Power demand side multi-load peak regulation strategy optimization method and device

Abstract

The invention discloses a power demand side multi-load peak regulation strategy optimization method and device, and belongs to the field of power system operation scheduling. The method comprises the following steps: according to a scheduling instruction sent by a power system, calculating a load reduction amount of a demand side load for responding to the scheduling instruction; according to the user behavior factors in the response process, calculating a reliability index used for representing the demand side load effective response probability; according to the load reduction amount, calculating to obtain a cost index corresponding to reduction of each demand side load; according to the load evaluation index of the demand side load under the finite rationality condition, establishing a load multi-dimensional foreground model used for representing the foreground value of the load calling scheme; and according to the load multi-dimensional foreground model, establishing a multi-load response decision model taking a reliability index and a cost index as decision variables, and solving the multi-load response decision model to obtain an optimal peak regulation strategy of the demand side load. According to the invention, the applicability and reliability of the demand side peak regulation strategy can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation and dispatch, and particularly to a method and device for optimizing a multi-load peak shaving strategy on the power demand side. Background Art

[0002] As the dependence of power system peak shaving on demand response continues to increase, the impact of demand response reliability on the power system peak shaving effect cannot be ignored. Therefore, it is urgent for load aggregators to quantitatively analyze the response reliability of diversified loads to achieve precise peak shaving on the demand side. The demand side of industrial parks is rich in resources. Among them, the air-conditioning load has a fast response speed and large adjustment potential; the production load has a large capacity and a strong willingness to participate in demand response, and both are ideal resources for participating in power system peak shaving.

[0003] In the related art, existing research results all use a deterministic model to describe the load demand response process on the demand side, ignoring the influence of factors such as the incentive level and user electricity consumption habits on the peak shaving strategy on the demand side, and not considering the load response reliability either, resulting in low applicability and reliability of the peak shaving strategy on the demand side.

[0004] Based on this, there is an urgent need for a method and device for optimizing a multi-load peak shaving strategy on the power demand side to solve the above technical problems. Summary of the Invention

[0005] The present invention provides a method and device for optimizing a multi-load peak shaving strategy on the power demand side, which can effectively improve the applicability and reliability of the peak shaving strategy on the demand side. The technical solutions are as follows:

[0006] On the one hand, a method for optimizing a multi-load peak shaving strategy on the power demand side is provided. The method includes:

[0007] Calculating the load reduction amount of the demand side load for responding to the dispatch instruction according to the dispatch instruction issued by the power system; wherein, the demand side load includes the air-conditioning load and industrial load in the industrial park;

[0008] Calculating a reliability index for characterizing the effective response probability of the demand side load according to the user behavior factors of the demand side load during the response process;

[0009] Calculating a cost index corresponding to each reduction of the demand side load according to the load reduction amount;

[0010] Establishing a multi-dimensional load prospect model for characterizing the prospect value of the load calling scheme according to the load evaluation index of the demand side load under the condition of bounded rationality;

[0011] According to the multi-dimensional load prospect model, a multi-load response decision model with the reliability index and the cost index as decision variables is established, and the optimal peak shaving strategy of the demand-side load is obtained by solving the multi-load response decision model.

[0012] On the other hand, a device for optimizing the multi-load peak shaving strategy on the power demand side is provided. The device includes:

[0013] A first calculation module, configured to calculate the load reduction amount of the demand-side load for responding to the scheduling instruction according to the scheduling instruction issued by the power system; wherein, the demand-side load includes the air-conditioning load and the industrial load in the industrial park;

[0014] A second calculation module, configured to calculate a reliability index for characterizing the effective response probability of the demand-side load according to the user behavior factors of the demand-side load during the response process;

[0015] A third calculation module, configured to calculate the cost index corresponding to each type of the demand-side load reduced according to the load reduction amount;

[0016] A first modeling module, configured to establish a multi-dimensional load prospect model for characterizing the prospect value of the load invocation scheme according to the load evaluation index of the demand-side load under the condition of bounded rationality;

[0017] A second modeling module, configured to establish a multi-load response decision model with the reliability index and the cost index as decision variables according to the multi-dimensional load prospect model, and solve the multi-load response decision model to obtain the optimal peak shaving strategy of the demand-side load.

[0018] On the other hand, a computer device is provided. The computer device includes a memory and a processor. The memory is used to store a computer program, and the processor is used to execute the computer program stored on the memory to implement the steps of the above-mentioned method for optimizing the multi-load peak shaving strategy on the power demand side.

[0019] On the other hand, a computer-readable storage medium is provided. The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for optimizing the multi-load peak shaving strategy on the power demand side are implemented.

[0020] On the other hand, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for optimizing the multi-load peak shaving strategy on the power demand side are implemented.

[0021] The technical solution provided by the present invention can at least bring the following beneficial effects: First, the reduction amounts of diversified loads are diversified, and the corresponding cost indexes are calculated. At the same time, the influence of user behavior on the reliability of load response is evaluated by defining the load response reliability index. Second, a multi-dimensional prospect model of bounded rationality load is constructed according to multi-attribute decision-making and cumulative prospect theory. Then, taking the load response reliability and cost as the decision variables of the load aggregator together, a multi-load response decision model is constructed and the optimal peak shaving strategy is obtained by solving. This method can provide a basis for application scenarios such as the interactive decision-making between load aggregators and power systems, promote the wide application of bounded rational user behavior in the power field, improve the applicability and reliability of demand-side peak shaving strategies, and enhance the flexibility of the power system by precisely adjusting demand-side resources. Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to these drawings.

[0023] Figure 1 It is a flowchart of a method for optimizing the multi-load peak shaving strategy on the power demand side provided by an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the typical daily air-conditioning power curve provided by an embodiment of the present invention;

[0025] Figure 3 It is a cluster of air-conditioning power curves showing the preferences of two different types of air conditioners at different power consumption times provided by an embodiment of the present invention;

[0026] Figure 4 a and Figure 4 b are schematic diagrams of the cost value function and the reliability value function provided by an embodiment of the present invention in sequence;

[0027] Figure 5 It is a schematic diagram of the comparison of load curves before and after response provided by an embodiment of the present invention;

[0028] Figure 6 a and Figure 6 b are schematic diagrams of the load response of the load aggregator under two strategies when the peak shaving instruction is 8 MW provided by an embodiment of the present invention;

[0029] Figure 6 c and Figure 6 d are schematic diagrams of the load response of the load aggregator under two strategies when the peak shaving instruction is 20 MW provided by an embodiment of the present invention;

[0030] Figure 7 It is the structure diagram of the device for optimizing the multi - load peak - shaving strategy on the power demand side provided by an embodiment of the present invention;

[0031] Figure 8 It is the hardware architecture diagram of a computer device provided by an embodiment of the present invention. Detailed implementation manners

[0032] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0033] As mentioned above, most of the existing research on the demand - side peak - shaving strategy is optimized with the goal of maximizing the expected utility of users or minimizing the electricity cost, without considering the influence of users' subjective feelings on electricity consumption behavior, resulting in the peak - shaving strategy being difficult to meet the actual application needs.

[0034] Based on this, the concept of the present invention is to consider the reliability of load response in the demand - side peak - shaving strategy by studying users' response behavior, providing a basis for formulating a more perfect demand response strategy and improving the applicability of the demand - side peak - shaving strategy.

[0035] The following describes the specific implementation manners of the above concept.

[0036] Please refer to Figure 1 , a method for optimizing the multi - load peak - shaving strategy on the power demand side provided by an embodiment of the present invention, the method includes:

[0037] Step 100, calculate the load reduction amount of the demand - side load for responding to the scheduling instruction according to the scheduling instruction issued by the power system; wherein, the demand - side load includes the air - conditioning load and industrial load in the industrial park;

[0038] Step 102, calculate the reliability index for characterizing the effective response probability of the demand - side load according to the user behavior factors of the demand - side load during the response process;

[0039] Step 104, calculate the cost - related index corresponding to each reduction of the demand - side load according to the load reduction amount;

[0040] Step 106, establish a multi - dimensional load prospect model for characterizing the prospect value of the load invocation plan according to the load evaluation index of the demand - side load under the condition of bounded rationality;

[0041] Step 108: According to the multi-dimensional load prospect model, establish a multi-load response decision model with the reliability index and the cost index as decision variables, and solve the multi-load response decision model to obtain the optimal peak shaving strategy for the demand-side load.

[0042] The following describes Figure 1 the execution manners of the respective steps shown.

[0043] First, for step 100, according to the scheduling instruction issued by the power system, calculate the load reduction amount of the demand-side load for responding to the scheduling instruction.

[0044] According to the scheduling instruction issued by the power system, the load aggregator must regulate the reliable response of each user in the industrial park within 1 hour and precisely reduce the user load. Therefore, in the embodiments of the present invention, the air-conditioning load of the park users and the industrial load with short-term response capabilities are selected as the peak shaving resources, and a load response characteristic model is established to calculate the corresponding load reduction amount.

[0045] In the embodiments of the present invention, the load reduction amount is calculated through the following steps: According to the change relationship between the room temperature and the air-conditioning power, calculate the total air-conditioning power reduction amount before and after the response; According to the operation process of the equipment, determine the total power reduction amount of the production line load in the industrial load; wherein, the production line load includes light industrial machine tools and steelmaking electric arc furnaces, the total airport power reduction amount of the light industrial machine tools is determined by interrupting the discrete production process, and the total electric arc furnace power reduction amount of the steelmaking electric arc furnace is determined by adjusting the working gear of the electric arc furnace; Equivalent the intermediate product reserve to an energy storage model, and determine the total equivalent energy storage power reduction amount of the equivalent energy storage load of the standby link in the industrial load.

[0046] Specifically, variable-frequency air conditioners have the advantages of flexible use and high temperature control accuracy, and are commonly found in the office buildings of industrial parks. Ignoring other heat sources in the room, the relationship between the room temperature change and the air-conditioning power is shown in the following formula:

[0047] dT in,i (t) / dt=[T out (t)-T in,i (t)] / (R in,i C in,i )-η AC P AC,i (t) / C in,i

[0048] The air conditioner can maintain the room temperature near the air-conditioning temperature set value by starting intermittently. When the indoor temperature is stable, the above formula can be rewritten as:

[0049] P AC,i (t)=[T out (t)-T in,i(t)] / (η AC R in,i )

[0050] Assume that the outdoor temperature remains stable during the demand response process, and the relationship between the air-conditioning power reduction and the room temperature before and after the response is shown in the following formula:

[0051] ΔP AC,i (t)=[T in,i (t)-T in,i (t-Δt)] / (η AC R in,i )

[0052]

[0053] The constraints of air-conditioning power and human comfort are shown in the following formula:

[0054]

[0055] In the above formula, T in,i (t) is the indoor temperature of air-conditioner i at time t; T out (t) is the outdoor temperature at time t; R in,i and C in,i are the indoor equivalent thermal resistance and equivalent heat capacity of air-conditioner i respectively; η AC is the operating efficiency of the air-conditioner; P AC,i (t) is the power of air-conditioner i at time t; ΔP AC,i (t) is the power reduction of air-conditioner i; Δt is the time interval; ΔP AC (t) is the total power reduction of the air-conditioner; N AC is the total number of air-conditioners; T AC,i (t) is the temperature set value of air-conditioner i at time t; and are the upper and lower limits of the temperature set value of air-conditioner i respectively; and are the upper and lower limits of the power of air-conditioner i respectively.

[0056] Furthermore, according to the assessment of the response speed for peak shaving in the power system, the embodiments of the present invention select light industrial machine tools and steelmaking electric arc furnaces in the production line load as the demand response resources in the industrial load. Among them, the light industrial machine tools can respond to the peak shaving instruction by reasonably interrupting the discrete production process and resume production after the response ends. The power reduction of the machine tool is shown in the following formula:

[0057]

[0058] In the formula, ΔP MT (t) is the power reduction of the machine tool; N MT is the total number of machine tools; α MT,iis a 0-1 variable, representing the working condition of machine tool i; when machine tool i is shut down, α MT,i = 0; P MT,i (t) is the power of machine tool i.

[0059] On the premise of ensuring that the total melting energy of the electric arc furnace remains unchanged, power reduction can be achieved by adjusting the working gear of the electric arc furnace:

[0060]

[0061] In the formula, ΔP SAF (t) is the power reduction amount of the electric arc furnace; N SAF is the total number of electric arc furnaces; α SAF,i (t) is the gear of electric arc furnace i at time t; P SAF,i is the power change amount when the adjacent gears of electric arc furnace i change at time t.

[0062] Furthermore, the industrial load also includes an equivalent energy storage device in the standby link. In case of an accident, the factory standby link can ensure the normal operation of sensitive equipment. The factory standby link includes the reserve of intermediate products, energy storage, etc. When two adjacent processes in the production line participate in demand response, the power of the process can be reduced in two ways. One way is to start the energy storage to meet the power demand of the upstream process. The other way is to reduce the power of the upstream process and maintain the operation of the downstream process by putting into the reserve of intermediate products. Equivalent the reserve of intermediate products to an energy storage model, then the equivalent energy storage model of the standby link and its constraint conditions to be satisfied are shown in the following formula:

[0063]

[0064] In the formula: S RR,i (t) is the remaining capacity of equivalent energy storage i at time t; is the discharge power of equivalent energy storage i at time t; is the discharge efficiency of equivalent energy storage i at time t; is the total change amount of equivalent energy storage power at time t; N RR is the total number of equivalent energy storage in the standby link; is the power change amount of equivalent energy storage i at time t; are the upper and lower limits of the remaining capacity of equivalent energy storage i respectively; are the upper and lower limits of the discharge power of equivalent energy storage i respectively.

[0065] Then, for step 102, according to the user behavior factors of the demand-side load in the response process, calculate the reliability index for characterizing the effective response probability of the demand-side load.

[0066] After receiving the power system peak shaving instruction, the load aggregator selects the loads participating in peak shaving and issues an incentive signal. During the demand response process, some loads cannot respond effectively due to user behavior factors, resulting in the actual response volume of the load aggregator falling short of expectations and reducing the actual effect of the demand-side resources participating in peak shaving.

[0067] There are significant differences in user behavior factors among different types of loads. User behavior factors of air-conditioning users during the demand response process include: First, the on / off state of the air conditioner changes, such as the user turning off the air conditioner when leaving; Second, the set temperature value of the air conditioner changes, such as temperature-sensitive users resetting the temperature. The user behavior of production lines means that the marginal cost of their participation in demand response increases as the response capacity increases, resulting in a decrease in their willingness to participate in demand response. The user behavior of standby links means that the economic loss of encountering risks during their participation in the demand response process increases as the standby resources decrease, resulting in a decrease in their willingness to participate in demand response.

[0068] To achieve precise peak shaving control of load resources, the embodiments of the present invention define a load response reliability index. When the types of loads participating in peak shaving are determined, the normalized indexes of various load response reliability influencing factors are calculated, and the geometric mean of all relevant indexes is taken as the response reliability index of the load. The load response reliability index characterizes the probability that the load effectively responds to the instructions of the load aggregator, and the load response reliability increases as the index value increases. When regulating the load resources to participate in peak shaving, the load aggregator can use the load response reliability index as one of the decision-making information to formulate a peak shaving instruction allocation plan that takes into account both response reliability and economy.

[0069] In the embodiments of the present invention, the reliability index includes an air-conditioning response reliability index, a production line response reliability index, and an equivalent energy storage response reliability index, where: the air-conditioning response reliability index is determined according to the power consumption period index and the comfort index, the power consumption period index is determined according to the matching degree between the user's habitual power consumption period and the demand response period, and the comfort index is determined according to the matching degree between the user's desired temperature and the set temperature value of the air conditioner; the production line response reliability index is determined according to the interruptible load volume of machine tools, the reducible load volume of electric arc furnaces, and the reduction value of production efficiency caused by shutdown during the demand response process; the equivalent energy storage response reliability index is determined according to the reserve volume of standby resources and the maintainable operation time of sensitive equipment.

[0070] First, for the air-conditioning load, its reliability index is jointly determined by the electricity consumption period index and the comfort index. The electricity consumption period index reflects the probability of the air-conditioning switch state changing during the demand response process. The smaller the electricity consumption period index, the lower the matching degree between the user's habitual electricity consumption period and the demand response period, and the greater the probability of the air-conditioning changing its switch state. According to the historical data of air-conditioning power, the K-means algorithm is used to obtain the power curves of each typical day of the air-conditioning. By performing secondary clustering on similar typical day power curves, the cluster of user habitual electricity consumption period preference curves can be summarized. By analyzing the distance between the peak load period of the central curve of the user habitual electricity consumption period preference curve cluster and the demand response period, the expression of the electricity consumption period index is as follows:

[0071]

[0072] In the formula: is the electricity consumption period index of air-conditioning i; t start,i , t end,i are the start time and end time of the peak load period respectively; P λ,i is the judgment power of the peak load period; T is the scheduling period.

[0073] The comfort index reflects the probability of the air-conditioning temperature setting value changing during the demand response process. The smaller the comfort index, the lower the matching degree between the temperature expected by the user and the air-conditioning temperature setting value, and the easier it is for the user to reset the temperature. The influence degree of the air-conditioning temperature setting value on the user's comfort is as follows:

[0074]

[0075] In the formula: is the comfort index of air-conditioning i; and are the upper and lower limits of the expected temperature of air-conditioning i respectively.

[0076] Then the load response reliability index model of air-conditioning i is as follows:

[0077]

[0078] Furthermore, the load response reliability index of the production line reflects the impact of its response capacity on the willingness to participate in demand response. The smaller the load response reliability index of the production line, the more interruptible load of machine tools or reducible load of electric arc furnaces during the demand response process. The greater the impact of pipeline stagnation and scrap raw material accumulation on the production efficiency of the enterprise, the smaller its willingness to participate in demand response.

[0079] Specifically, it is as follows:

[0080]

[0081] In the formula: is the response reliability index of production line load i; S PL,i (t) is the number of products stacked on the assembly line or the stock of scrap steel to be melted in the electric arc furnace; ΔP PL,i (t) is the power reduction of the machine tool or the electric arc furnace; is the upper limit of the number of products stacked on the assembly line or the stock of scrap steel to be melted in the electric arc furnace; is the upper limit of the power reduction of the machine tool or the electric arc furnace.

[0082] Furthermore, the reliability index of the equivalent energy storage load reflects the risk of insufficient standby caused by the participation of the standby link in demand response during factory production. The smaller the reliability index of the equivalent energy storage response in the standby link, the less the standby resource reserve. In the event of an unexpected power outage, the shorter the time that sensitive equipment can be maintained in operation, and the greater the potential economic risk. The expression of the reliability index of the equivalent energy storage response in the standby link is as follows:

[0083]

[0084] In the formula: is the response reliability index of the equivalent energy storage i in the standby link.

[0085] For step 104, according to the load reduction amount, the cost index corresponding to each of the demand-side loads is calculated.

[0086] In the embodiment of the present invention, the cost index includes the air-conditioning load response cost, the production line load response cost, and the equivalent energy storage load response cost. Among them, the air-conditioning load response cost C AC,n is calculated through the following formula:

[0087]

[0088] In the formula, ρ AC,i is the compensation price of air conditioner i; ρ TOU (t) is the real-time electricity price; N AC is the total number of air conditioners; ΔP AC,i (t) is the power reduction of air conditioner i;

[0089] The production line load response cost C MLS,n is calculated through the following formula:

[0090] C MLS,n =ρ MT ΔP MT (t)+ρ SAF ΔP SAF (t)-ρ TOU (t)[ΔP MT (t)+ΔP SAF (t)]

[0091] where ρ MT and ρ SAF are the compensation prices of the machine tool and the electric arc furnace respectively; ΔP MT (t) is the power reduction of the machine tool; ΔP SAF (t) is the power reduction of the electric arc furnace;

[0092] The equivalent energy storage load response cost C RR,n is calculated by the following formula:

[0093]

[0094] where ρ RR,i is the compensation price of the equivalent energy storage i; N RR is the total number of equivalent energy storages in the standby link; is the power change of the equivalent energy storage i at time t.

[0095] For step 106, according to the load evaluation index of the demand-side load under the condition of bounded rationality, a load multi-dimensional prospect model is established to characterize the prospect value of the load invocation plan.

[0096] Cumulative prospect theory emphasizes the impact of cumulative effects on decision-making and replaces the weights of decision-making in prospect theory with cumulative weights. Assume that there are N schemes for the evaluation index of load m, arranged in ascending order as {x m,1 ,…,x m,n-1 ,x m,n ,…,x m,N}, and the occurrence probabilities are {p m,1 ,…,p m,n-1 ,p m,n ,…,p m,N}. Among them, (x m,1 ,…,x m,n-1 ) represents losses, and (x m,n ,…,x m,N ) represents gains.

[0097] It should be noted that the load evaluation index here refers to the reliability index or the cost index, that is, both can be substituted into this general model. The schemes of the index are alternative options that can be selected for load response. These options are sorted according to the index values for the subsequent multi-load response decision model to select.

[0098] The cumulative weight ω(p m,n ) is calculated by the following formula:

[0099]

[0100] ω(p m,1 ) = π(p m,1 ), ω(p m,N) = π(p m,N )

[0101] where: x m,n represents the index of load m in scenario n; p m,n represents the occurrence probability of the index of load m in scenario n; ω(·) is the cumulative weight function; π(·) is the initial weight function.

[0102] Multi-attribute decision-making refers to the situation where there are multiple scenarios, and the decision-maker needs to balance multiple conflicting attributes to select the optimal scenario. In the demand response process, the load aggregator faces the choice of multiple load combination scenarios, and the decision-making involves cost and reliability indicators, which belongs to a typical multi-attribute decision-making problem.

[0103] Therefore, the embodiment of the present invention introduces a multi-attribute decision-making framework into the cumulative prospect theory. In the demand response process, the load aggregator first evaluates each load index in the scenario according to the cumulative prospect theory and sets the index weights, obtains the prospect value of the scenario through linear weighting, and makes a decision after weighing all scenarios. The value function of the index of load m is as follows:

[0104]

[0105] where, x m,n represents the evaluation index of load m in scenario n; w m is the weight of the evaluation index of load m; is the reference value of the evaluation index of load m; κ m , β m and γ m are all index parameters;

[0106] Establish the prospect value function of the index of load m in scenario n:

[0107]

[0108] According to the above formula, establish the multi-dimensional prospect model V n (H, W, X, P, X 0 ):

[0109]

[0110] where: w m is the weight of the index of load m; is the reference value of the index; ε mn is the error term; H is the index parameter vector; W is the index weight vector; X and P represent the vectors of the index values and their occurrence probabilities in scenario n; X 0 is the reference value vector of the index; M is the total number of loads; U n , ε nThey are the deterministic term and the error term in the foreground value function respectively.

[0111] It should be noted that the uncertainty in the load aggregator's scheme selection is reflected as the differentiation of parameters H and W in the multi-dimensional load foreground model. Therefore, parameter identification of these two parameters is required. Specifically, since different types of loads have different sensitivities to electricity prices, it is necessary to analyze the historical load data, reasonably set the reference points of the load indicators, and reduce the impact of the reference points on the foreground value. According to the utility maximization theory, the probability that the load aggregator selects scheme n is equal to the probability that scheme n brings greater utility to the load aggregator than any other scheme. After using the foreground value to replace the traditional utility, the probability that the load aggregator selects scheme n is as follows:

[0112]

[0113] The mixed Logit model is a statistical model used to model individual behavior choices and can capture the uncertainty in the load aggregator's selection behavior. Based on the mixed Logit model, the probability that the load aggregator selects scheme n is determined as follows:

[0114]

[0115] In the formula, pb n is the probability that the load aggregator selects scheme n; A(χ) is the Logit probability function; f(χ|θ) is the Logit density function; χ is the set of index parameters to be estimated; θ is the characteristic parameter, such as the mean and variance of the normal distribution, etc.

[0116] The selection probability of the mixed Logit model is a non-closed integral. The maximum likelihood estimation method needs to be used to calculate the probability and determine the best parameter distribution. First, a random vector is sampled from the probability density function f(χ|θ) using Latin hypercube sampling, denoted as χ. Substituting it into the probability calculation equation for the load aggregator to select scheme n can obtain the probability values for the load aggregator to select any scheme. Then, assuming that the load aggregator faces Q decision scenarios, the maximum likelihood function of the load aggregator is as shown in the following formula:

[0117]

[0118] After taking the logarithm, it can be changed to:

[0119]

[0120] In the formula, y q represents the decision made by the load aggregator in scenario q. If the user selects scheme n, the value of the expression (y q = n) is 1, otherwise it is 0.

[0121] Finally, the genetic algorithm is used to solve for the optimal θ, such that the maximum likelihood operator reaches its maximum value. Substituting the obtained θ back into the above formula, the H and W parameters in the load multi-dimensional prospect model can be calculated.

[0122] Regarding step 108, based on the load multi-dimensional prospect model, a multi-load response decision model with the reliability index and the cost index as decision variables is established, and the optimal peak shaving strategy for the demand-side load is obtained by solving the multi-load response decision model.

[0123] After the electricity price information changes, the load aggregator will make a load demand response decision according to the new electricity price. Studying the real response behavior of the load aggregator can provide an important basis for formulating the demand response mechanism. In order to more accurately depict the bounded rational response behavior of the load aggregator under the demand response mechanism, the embodiments of the present invention are based on the load multi-dimensional prospect model under bounded rationality established in the above steps. A multi-load response decision model under the bounded rationality of the load aggregator is constructed by comprehensively considering cost and reliability:

[0124]

[0125] In the formula, maxV n is the maximum value of the multi-objective response decision function; and respectively represent the prospect values of the cost and reliability indexes of the load aggregator; C m,n represents the response cost of load m in plan n, that is, the corresponding cost index; ω(C m,n ) represents the cumulative weight function of the cost index of load m in plan n; is the reference value of the cost index; w COST,m is the weight of the cost index; is the reliability index of load m in plan n; is the cumulative weight function of the reliability index of load m in plan n; is the reference value of the reliability index; w LRR,m is the weight of the reliability index.

[0126] The calculation process of obtaining the optimal peak shaving strategy for the demand load side according to this model can be implemented by those skilled in the art themselves, and will not be elaborated here.

[0127] It should be noted that the above multi-load response decision model can accurately map the law of the influence of the changes in subjective and objective factors on the decision-making of the load aggregator after the integration of social elements. In addition, the model also considers the demand response cost, reliability, and bounded rational behavior of the load aggregator, and can effectively characterize the real response behavior of the load aggregator under the demand response mechanism. In a complex market environment, the effective regulation of the load can be achieved by tapping the adjustable potential of the massive loads on the demand side. To verify the rationality of the established bounded rational response decision model of the load aggregator, a response decision model of the load aggregator based on the assumption of perfect rationality was constructed for comparative analysis. Under the assumption of perfect rationality, the load aggregator usually adopts the expected utility theory in traditional economics and makes response decisions with the goal of maximizing utility. The user response decision model is divided into two parts, including the electricity cost function and the reliability level function, as follows:

[0128] maxV = w C C all + w R R all

[0129] In the formula, V is the total utility function; C all , w C are the load response cost of the load aggregator and its weight respectively; R all , w R are the load response reliability index of the load aggregator and its weight respectively.

[0130] Next, an actual industrial park is used as an example scenario to verify the effectiveness of the above method.

[0131] The electricity price information is shown in Table 1. The industrial park includes 40 air conditioners, 12 machine tools, 3 steelmaking electric arc furnaces, and the equivalent energy storage of 3 factory standby links. The load parameters of various types are shown in Table 2. The outdoor temperature is 33 °C, the initial indoor temperature follows a uniform distribution, which is U(21, 32) °C, and the air conditioner temperature setting value follows a uniform distribution, which is U(22, 29) °C.

[0132] Table 1

[0133]

[0134] Table 2

[0135]

[0136] According to the historical data of the air conditioner power, the K-means clustering algorithm is used to obtain the typical daily power curve of the air conditioner, as Figure 2 shown. Then, by performing secondary clustering on the 42 typical daily power consumption curves, two curve clusters with significantly different electricity consumption period preferences can be obtained, as Figure 3As shown. From the above calculation formula of the electricity consumption period index, it can be seen that the peak load periods of these two types of air conditioners are 09:00 - 19:00 and 00:00 - 16:00 respectively. Therefore, the first type of air conditioner is suitable for responding to the daytime load reduction plan, and the second type of air conditioner is suitable for responding to the load reduction plans in the early morning and afternoon. Based on the clustering analysis results of the air conditioner power, the load aggregator can make decisions according to the electricity consumption period index and comfort index calculated for the dispatching period.

[0137] Taking the parameter identification of the air conditioner as an example for illustration, the parameter identification processes of machine tools, electric arc furnaces, and the equivalent energy storage in the standby link are similar, and will not be elaborated in this article. Existing literature points out that the weight functions of benefits and losses are similar. To avoid increasing the model complexity by using different weight functions, this article does not distinguish the weight functions of benefit and loss probabilities. The parameter identification evaluation criterion uses McFadden's R 2 . The air conditioner temperature set values in summer can be divided into 22 - 23 degrees, 24 - 25 degrees, 26 - 27 degrees, and above 28 - 29 degrees. They are divided into categories A, B, C, and D for parameter identification, and each type of air conditioner needs to determine the temperature set value according to different electricity prices. Considering the parameter differences in cost and reliability, a method combining cumulative prospect theory and mixed Logit model is used to identify the air conditioner parameters, and the results are shown in Table 3. The value functions of cost and reliability are as Figure 4 shown.

[0138] Table 3

[0139]

[0140] As shown in Table 3 and Figure 4 (a), it can be seen that the average value of the sensitivity coefficient of cost is about 1, and the value perception of the load aggregator has an almost linear relationship with the cost. Because the initial powers of various air conditioners are different, there are differences in the cost reference points of different air conditioners. The increase in cost has the greatest impact on air conditioners with low temperature set values, and their value perception is lower when the cost rises. As shown in Table 3 and Figure 4 (b), it can be seen that the sensitivity coefficient of reliability is about 1.8, and the value function is a convex function on the benefit side (saving regulation cost or high reliability) and a concave function on the loss side (increase in regulation cost or decrease in reliability). This result conforms to the inverse S-shaped value function, and the more the electricity consumption decision deviates from the habit, the more repulsive it is to users. In summer, compared with the consideration of cost, the load aggregator is significantly more sensitive to reliability, and the load aggregator is more inclined to maintain the use of the air conditioner and appropriate temperature set values during the decision-making process.

[0141] In the electricity market, the power system can guide the load aggregator to make load response decisions through time-of-use electricity prices. The load change situations before and after the load aggregator responds are as Figure 5As shown in the figure. Compared with the initial load, the peak load of the load aggregator has been significantly adjusted during the peak load period after demand response, and most of the load in the evening has been transferred to the period with lower electricity prices. Under the assumption of perfect rationality, the load aggregator can fully understand and predict the results of different decisions, and the selected plan can bring the maximum utility. Compared with before demand response, the peak load of the load aggregator has decreased by 36.68%, and the daily load rate has increased by 16.79%. The limited rationality of users determines that users can only optimize decisions within the scope of rationality. Compared with before demand response, the peak load of the load aggregator with limited rationality has decreased by 40.49%, and the daily load rate has increased by 12.8%. The optimal decision result of limited rationality has gradually approached perfect rationality.

[0142] Therefore, although the load aggregator with perfect rationality can make the optimal load response decision according to the time-of-use electricity price, this assumption is too idealistic. In reality, these prerequisite conditions are often difficult to meet. Facing complex and uncertain situations, the actual decision-making of the load aggregator will be affected by limited rationality, and usually simplified decision rules or heuristic methods will be adopted instead of complete optimization analysis. Therefore, limited rationality is more in line with the decision-making mode of the load aggregator in real life.

[0143] Furthermore, through the comparison of two scenarios, the necessity of introducing the reliability index in the demand response process of the load aggregator is illustrated. Strategy 1 does not consider the reliability of load response. The load aggregator takes the minimum cost as the optimization goal and distributes the power to be curtailed to various types of loads. Strategy 2 adopts the load response decision proposed in this paper to find the optimal distribution plan of the power to be curtailed among various types of loads.

[0144] After receiving the peak shaving instruction from the power system, the load aggregator cuts the load power in the next 1 hour according to different strategies. Assume that at 20:00, the load aggregator receives peak shaving instructions of 8MW and 20MW respectively. The load response of the load aggregator under different strategies is as Figure 6 shown. From Figure 6 (a) and Figure 6 (b), it can be seen that when the peak shaving instruction is relatively low, the load aggregator gives priority to cutting the electric arc furnace and air conditioner, and the equivalent energy storage in the standby link remains basically unchanged. Among them, some machine tools and air conditioners in Strategy 1 fail to respond reliably, resulting in the actual curtailment being 0.3MW less than expected. From Figure 6 (c), it can be seen that in Strategy 1, in order to meet the peak shaving instruction of 20MW, the load aggregator issues curtailment instructions to all loads. Due to ignoring the impact of excessive curtailment power on the load response willingness, the machine tools are curtailed 0.4MW less than expected. In addition, some air conditioners do not respond effectively, resulting in a large deviation between the expected curtailment and the actual curtailment. From Figure 6As can be seen from (d), in Strategy 2, the load aggregator takes into account the impact of power curtailment on the willingness of load response, distributes the power curtailment amount more evenly, and improves the reliability of each load response.

[0145] Furthermore, different peak shaving instructions are sent to the load aggregator at 19:00, 20:00, and 21:00 respectively. The average values of the indicators of each peak shaving instruction at different times are shown in Table 4. When the peak shaving instruction is 20 MW, the actual adjustment effect of Strategy 2 is very close to the peak shaving instruction. However, the cost performance of Strategy 1 is better than that of Strategy 2. The main reason is that Strategy 2 is limited by the reliability index, and the load response plan is not as radical as that of Strategy 1. As the peak shaving instruction decreases, the load response reliability of Strategy 2 can be maintained at 96%. Since Strategy 1 lacks the evaluation of the load response state and fails to accurately curtail the load, the cost performance index is relatively high. When the peak shaving instruction is as low as 8 MW, the differences in cost performance and reliability between the two strategies are not significant, but the reliability of Strategy 2 is still better than that of Strategy 1.

[0146] Table 4

[0147]

[0148] Therefore, the load response strategy proposed by this method has advantages in both reliability and cost performance, can better complete load curtailment, and cooperate with the power system to complete the peak shaving task, verifying the superiority of the proposed strategy. However, as the peak shaving instruction increases, the advantages of the strategy proposed by this method gradually decrease. When the peak shaving instruction is large, the strategy proposed by this method will be more conservative than the strategy that only considers cost performance. Therefore, to ensure the reliability and cost performance of load response, the strategy proposed by this method is more suitable for the situation where the peak shaving instruction does not exceed the adjustable capacity of the load aggregator.

[0149] Please refer to Figure 7 , an embodiment of the present invention provides an optimization device for a multi-load peak shaving strategy on the power demand side. The device includes:

[0150] A first calculation module 700, configured to calculate the load curtailment amount of the demand-side load for responding to the scheduling instruction according to the scheduling instruction issued by the power system; wherein, the demand-side load includes the air-conditioning load and industrial load in the industrial park;

[0151] A second calculation module 702, configured to calculate a reliability index for characterizing the effective response probability of the demand-side load according to the user behavior factors of the demand-side load during the response process;

[0152] A third calculation module 704, configured to calculate the cost performance index corresponding to each demand-side load reduction according to the load curtailment amount;

[0153] The first modeling module 706 is configured to establish a multi-dimensional load prospect model for characterizing the prospect value of a load invocation plan according to the load evaluation index of the demand-side load under bounded rationality conditions;

[0154] The second modeling module 708 is configured to establish a multi-load response decision model with the reliability index and the cost index as decision variables according to the multi-dimensional load prospect model, and solve the multi-load response decision model to obtain the optimal peak shaving strategy of the demand-side load.

[0155] In an embodiment of the present invention, calculating the load reduction amount of the demand-side load for responding to the dispatching instruction sent by the power system includes: calculating the total reduction amount of the air-conditioning power before and after response according to the change relationship between the room temperature and the air-conditioning power; determining the total reduction amount of the power of the production line load in the industrial load according to the operation process of the equipment; wherein, the production line load includes light industrial machine tools and steelmaking electric arc furnaces, the total reduction amount of the airport power of the light industrial machine tools is determined by interrupting the discrete production process, and the total reduction amount of the electric arc furnace power of the steelmaking electric arc furnace is determined by adjusting the working gear of the electric arc furnace; equivalent the intermediate product reserve to an energy storage model, and determine the total reduction amount of the equivalent energy storage power of the equivalent energy storage load of the standby link in the industrial load.

[0156] In an embodiment of the present invention, the reliability index includes an air-conditioning response reliability index, a production line response reliability index, and an equivalent energy storage response reliability index, wherein: the air-conditioning response reliability index is determined according to the power consumption period index and the comfort index, the power consumption period index is determined according to the matching degree between the user's habitual power consumption period and the demand response period, and the comfort index is determined according to the matching degree between the temperature expected by the user and the air-conditioning temperature setting value; the production line response reliability index is determined according to the interruptible load amount of the machine tool, the reducible load amount of the electric arc furnace, and the reduction value of the production efficiency caused by shutdown during the demand response process; the equivalent energy storage response reliability index is determined according to the reserve amount of the standby resource and the maintainable operation time of the sensitive equipment.

[0157] In an embodiment of the present invention, the cost index includes an air-conditioning load response cost, a production line load response cost, and an equivalent energy storage load response cost, wherein:

[0158] The air-conditioning load response cost C AC,n is calculated by the following formula:

[0159]

[0160] In the formula, ρ AC,i is the compensation price of the air conditioner i; ρ TOU (t) is the real-time electricity price; N AC is the total number of air conditioners; ΔPAC,i (t) is the power reduction of air conditioner i;

[0161] The production line load response cost C MLS,n is calculated by the following formula:

[0162] C MLS,n = ρ MT ΔP MT (t) + ρ SAF ΔP SAF (t) - ρ TOU (t)[ΔP MT (t) + ΔP SAF (t)]

[0163] In the formula, ρ MT and ρ SAF are the compensation prices of the machine tool and the electric arc furnace respectively; ΔP MT (t) is the power reduction of the machine tool; ΔP SAF (t) is the power reduction of the electric arc furnace;

[0164] The equivalent energy storage load response cost C RR,n is calculated by the following formula:

[0165]

[0166] In the formula, ρ RR,i is the compensation price of the equivalent energy storage i; N RR is the total number of equivalent energy storage in the standby link; is the power change of the equivalent energy storage i at time t.

[0167] In the embodiments of the present invention, the load multi-dimensional prospect model for characterizing the prospect value of the load invocation plan is established according to the load evaluation index of the demand-side load under the condition of bounded rationality, including:

[0168] According to the occurrence probability of each load evaluation index in each load invocation plan, the cumulative weight ω(p m,n ) of each load evaluation index is calculated:

[0169]

[0170] Among them, p m,n represents the occurrence probability of the load m evaluation index in plan n; i is the i-th plan among the N plans; ω(·) is the cumulative weight function; π(·) is the initial weight function;

[0171] According to the losses and benefits of each load evaluation index in each load invocation plan, the value function of each load evaluation index is calculated:

[0172]

[0173] Among them, x m,n represents the evaluation index of load m in scenario n; w m is the weight of the evaluation index of load m; is the reference value of the evaluation index of load m; κ m , β m and γ m are all index parameters;

[0174] According to the cumulative weight and the value function, a multi-dimensional load prospect model V n (H, W, X, P, X 0 ) for characterizing the prospect value of the load call scenario n is established:

[0175]

[0176] H = {(β m , γ m , κ m ) | m = 1, 2,..., M}

[0177] W = {w m | m = 1, 2,..., M}

[0178] Among them, H is the index parameter vector; W is the index weight vector; X and P represent the vectors of the index values and their occurrence probabilities in scenario n; X 0 is the index reference value vector; M is the total number of loads; ε n is the error term in the multi-dimensional load prospect model.

[0179] In the embodiment of the present invention, the multi-load response decision model is established by the following formula:

[0180]

[0181] In the formula, maxV n is the maximum value of the multi-objective response decision function; and respectively represent the prospect values of the cost and reliability indexes of the load aggregator; C m,n represents the cost index of load m in scenario n; ω(C m,n ) represents the cumulative weight function of the cost index of load m in scenario n; is the reference value of the cost index; w COST,m is the weight of the cost index; is the reliability index of load m in scenario n; is the cumulative weight function of the reliability index of load m in scenario n; is the reference value of the reliability index; w LRR,m is the weight of the reliability index.

[0182] It should be noted that: The power demand side multi-load peak shaving strategy optimization device provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the power demand side multi-load peak shaving strategy optimization device provided in the above embodiments and the embodiments of the power demand side multi-load peak shaving strategy optimization method belong to the same concept. The specific implementation process can be seen in the method embodiments and will not be elaborated here.

[0183] The embodiments of the present application also provide a computer device. Please refer to Figure 8 , the computer device includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory. At least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the power demand side multi-load peak shaving strategy optimization method provided in each of the above method embodiments.

[0184] The embodiments of the present application also provide a computer-readable storage medium. At least one instruction, at least one program, a code set or an instruction set is stored on the computer-readable storage medium. At least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the power demand side multi-load peak shaving strategy optimization method provided in each of the above method embodiments.

[0185] The embodiments of the present application also provide a computer program product. The computer program product includes a computer program. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program to enable the computer device to execute the power demand side multi-load peak shaving strategy optimization method described in any one of the above embodiments.

[0186] Finally, it should also be noted that in this article, relational terms such as first, second, third, and fourth are only used 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.

[0187] The above is only the preferred embodiment of the present application. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. An optimization method for multi-load peak shaving strategies on the power demand side, characterized in that, The method includes: Calculating the load reduction amount of the demand-side load for responding to the dispatching instruction issued by the power system; wherein, the demand-side load includes the air-conditioning load and industrial load in the industrial park; Calculating a reliability index for characterizing the effective response probability of the demand-side load according to the user behavior factors of the demand-side load during the response process; Calculating a cost index corresponding to each type of the demand-side load according to the load reduction amount; Establishing a multi-dimensional load prospect model for characterizing the prospect value of the load calling scheme according to the load evaluation index of the demand-side load under bounded rationality conditions; Establishing a multi-load response decision model with the reliability index and the cost index as decision variables according to the multi-dimensional load prospect model, and solving the multi-load response decision model to obtain the optimal peak shaving strategy of the demand-side load.

2. The method according to claim 1, wherein The calculating the load reduction amount of the demand-side load for responding to the dispatching instruction issued by the power system includes: Calculating the total reduction amount of the air-conditioning power before and after the response according to the change relationship between the room temperature and the air-conditioning power; Determining the total reduction amount of the power of the production line load in the industrial load according to the operation process of the equipment; wherein, the production line load includes light industrial machine tools and steelmaking electric arc furnaces, the total reduction amount of the airport power of the light industrial machine tools is determined by interrupting the discrete production process, and the total reduction amount of the electric arc furnace power of the steelmaking electric arc furnace is determined by adjusting the working gear of the electric arc furnace; Equivalently modeling the intermediate product reserve as an energy storage model, and determining the total reduction amount of the equivalent energy storage power of the equivalent energy storage load of the standby link in the industrial load.

3. The method according to claim 2, characterized in that, The reliability index includes an air-conditioning response reliability index, a production line response reliability index, and an equivalent energy storage response reliability index, wherein: The air-conditioning response reliability index is determined according to the power consumption period index and the comfort index, the power consumption period index is determined according to the matching degree between the user's habitual power consumption period and the demand response period, and the comfort index is determined according to the matching degree between the temperature expected by the user and the air-conditioning temperature setting value; The production line response reliability index is determined according to the interruptible load amount of the machine tools, the reducible load amount of the electric arc furnace, and the reduction value of the production efficiency caused by shutdown during the demand response process; The equivalent energy storage response reliability index is determined according to the reserve amount of the standby resource and the maintainable operation time of the sensitive equipment.

4. The method according to claim 2, wherein The cost index includes the air-conditioning load response cost, the production line load response cost, and the equivalent energy storage load response cost, wherein: The air-conditioning load response cost C AC,n is calculated by the following formula: where ρ AC,i is the compensation price of air conditioner i; ρ TOU (t) is the real-time electricity price; N AC is the total number of air conditioners; ΔP AC,i (t) is the power reduction of air conditioner i; The load response cost C of the production line MLS,n is calculated by the following formula: C MLS,n = ρ MT ΔP MT (t) + ρ SAF ΔP SAF (t) - ρ TOU (t)[ΔP MT (t) + ΔP SAF (t)] where ρ MT and ρ SAF are the compensation prices of the machine tool and the electric arc furnace respectively; ΔP MT (t) is the power reduction of the machine tool; ΔP SAF (t) is the power reduction of the electric arc furnace; The equivalent energy storage load response cost C RR,n is calculated by the following formula: where ρ RR,i is the compensation price of the equivalent energy storage i; N RR is the total number of equivalent energy storages in the reserve link; is the power change amount of the equivalent energy storage i at time t.

5. The method according to claim 1, wherein The establishing a multi-dimensional load prospect model for characterizing the prospect value of the load calling scheme according to the load evaluation index of the demand-side load under bounded rationality conditions includes: According to the occurrence probability of each load evaluation index in each load invocation scheme, the cumulative weight ω(p m,n ) of each load evaluation index is calculated: Among them, p m,n represents the occurrence probability of the evaluation index of load m in scenario n; i is the i-th scenario among the N scenarios in total; ω(·) is the cumulative weight function; π(·) is the initial weight function. Calculating a value function of each load evaluation index according to the losses and benefits of each load evaluation index in each load calling scheme; Among them, x m,n represents the evaluation index of load m in solution n; w m is the weight of the evaluation index of load m; is the reference value of the evaluation index of load m; κ m , β m and γ m are all index parameters; Based on the cumulative weight and the value function, a load multi-dimensional prospect model V for characterizing the prospect value of the load invocation scenario n is established n (H,W,X,P,X 0 ): H = { (β m , γ m , κ m ) | m = 1, 2, …, M} W = {w m | m = 1, 2, …, M} Among them, H is the index parameter vector; W is the index weight vector; X and P represent the vectors of the index values and their occurrence probabilities in scheme n; X 0 is the index reference value vector; M is the total number of loads; ε n is the error term in the multi-dimensional foreground model of the load.

6. The method according to claim 5, wherein The multi-load response decision model is established by the following formula: where maxV n is the maximum value of the multi-objective response decision function; and respectively represent the prospect values of the cost and reliability indicators of the load aggregator; C m,n represents the cost indicator of load m in scenario n; ω(C m,n ) represents the cumulative weight function of the cost indicator of load m in scenario n; is the reference value of the cost indicator; w COST,m is the weight of the cost indicator; is the reliability indicator of load m in scenario n; is the cumulative weight function of the reliability indicator of load m in scenario n; is the reference value of the reliability indicator; w LRR,m is the weight of the reliability indicator.

7. An optimization device for multi-load peak shaving strategies on the power demand side, characterized in that, The device includes: A first calculation module, configured to calculate the load reduction amount of the demand-side load for responding to the dispatching instruction issued by the power system; wherein, the demand-side load includes the air-conditioning load and industrial load in the industrial park; A second calculation module, configured to calculate a reliability index for characterizing the effective response probability of the demand-side load according to the user behavior factors of the demand-side load during the response process; A third calculation module, configured to calculate a cost index corresponding to each type of the demand-side load cut according to the load cut amount; A first modeling module, configured to establish a multi-dimensional load prospect model for characterizing the prospect value of the load invocation plan according to the load evaluation index of the demand-side load under bounded rationality conditions; A second modeling module, configured to establish a multi-load response decision model with the reliability index and the cost index as decision variables according to the multi-dimensional load prospect model, and solve the multi-load response decision model to obtain the optimal peak shaving strategy of the demand-side load.

8. A computer device, characterized in that, The computer device includes a memory and a processor. The memory is used to store a computer program, and the processor is used to execute the computer program stored on the memory to implement the steps of the method according to any one of claims 1-6 above.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1-6 are implemented.

10. A computer program product, characterized in that, It includes a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1-6 are implemented.

Images