay group, the start and stop times are set to the zero time of unordered recovery plus the optimal delay amount of the deep delay group; The output includes the instruction sequence list of each target node in the rigid recovery group, elastic adjustment group, and deep delay group, which is the optimal start and stop time group for each target node.

[0029] S50: Execute the optimal start-stop time group and collect the latent cooling potential energy and remaining sensible heat demand in the put-in electrical energy in real time. Perform conflict analysis on the latent cooling potential energy and remaining sensible heat demand, identify the target nodes in the optimal start-stop time group that have not executed actions, and perform predictive interception. The method for executing the optimal start-stop timing group and collecting in real time the latent cooling potential energy and remaining sensible heat demand in the already invested electrical energy is as follows: The energy flow within the virtual power plant system is analyzed to extract the latent cooling potential energy of the emitted but not yet perceived by room temperature. The preferred method for parsing is: Within the execution cycle of the optimal start-stop time group, a two-way monitoring channel is established through the smart meter and temperature sensor network to collect the cumulative power consumption data of the started nodes and the actual indoor temperature drop data in real time. Based on the principle of energy conservation and the characteristics of heat conduction delay, an algorithm combining time integration and difference operations is used to calculate the latent cooling potential energy. It is understandable that the method for calculating the latent cooling potential energy is as follows: perform time-domain integration on the real-time power of the activated node in the current operating cycle, and multiply the integration result with the corrected comprehensive energy efficiency ratio in S10 to obtain the theoretically generated cumulative cooling capacity. Obtain the reference temperature at the start of recovery and the real-time temperature at the current sampling time, calculate the difference between the two values, and obtain the realized physical temperature drop. The realized physical temperature drop is multiplied by the equivalent heat capacity after correction of the transient thermal response coefficient to obtain the apparent sensible heat removal. Calculate the numerical residual between the total accumulated cooling volume and the apparent sensible heat removal volume as the latent cooling potential energy; It should be noted that latent cooling potential energy is a dynamically changing physical quantity. Its physical meaning is to characterize the amount of cold energy that is currently stagnant in the pipe medium, wall heat storage layer, and air convection dead zone and has not yet been converted into temperature sensor readings. The transient thermal response coefficient is a dimensionless value less than 1 (e.g., 0.05-0.1), used to extract the component that characterizes only the indoor air heat load capacity from the overall equivalent heat capacity of the building established by S10, thereby reducing the calculation deviation caused by the deep thermal inertia of the wall. Obtain the target comfort temperature set by the user, and calculate the remaining temperature difference gap between the real-time temperature at the current sampling time and the target comfort temperature; The remaining temperature difference gap value is multiplied by the equivalent heat capacity of the building envelope to quantify the remaining sensible heat requirement required to bring the indoor environment back to the target state. The method for conflict analysis of latent cooling potential energy and remaining sensible heat demand, identifying target nodes that have not yet performed actions in the optimal start-stop time group, and predictively intercepting them is as follows: Preset thermal saturation confidence coefficient; It should be noted that the thermal saturation confidence coefficient is a dimensionless scalar with a value less than 1, used to define the safety margin of the control. The remaining sensible heat demand is multiplied by the heat saturation confidence coefficient to generate the corrected demand determination threshold. like Figure 2 As shown, the latent cooling potential energy is compared with the demand judgment threshold. When the value of the latent cooling potential energy is greater than or equal to the corrected demand judgment threshold, the system triggers a state reversal and marks the current operating state as a false scarcity state. If the value of latent cooling potential is lower than the corrected demand judgment threshold, the latent cooling potential will be continuously monitored. If the latent cooling potential falls back to a certain percentage (e.g., 80%) below the demand judgment threshold, or if the indoor temperature rebounds significantly, the false shortage state will be lifted to prevent high-frequency oscillation of control commands. It should be noted that the false scarcity state reflects a specific thermodynamic transient: at this time, although the real-time temperature of the physical environment has not yet reached the user-set value (i.e., there is still demand on the surface), the energy that has been put into the system and is in the process of transmission is sufficient to cover the remaining temperature gap value. Continuing to put in new loads will lead to future temperature overshoot and energy waste. Based on the trigger signal of the false scarcity state, the system immediately scans the instruction queue of the optimal start and stop time group, identifies the remaining target nodes that are arranged after the current time and whose priority is lower than the preset threshold (such as the elastic adjustment group and the deep delay group), and establishes an oversaturated redundancy group. For oversaturated redundancy groups, time window freeze control commands are generated, and dynamic reduction processing is performed on optimal start and stop time groups. Preferably, the dynamic reduction process is performed as follows: Ignore the original time plan in the optimal start and stop time group, deprive the oversaturated redundancy group of its start permission, and lock the control status of the relevant nodes to forced standby mode to achieve predictive interception; The system remains locked in the forced standby mode until the latent cooling potential is fully released and the indoor temperature trajectory shows a trend of convergence or rebound towards the target value. Then, the system is unfrozen or re-planned based on the actual residual load. It should be noted that the dynamic reduction process reduces the grid rebound peak caused by overresponse by utilizing the thermal inertia hysteresis characteristics of the system without sacrificing the final thermal comfort.

[0030] Example 3

[0031] Please see Figure 3 As shown, a virtual power plant optimization operation system based on multi-energy coupling includes the following modules: Demand Analysis Module: Used to collect thermal attenuation parameters of each target node in the building complex during the response period, extract the thermal damping characteristics of the thermal attenuation parameters, and obtain a thermal inertia feature set; based on the thermal inertia feature set, perform equivalent conversion of heat deficit for end users to obtain the energy repayment demand value of the target node. Collision analysis module: Based on the repayment demand value, simulate the power superposition effect under the synchronous start-up of multiple nodes, generate the rebound peak curve, and perform safety constraint verification of transformer capacity and curve ramp rate respectively, and establish the time window of rebound risk; Response grading module: Based on the time window of rebound risk, it assesses the user's thermal comfort tolerance limit and energy repayment requirement, and establishes a recovery urgency matrix; based on the recovery urgency matrix, it identifies the delayed response potential of different target nodes and establishes a recovery tier list with different priorities. Follow-up optimization module: It is used to combine the maximum allowable ramp rate of the power grid with the total energy repayment demand to perform power planning and obtain the target power evolution trajectory; and to perform time-series follow-up optimization based on evolution trajectory approximation on the heterogeneous recovery echelon to determine the optimal start-up and shutdown time group of the target node. Predictive Interception Module: This module is used to execute the optimal start-stop time group and collect the latent cooling potential energy and remaining sensible heat demand in the put-in electrical energy in real time. It performs conflict analysis on the latent cooling potential energy and remaining sensible heat demand, identifies the target nodes in the optimal start-stop time group that have not executed actions, and performs predictive interception.

[0032] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for optimizing the operation of a virtual power plant based on multi-energy coupling, characterized in that, Includes the following steps: The thermal attenuation parameters of each target node in the building complex are collected during the response period. The thermal damping characteristics of the thermal attenuation parameters are extracted to obtain a thermal inertia feature set. Based on the thermal inertia feature set, the equivalent conversion of heat deficit for end users is performed to obtain the energy repayment demand value of the target node. Based on the repayment demand value, the power superposition effect under the synchronous start-up of multiple nodes is simulated, a rebound peak curve is generated, and the safety constraint verification of transformer capacity and curve ramp rate is performed respectively to establish the time window of rebound risk. Based on the time window of rebound risk, a demand assessment is conducted on the user's thermal comfort tolerance limit and energy repayment requirement value to establish a recovery urgency matrix; based on the recovery urgency matrix, the delayed response potential of different target nodes is identified, and a recovery echelon list with different priorities is established. By combining the maximum allowable ramp rate of the power grid with the total energy repayment demand, power planning is performed to obtain the target power evolution trajectory; time-series following optimization based on evolution trajectory approximation is performed on the heterogeneous recovery echelon to determine the optimal start-stop time group of the target node.

2. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 1, characterized in that: The equivalent conversion is performed as follows: Obtain the temperature difference gap between the actual indoor temperature and the critical comfort threshold at the end of the response; Extract the equivalent heat capacity and equivalent thermal resistance from the thermal inertia feature set, and multiply the equivalent heat capacity with the temperature difference gap value to obtain the required total static sensible heat. Predict the intrusive heat flow within the recovery time window based on equivalent thermal resistance; The dynamic heat loss is obtained by integrating the intrusive heat flow over the time axis. The total static sensible heat and the dynamic heat loss are summed to obtain the total heat energy deficit. Obtain the comprehensive energy efficiency ratio of the air conditioning equipment under the current operating conditions, and establish an electrical-thermal equivalent conversion logic based on the total heat energy deficit to obtain the energy repayment demand value.

3. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 1, characterized in that: The process of performing the aforementioned security constraint verification is as follows: Numerical traversal analysis was performed on the theoretical rebound peak curve, and capacity over-limit detection and rate over-limit detection were performed respectively to obtain the overload risk time set and the impact risk time set. By performing a union operation on the overload risk time set and the impact risk time set, a discrete time point set containing all potential dangerous moments is obtained; Extract the earliest time from the set of discrete time points as the starting point of the window, and extract the latest time as the ending point of the window; Establish a time interval based on the start and end points of the window, which is the time window for rebound risk.

4. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 3, characterized in that: The method for performing the rate limit violation detection is as follows: Perform a first-order differential operation on the theoretical rebound peak curve to obtain the instantaneous ramp rate curve; Establish velocity collision inequalities, select the time points of velocity collision inequalities, and form a time set of impact risk.

5. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 1, characterized in that: The method for conducting the aforementioned requirements assessment is as follows: Obtain the real-time indoor temperature values ​​of each target node within the time window of rebound risk, and establish the critical comfort boundary of the target node; The thermal tolerance is obtained by calculating the residual physical temperature difference between the real-time indoor temperature value and the critical boundary of comfort. Collect the energy repayment demand value of each target node, perform reverse normalization mapping on the heat tolerance index to obtain the normalized urgency value, and perform forward normalization mapping on the energy repayment demand value to obtain the normalized demand value. A two-dimensional feature vector of heat and electricity is constructed using normalized demand value as the horizontal axis feature and normalized urgency value as the vertical axis feature. Arrange the feature vectors of each target node in rows to construct a dataset representing the global recovery situation, which is the recovery urgency matrix.

6. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 1, characterized in that: The process of performing the timing follow-up optimization is as follows: Extract the rigid recovery group, elastic adjustment group, and deep delay group from the recovery echelon list, and use the discrete time-shift traversal matching strategy to obtain the start and stop times of the rigid recovery group and the optimal delay amount of the elastic adjustment group and the deep delay group. Based on the start and stop times of the rigid recovery group, the optimal delay amounts of the elastic adjustment group and the deep delay group are used to dynamically plan the start and stop times, thereby obtaining the optimal start and stop time groups for different target nodes.

7. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 6, characterized in that: The process of performing the discrete time-shift traversal matching is as follows: The start-up time of all target nodes in the rigid recovery group is forcibly locked at the zero time of disordered recovery. The transient power response function constructed during the power superposition effect analysis is called to calculate the aggregated power curve after superposition of all nodes in the rigid recovery group, which serves as the basic superposition layer. Establish temporary test curves for the elastic adjustment group and optimize the delay to obtain the optimal delay amount for the elastic adjustment group, and update the base overlay layer; Trajectory deviation analysis is performed on the deep delay group based on the updated base overlay to obtain the cumulative absolute deviation value between the total power curve and the target power evolution trajectory. Based on the cumulative absolute deviation value, the optimal delay amount of the deep delay group is obtained.

8. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 1, characterized in that: The system executes the optimal start-stop timing group and collects the latent cooling potential energy and remaining sensible heat demand in the put-in electrical energy in real time. It performs conflict analysis on the latent cooling potential energy and remaining sensible heat demand, identifies the target nodes in the optimal start-stop timing group that have not executed actions, and performs predictive interception.

9. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 8, characterized in that: The method for obtaining the latent cooling potential energy is as follows: Within the execution cycle of the optimal start-stop time group, a two-way monitoring channel is established through the smart meter and temperature sensor network to collect the cumulative power consumption data of the started nodes and the actual indoor temperature drop data in real time. An algorithm combining time integration and difference operations is used to calculate the latent cooling potential energy.

10. The method for optimizing the operation of a virtual power plant based on multi-energy coupling according to claim 8, characterized in that: The method for conducting conflict analysis is as follows: The preset heat saturation confidence coefficient is used to multiply the remaining sensible heat demand with the heat saturation confidence coefficient to generate a corrected demand judgment threshold. The latent cooling potential energy is compared with the demand judgment threshold. When the value of the latent cooling potential energy is greater than or equal to the corrected demand judgment threshold, the current operating state is marked as a false scarcity state. Based on the trigger signal of the false scarcity state, the system scans the instruction queue of the optimal start and stop time group to identify the remaining target nodes that are arranged after the current time and have not yet issued a start instruction.

Images