Demand side response potential real-time calculation method based on flexible load aggregation

Abstract

The invention relates to a demand side response potential real-time calculation method based on flexible load aggregation, and the method comprises the steps: building a demand side maximum adjustable capacity calculation model based on flexible load aggregation through obtaining the operation load data, switch start-stop state data, environmental meteorological data and power time-sharing price of a flexible load in a region; and solving the demand side maximum adjustable capacity calculation model by constructing a flexible load refined prediction model and a switch state transition probability evaluation model, and finally obtaining an upper limit value and a lower limit value of the demand side flexible load adjustable capacity. The method can improve the effective utilization of the flexible load, assists the realization of the dual-carbon target, promotes the transformation and upgrading of the regional power grid to the energy internet, and has a great practical value.

Description

Technical Field

[0001] The present invention belongs to the technical field of power load aggregation and multi-source collaboration, and specifically relates to a real-time calculation method for demand-side response potential based on flexible load aggregation. Background Art

[0002] Currently, with the accelerated construction of new power systems, modern power systems are gradually evolving towards source-load interaction and the large-scale integration of renewable energy. This presents a trend of increasingly abundant power resources and flexible and volatile end-user load demands. On the one hand, the large-scale integration of renewable energy has led to significant spatial and temporal imbalances in the power balance, significantly increasing the pressure on peak and frequency regulation. Furthermore, the continued rise in peak loads has placed higher demands on the power system's regulation capabilities. The traditional generation-following-load model is no longer able to meet the needs of grid development. Furthermore, load-side equipment is becoming increasingly diverse. Large-scale, adjustable load resources, or flexible loads, are becoming key regulatory resources for scenarios such as peak load shaving, smoothing renewable energy fluctuations, and providing ancillary services, due to their large number, rapid response, and flexible control. On the demand side, there is a clear trend towards diversification of load-side equipment. Flexible loads such as air conditioners, electric vehicles, and distributed energy storage not only offer advantages such as large scale, rapid response, and high flexibility, but also have the disadvantages of small individual capacity, large number, and high randomness. This makes it difficult to accurately assess the demand-side response potential, hindering direct participation in power regulation.

[0003] Based on this, the present invention proposes a real-time calculation method for demand-side response potential that can overcome the above-mentioned defects. It is used to aggregate model flexible loads, calculate the maximum adjustable capacity in real time, guide demand-side flexible loads to deeply participate in the operation control of the distribution network, and collaboratively participate in dynamic demand-side response to address the above-mentioned problems.

[0004] It will be understood that the above statements merely provide background technology related to the present invention and do not necessarily constitute prior art. Summary of the Invention

[0005] The purpose of this invention is to provide a real-time calculation method for demand-side response potential based on flexible load aggregation, so as to improve the effective utilization of flexible loads, assist in the realization of dual carbon goals, and promote the transformation and upgrading of regional power grids to energy Internet.

[0006] To achieve the above-mentioned objectives, the present invention provides a real-time calculation method for demand-side response potential based on flexible load aggregation, comprising the following steps: S1, obtaining operating load data, switch start-stop status data, environmental meteorological data and electricity time-of-use prices of flexible loads in a region; S2, constructing a demand-side maximum adjustable capacity calculation model based on flexible load aggregation; S3, constructing a flexible load refined prediction model, and inputting the operating load data, environmental meteorological data and electricity time-of-use prices of each flexible load into the prediction model, and performing intraday ultra-short-term load prediction for each flexible load; S4, using a Markov chain state transition probability matrix, constructing a switch state transfer probability evaluation model, and inputting the switch start-stop status data of each flexible load into the evaluation model, and evaluating the switch state probability corresponding to each flexible load in different time periods; S5, based on the flexible load intraday ultra-short-term load prediction and switch state probability evaluation results, solving the maximum adjustable capacity calculation model constructed in S2 to obtain the upper and lower limits of the adjustable capacity of the demand-side flexible load.

[0007] Optionally, in S1, the flexible load includes a temperature control load, an electric vehicle, and an energy storage device.

[0008] Optionally, in S2, the establishment of a demand-side maximum adjustable capacity calculation model based on flexible load aggregation requires clarifying the operating scenario and switching status of the flexible load.

[0009] Optionally, the calculation formulas for the on / off states of the temperature control load, electric vehicle, and energy storage device are respectively expressed as:

[0010]

[0011] Where δ is the temperature offset acceptable to the user, T set is the set comfort temperature value, is the actual indoor temperature at time t; TC, EV, and ES are the temperature control load, electric vehicle, and energy storage equipment in the flexible load, respectively; They are the switch start and stop status of the temperature control load, electric vehicle and energy storage equipment at time t.

[0012] Optionally, the calculation formula of the demand-side maximum adjustable capacity calculation model based on flexible load aggregation is:

[0013]

[0014] Where, and are the calculated values ​​of the maximum adjustable increase and decrease capacity of the flexible load respectively; N TC 、N EV and N ESare the corresponding load agent numbers for temperature control load, electric vehicle and energy storage equipment respectively; and They are respectively the predicted value of the operating power of the temperature control load at time t, the predicted value of the charging power of the electric vehicle, and the predicted value of the charging and discharging power of the energy storage device; are the start / stop status evaluation values ​​of the temperature control load, electric vehicle, and energy storage equipment at time t; s is the start / stop status of the flexible load switch, 1 represents the running or charging state, 0 represents the closed state, and -1 represents the discharging state.

[0015] Optionally, in S3, constructing the flexible load refined prediction model includes the following steps: S31, using a BiLSTM model to capture the bidirectional temporal dependency of the flexible load on the data, and obtaining the ultra-short-term load prediction results of each flexible load within the day; S32, using a sparrow search algorithm to optimize the parameters of the BiLSTM model in S31, and obtaining the ultra-short-term load prediction results of each flexible load within the day at different time scales.

[0016] Optionally, in S31, the input part of the BiLSTM model is historical environmental data, and the output part is the respective load forecast values ​​of the temperature control load, electric vehicle and energy storage equipment under the same time scale conditions.

[0017] Optionally, the Markov chain state transition probability matrix has convergence. Regardless of the initial state, as long as the state transition probability matrix does not change, after sufficient state transitions, the probability corresponding to the start and stop state of the flexible load will converge to a fixed value.

[0018] Optionally, the calculation formula for converging to a fixed value is:

[0019]

[0020] In the formula, x(t) and y(t) represent different initial states, Indicates that after infinite state transitions, S N×1 is the probability convergence value of different states.

[0021] Optionally, the temperature control load includes an electric water heater and a central air conditioner; the electric vehicle includes a pure electric vehicle and a plug-in hybrid vehicle; and the energy storage device includes a distributed energy storage device and a centralized energy storage device.

[0022] To sum up, compared with the existing technology, the present invention provides a real-time calculation method for demand-side response potential based on flexible load aggregation, which fully considers historical load, temperature and other environmental factors, and performs time-series prediction on the operating power of demand-side temperature-controlled loads, electric vehicles and energy storage equipment, and evaluates the switching status, thereby establishing a real-time maximum adjustable capacity calculation method for flexible loads based on load power and switching status, providing support for flexible loads to participate in dynamic demand-side response, improving the effective utilization of user-side load resources, helping to achieve the dual carbon goals, and promoting the transformation and upgrading of regional power grids to energy Internet. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Schematic diagram of the flow of the real-time calculation method of demand-side response potential based on flexible load aggregation in the present invention;

[0024] Figure 2 This is an optimization flow chart of the sparrow optimization algorithm in the present invention;

[0025] Figure 3 It is a diagram of the Markov chain state transition process in the present invention;

[0026] Figure 4 This is a diagram showing the total load prediction results of flexible loads in a low-carbon park in one embodiment of the present invention;

[0027] Figure 5 Schematic diagram of the maximum increase and decrease capacity values ​​after flexible load aggregation in a low-carbon park in one embodiment of the present invention;

[0028] Figure 6 These are the upper and lower limits of the total capacity after flexible load aggregation in a low-carbon park in one embodiment of the present invention. DETAILED DESCRIPTION

[0029] The following is combined with Figure 1 ~Attached Figure 6 , the present invention is further explained by describing a preferred specific embodiment in detail.

[0030] It should be noted that the drawings are in a very simplified form and use non-precise proportions. They are only used to conveniently and clearly assist in explaining the embodiments of the present invention, and are not used to limit the conditions for the implementation of the present invention. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical content disclosed in the present invention without affecting the efficacy and purpose that can be achieved by the present invention.

[0031] It should be noted that, in the present invention, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only the elements explicitly listed, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0032] like Figure 1 As shown, the present invention provides a real-time calculation method for demand-side response potential based on flexible load aggregation, which specifically includes the following steps:

[0033] S1. Obtaining operating load data, switch start / stop status data, environmental meteorological data, and time-of-use electricity prices for flexible loads in a region; wherein the flexible loads include temperature-controlled loads, electric vehicles, and energy storage devices;

[0034] S2. Construct a demand-side maximum adjustable capacity calculation model based on flexible load aggregation;

[0035] S3. Construct a refined forecasting model for flexible loads, input the operating load data, environmental meteorological data, and time-of-use electricity price of each flexible load into the forecasting model, and perform intraday ultra-short-term load forecasting for each flexible load;

[0036] S4. Using the Markov chain state transition probability matrix, a switch state transition probability evaluation model is constructed. The switch start and stop state data of each flexible load is input into the evaluation model to evaluate the switch state probability of each flexible load at different time periods.

[0037] S5. Based on the ultra-short-term load forecast and switch state probability assessment results of the flexible load within a day, the maximum adjustable capacity calculation model constructed in S2 is solved to obtain the upper and lower limits of the demand-side flexible load adjustable capacity.

[0038] Furthermore, in S1, the temperature control load includes an electric water heater and a central air conditioner; the electric vehicle includes a pure electric vehicle and a plug-in hybrid vehicle; and the energy storage device includes a distributed energy storage device and a centralized energy storage device.

[0039] Furthermore, in S2, the establishment of a demand-side maximum adjustable capacity calculation model based on flexible load aggregation requires clarifying the flexible load's operating scenarios and on / off states. For flexible loads such as temperature-controlled loads and electric vehicles, load reduction scenarios are not possible when they are off. Load reduction can only be achieved by shutting down the load for a short period of time when on, or by starting the device when off. Energy storage devices, on the other hand, possess bidirectional energy exchange capabilities, allowing for flexible switching between charging and discharging modes based on grid demand.

[0040] Specifically, the calculation formulas for the on / off states of the temperature control load, electric vehicle, and energy storage device are respectively expressed as:

[0041]

[0042] Where δ is the temperature offset acceptable to the user, T set is the set comfort temperature value, is the actual indoor temperature at time t; TC, EV, and ES are the temperature control load, electric vehicle, and energy storage equipment in the flexible load, respectively; They are the switch start and stop status of the temperature control load, electric vehicle and energy storage equipment at time t.

[0043] Based on the above content, the demand-side maximum adjustable capacity calculation model based on flexible load aggregation can be calculated as follows:

[0044]

[0045] Where, and are the calculated values ​​of the maximum adjustable increase and decrease capacity of the flexible load respectively; N TC 、N EV and N ES are the corresponding load agent numbers for temperature control load, electric vehicle and energy storage equipment respectively; and They are respectively the predicted value of the operating power of the temperature control load at time t, the predicted value of the charging power of the electric vehicle, and the predicted value of the charging and discharging power of the energy storage device; are the start / stop status evaluation values ​​of the temperature control load, electric vehicle, and energy storage equipment at time t; s is the start / stop status of the flexible load switch, 1 represents the running or charging state, 0 represents the closed state, and -1 represents the discharging state.

[0046] Furthermore, in S3, constructing the flexible load refined prediction model specifically includes the following steps:

[0047] S31. Use the BiLSTM (bidirectional long short-term memory) model to capture the bidirectional temporal dependency of flexible loads on data and obtain the intraday ultra-short-term load forecast results for each flexible load.

[0048] S32. Use SSA (Sparrow Search Algorithm) to optimize the parameters of the BiLSTM model in S31 to obtain the intraday ultra-short-term load forecast results of each flexible load under different time scales.

[0049] In the above S31, the refined load forecasting model performs intraday ultra-short-term load forecasting on the temperature control load, electric vehicle and energy storage equipment based on the bidirectional long short-term memory network (BiLSTM), and obtains the predicted value of the temperature control load operating power at time t. Predicted value of electric vehicle charging power and the predicted value of the charging and discharging power of the energy storage device

[0050] The cyclic state of the basic unit of the long short-term memory network can be expressed as:

[0051] h t =f(h t-1 ,x t )

[0052] Where h t represents the hidden layer state at time t, x t Represents the input vector value at time t, h t-1 Represents the hidden layer state at time t-1.

[0053] The calculation process of iterative update of the basic unit state of the long short-term memory network is:

[0054] i t =σ(W i [h t-1 ,x t ]+b i )

[0055] f t =σ(W f [h t-1 ,x t ]+b f )

[0056] o t =σ(W o [h t-1 ,x t ]+b o )

[0057] g t =tanh(W g [h t-1 ,x t ]+bg )

[0058] C t =f t C t +i t g t

[0059] h t =o t C t +tanC t

[0060] Where i t 、f t and o t They are the input gate, forget gate and output gate states respectively, W and b are the parameters of the LSTM basic unit respectively, C t is the current basic unit state, g t is the new candidate value for the state of the basic unit.

[0061] It should be noted that in this embodiment, the input of the bidirectional long short-term memory network is historical load, temperature and other environmental data, and the forward and reverse LSTM are introduced to process the forward and reverse data flows respectively, and the risk of underfitting is reduced through the weight sharing mechanism. The output of the bidirectional long short-term memory network is the respective load forecast values ​​of the temperature control load, electric vehicle and energy storage equipment under the same time scale conditions, which are respectively and

[0062] In the S32, as Figure 2 As shown, the sparrow search algorithm (SSA) includes the following steps:

[0063] Step S00: Initialize the sparrow population: a population of n sparrows is formed. If the dimension of the variable to be optimized is m, the population can be expressed as:

[0064]

[0065] Step S01: Calculate fitness and sort: Sparrows with strong adaptability will get food first, and will also attract other individuals to join and expand the search range. The corresponding fitness is expressed as:

[0066]

[0067] Step S02: Update the discoverer's position: The discoverer continuously updates its position. The iterative process can be expressed as:

[0068]

[0069] in, is the current position of the finder, d is the number of iterations, α is a random number, iter max is the maximum number of iterations, Q is a random number that follows a normal distribution, R2 is a warning value, and ST is a safety value.

[0070] Step S03: Update the joiner position: When the discoverer searches for better food, other sparrows will join in the fight. If they succeed, they become joiners and update their positions as follows:

[0071]

[0072] in, is the global worst position, is the optimal position of the finder in the iteration. When i>n / 2, it means that the adaptability of the i-th joiner is poor and it needs to fly to other locations to find food.

[0073] Step S04: Update the position of alert sparrows: In a group of sparrows, there is a group of alert sparrows that will notice the threats around them. They will inadvertently increase their vigilance to protect their own population.

[0074]

[0075] in, is the current global optimal position, β is the step size control parameter, f i and f g is the adaptation value between the current sparrow position and the global optimal position, f w is the global worst adaptation value, K is a random number in [-1,1], and the random number ε is close to 0.

[0076] Step S05: Calculate the fitness value and update the sparrow position to determine whether the stop condition is met. If so, exit and output the result; otherwise, repeat steps 2-6.

[0077] Furthermore, in the S4, as Figure 3 As shown, assuming that the three states of the Markov chain are Q1, Q2, and Q3, it is known that represents the state at time t. If the state at time t+1 is Q1, then the state transition probability is A 21 ; If it is Q2, then the state transition probability is A 22 ; If it is Q3, then the state transition probability is A 23 .

[0078] Among them, if the three states Q1, Q2, and Q3 are randomly converted, the corresponding state conversion probability matrix, that is, the switch state transition probability evaluation model, can be expressed as:

[0079]

[0080] Among them, the Markov chain state transition probability matrix can be used to construct the switching state transition probability of the flexible load in different time periods and evaluate the switching state of the flexible load in different time periods;

[0081] It can be understood that the input of the switch state transition probability evaluation model is the switch start and stop state data of each flexible load at different time periods, which is a random discrete process. Therefore, in this embodiment, with a sufficient number of statistical samples, the state transition frequency of the flexible load is used to replace the probability by the law of large numbers and conditional probability, as shown in the following formula:

[0082]

[0083] In this embodiment, taking the energy storage device as an example, the switch state transition probability evaluation model is constructed as follows:

[0084]

[0085] And because the Markov chain state transition probability matrix has convergence, no matter what the initial state is, as long as the state transition probability matrix T ES If there is no change, then after enough state transitions, the probability corresponding to the start and stop state of the energy storage device will converge to a fixed value, as follows:

[0086]

[0087] Where, It is the initial start and stop state of the energy storage device.

[0088] Similarly, as long as the state transition probability matrix T does not change, that is, the switch state transition probability evaluation model of each type of flexible load does not change, then after a sufficient number of switch start-stop state transitions, the probabilities corresponding to the start-stop states of the three types of flexible loads will converge to fixed values ​​and are independent of the initial state, that is:

[0089]

[0090] In the formula, x(t) and y(t) represent different initial states, Indicates that after infinite state transitions, S N×1 is the probability convergence value of different states.

[0091] It is understandable that the probability convergence value of the flexible load switch start and stop state is used for the switch state evaluation value corresponding to the flexible load in different time periods. In the above S4, the start and stop state evaluation value of the temperature control load at time t can be obtained respectively. Evaluation value of start-stop state of electric vehicles and the start-stop status evaluation value of the energy storage equipment

[0092] Furthermore, in said S5, the predicted value of the temperature control load operating power at time t obtained in said S3 is respectively Predicted value of electric vehicle charging power and the predicted value of the charging and discharging power of the energy storage device And the start-stop state evaluation value of the temperature control load at time t obtained in S4 Evaluation value of start-stop state of electric vehicles and the start-stop status evaluation value of the energy storage equipment Substitute it into the demand side maximum adjustable capacity calculation model in S2, and the calculated value of the maximum adjustable increase of flexible load is obtained. and the calculated value of the maximum adjustable capacity

[0093] In a preferred embodiment of the present invention, a low-carbon demonstration park is used as an example. First, different proportions of temperature-controlled loads, electric vehicles, and energy storage devices are configured according to demand, and the operating states of each flexible load are independent of each other.

[0094] Secondly, after obtaining the operating load data, switch start and stop status data, and ambient meteorological data for flexible loads such as temperature-controlled loads, electric vehicles, and energy storage equipment within the low-carbon park, the model then inputs historical load data, temperature data, electricity prices, incentives, and other time-series data into the flexible load refinement forecast model for standardization and "sliding window" processing. The model then performs intraday ultra-short-term forecasts for each of the three flexible load types mentioned above. The electricity price adopts a time-of-use pricing strategy, and the incentive coefficients for flexible loads participating in load regulation vary at different times, resulting in different compensation prices (yuan / kW·h).

[0095] Further, such as Figure 4 As shown in the figure, the horizontal axis is the time axis with a time granularity of 1 hour, and the vertical axis is the total load forecast for the day, which is the sum of the three types of flexible load forecast values. This shows that the above three types of flexible load components have typical differences in their proportions and power change trends, and the total load exhibits a "bimodal" characteristic.

[0096] Further, such as Figure 5 As shown, the horizontal axis represents time, with a 1-hour granularity, and the vertical axis represents the maximum increase and decrease in capacity after flexible load aggregation. This shows that the maximum daily capacity adjustment for flexible load aggregation is approximately -2000 to 2000 kW, but the adjustment capacity varies significantly across different time periods. The evening peak period has the greatest potential for capacity adjustment, indicating that electric vehicles and energy storage are more flexible in charging and discharging during this period, and their demand response is faster.

[0097] Further, such as Figure 6As shown in the figure, the horizontal axis is the time axis with a time granularity of 1 hour, and the vertical axis is the upper and lower limits of the total capacity after flexible load aggregation. This shows that flexible load aggregation can flexibly adjust the capacity threshold in real time, following the dispatch plan in real time within the safe range of the upper and lower limits. This is especially true during peak hours, when demand-side response potential is the strongest.

[0098] To sum up, the present invention provides a real-time calculation method for demand-side response potential based on flexible load aggregation, which fully considers historical load, temperature and other environmental factors, and performs time-series prediction on the operating power of demand-side temperature-controlled loads, electric vehicles and energy storage equipment, and evaluates the switch status, thereby establishing a real-time maximum adjustable capacity calculation method for flexible loads based on load power and switch status, providing support for flexible loads to participate in dynamic demand-side response, improving the effective utilization of user-side load resources, helping to achieve the dual carbon goals, and promoting the transformation and upgrading of regional power grids to energy Internet.

[0099] Although the present invention has been described in detail through the above preferred embodiments, it should be understood that the above description is not intended to limit the present invention. After reading the above description, various modifications and substitutions of the present invention will become apparent to those skilled in the art. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims

1. A real-time calculation method for demand-side response potential based on flexible load aggregation, characterized in that: The following steps are involved: S1. Obtaining operating load data, switch start / stop status data, environmental meteorological data, and time-of-use electricity prices for flexible loads in a region; S2. Construct a demand-side maximum adjustable capacity calculation model based on flexible load aggregation; S3. Construct a refined forecasting model for flexible loads, input the operating load data, environmental meteorological data, and time-of-use electricity price of each flexible load into the forecasting model, and perform intraday ultra-short-term load forecasting for each flexible load; S4. Using the Markov chain state transition probability matrix, a switch state transition probability evaluation model is constructed. The switch start and stop state data of each flexible load is input into the evaluation model to evaluate the switch state probability of each flexible load at different time periods. S5. Based on the ultra-short-term load forecast and switch state probability assessment results of the flexible load, the maximum adjustable capacity calculation model constructed in S2 is solved to obtain the upper and lower limits of the demand-side flexible load adjustable capacity.

2. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 1, characterized in that: In the above-mentioned S1, the flexible load includes a temperature-controlled load, an electric vehicle, and an energy storage device.

3. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 2, characterized in that: In S2, the establishment of a demand-side maximum adjustable capacity calculation model based on flexible load aggregation requires clarifying the operating scenario and switching status of the flexible load.

4. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 3, characterized in that: The calculation formulas for the switching states of the temperature control load, electric vehicle and energy storage device are respectively expressed as: Where δ is the temperature offset acceptable to the user, T set is the set comfort temperature value, is the actual indoor temperature at time t; TC, EV and ES are temperature control load, electric vehicle and energy storage equipment in flexible load respectively; They are the switch start and stop status of the temperature control load, electric vehicle and energy storage equipment at time t.

5. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 4, characterized in that: The calculation formula of the demand-side maximum adjustable capacity calculation model based on flexible load aggregation is: Where, and are the calculated values ​​of the maximum adjustable increase and decrease capacity of the flexible load respectively; N TC 、N EV and N ES are the corresponding load agent numbers for temperature control load, electric vehicle and energy storage equipment respectively; and They are respectively the predicted value of the operating power of the temperature control load at time t, the predicted value of the charging power of the electric vehicle, and the predicted value of the charging and discharging power of the energy storage device; are the start / stop status evaluation values ​​of the temperature control load, electric vehicle, and energy storage equipment at time t; s is the start / stop status of the flexible load switch, 1 represents the running or charging state, 0 represents the closed state, and -1 represents the discharging state.

6. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 2, characterized in that: In the above-mentioned S3, constructing the flexible load refined prediction model includes the following steps: S31. Use the BiLSTM model to capture the bidirectional temporal dependency of flexible loads on data and obtain the intraday ultra-short-term load forecast results for each flexible load. S32. Use the sparrow search algorithm to optimize the parameters of the BiLSTM model in S31 to obtain the intraday ultra-short-term load forecast results of each flexible load at different time scales.

7. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 6, characterized in that: In the above-mentioned S31, the input part of the BiLSTM model is historical environmental data, and the output part is the respective load forecast values ​​of the temperature control load, electric vehicle and energy storage equipment under the same time scale conditions.

8. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 2, characterized in that: The Markov chain state transition probability matrix has convergence. Regardless of the initial state, as long as the state transition probability matrix does not change, the probability corresponding to the start and stop state of the flexible load will converge to a fixed value after a sufficient number of state transitions.

9. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 8, characterized in that: The calculation formula for converging to a fixed value is: In the formula, x(t) and y(t) represent different initial states, Indicates that after infinite state transitions, S N×1 is the probability convergence value of different states.

10. The method for real-time calculation of demand-side response potential based on flexible load aggregation according to claim 2, characterized in that: The temperature control loads include electric water heaters and central air conditioners; the electric vehicles include pure electric vehicles and plug-in hybrid vehicles; and the energy storage devices include distributed energy storage devices and centralized energy storage devices.

Images