Provincial and local load collaborative intra-day scheduling method based on orderly power utilization scene

Abstract

The invention provides a provincial and regional load collaborative intra-day scheduling method based on an orderly power utilization scene, and the method comprises the steps: building a provincial and regional collaborative orderly power utilization framework in consideration of a regional power utilization level; an interaction model of a regional power grid and a demand response user is established in combination with the uncertainty of the load side user consciousness of consciousness response and energy-saving awareness, and a user refined screening strategy is provided; meanwhile, the daily supply and demand level and demand response execution condition are considered, and a provincial and regional power grid orderly power utilization load cooperative control and order recovery strategy is provided. By adopting the technical scheme provided by the invention, the orderly power utilization development efficiency can be improved, the collaborative intra-day scheduling of provincial and regional power grid loads can be realized, the balance of supply and demand can be maintained, the stable operation of the power grid can be ensured, and a relatively good application effect can be achieved.

Description

Technical Field

[0001] The present invention relates to the field of power system load management, and in particular to a provincial and local load coordinated intraday dispatching method based on an orderly power consumption scenario. Background Art

[0002] In the operation and management of today's power system, with the continuous increase in power demand and the increasingly complex power consumption environment, optimizing power resource allocation and improving energy efficiency have become key challenges. Especially with the development of orderly power consumption, how to achieve coordinated daily dispatch of provincial and local loads to maintain supply and demand balance, ensure stable operation of the power grid, and take into account economic benefits and environmental sustainability has become a technical bottleneck that the power industry urgently needs to break through.

[0003] Traditional power load dispatching methods are mainly based on empirical prediction and static strategies, which are incapable of coping with rapid fluctuations in electricity demand. Therefore, with the advancement of smart grid technology and the promotion of demand response (DR) mechanisms, power systems have begun to explore new, more flexible and interactive ways of load management. Demand response allows power users to adjust their electricity consumption behavior according to the needs and incentive mechanisms of the power grid. The State Grid also proposed to adhere to the principle of "demand response first, orderly electricity consumption as a bottom line, and energy conservation assistance" in power supply services. However, how to efficiently integrate demand response resources in combination with users' energy-saving awareness, realize effective coordinated dispatch between provincial and local power grids, and carry out coordinated dispatch and order restoration of orderly electricity loads in provincial and local power grids based on the execution of demand response are still technical problems that need to be overcome in the field of power systems. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a method for coordinated daytime scheduling of provincial and local loads based on an orderly power consumption scenario, comprising the following steps:

[0005] S1, the provincial power grid will combine the load benchmark curves of each node in the power system to predict the power consumption level of each area, and send power consumption level information and demand response instructions to the regional power grid;

[0006] S2, after receiving the power consumption level information and demand response instructions sent by the provincial power grid, the regional power grid selects users who have registered to participate in the demand response and sends them the power consumption level information and demand response instructions;

[0007] S3, after receiving the electricity consumption level information and demand response instructions, the selected user who has registered to participate in the demand response chooses whether to participate in the demand response. If the user chooses to participate, when the demand response starts, the regional power grid will remotely control the user's smart home temperature control system to reduce the electricity load;

[0008] S4: The regional power grid calculates the execution status of the current demand response every 15 minutes and reports it to the provincial power grid. The provincial power grid generates a strategy for coordinated control and orderly restoration of the provincial and local power grid's orderly power load for the next 4 hours every 15 minutes.

[0009] Furthermore, the calculation method of the power consumption level in S1 is:

[0010] Assuming that the demand response at node j will be carried out at time p on a certain day and last for h hours, the electricity consumption level at node j is:

[0011]

[0012] in, is the load baseline value of node j at time l in the tth round of demand response day; and They are electricity consumption level thresholds, which are set by the provincial power grid based on the historical electricity consumption data of the province.

[0013] Furthermore, the demand response information in S1 includes the demand response start time, end time and the load amount that the user is required to reduce.

[0014] Furthermore, the method by which the regional power grid in S2 selects users who have registered to participate in demand response is:

[0015] Assuming that orderly electricity consumption is carried out T times, for the regional power grid at node j, its objective function is to select the user group with the largest expected reward to maximize long-term benefits:

[0016]

[0017] Among them, in the tth round of demand response, S j，t is the set of users selected at node j; ΔP i，t is the load reduction required for user i; i，t ∈{0, 1} is the final response status of user i. If it is 1, it means that user i participates in demand response, and if it is 0, it means that user i exits demand response; C i，t is the incentive that user i gets after the response is completed; B t Represents the budget for this round of demand response.

[0018] Furthermore, the final response status described in S3 is determined based on the total load reduced in the demand response. During the demand response process, the selected user has the right to participate in or exit the demand response at any time. However, as long as the load reduced reaches or exceeds the load required to be reduced, it is deemed that the user participates in the demand response, otherwise it is deemed to exit the demand response.

[0019] Furthermore, the method for coordinated intraday dispatching of provincial and local loads based on orderly power consumption scenarios is characterized in that the method for the regional power grid in S2 to select users who have registered to participate in demand response is:

[0020] In the tth round of demand response, the regional power grid solves the following optimization problem to determine the optimal user participation combination:

[0021]

[0022] Among them, in the tth round of demand response, the estimated response probability of any user i at node j is:

[0023]

[0024] In the first round of demand response, all users have not been selected. For any user i, there is A i，0 =1, b i，0 =0

[0025] After the tth round of demand response, by updating user i’s A i and b i To achieve the learning of user response behavior, for any selected user i at node j, there is

[0026] A i，t+1 =A i，t +(δ j，t )2

[0027] b i，t+1 =b i，t +δ j，t r i，t

[0028] Among them, r i，t is the reward of user i. If user i is not selected, then A i and b i constant.

[0029] Furthermore, the orderly coordinated control and order restoration strategy of the provincial and local power grids in S4 is to perform coordinated control and downward adjustment of the provincial and local power grids when demand response is not fully implemented and the load is greater than the power supply capacity; in the intraday power tight balance link, there is sufficient power supply capacity to perform orderly power load order restoration.

[0030] Furthermore, the optimization goal of the coordinated control and order restoration of provincial and local power grids is to minimize the operation and maintenance costs of thermal power units and the economic losses of power outages at system nodes:

[0031]

[0032] Among them, q is a certain period of time within a day; the total cost is C q ; aj 、b j and c j is a constant. j，q and Respectively represent the power generation and load of node j; Ω G and Ω fh They are the node set where thermal generators are installed and the load node set. fh is the load outage cost coefficient of node j; represents the power supply status of node j as a binary variable, where Indicates power supply, Indicates a power outage.

[0033] The constraints of the objective function include thermal power unit output constraints, ramp constraints, system flow constraints, and branch flow safety constraints:

[0034]

[0035] in, Respectively represent the set of thermal power units and renewable energy units at node j; Ω j is the set of all nodes connected to node j. ij，q is the active power flow from node i to node j during period q. j min and P j max are the minimum and maximum output limits of the generator at node j, respectively. j u and r j d is the minimum and maximum ramp rate of the thermal generator at node j. j，q is the phase angle of node j in period q. ij is the reactance of the branch between nodes i and j. is the aggregated load of the regional power grid at node j during period q.

[0036] The provincial and local load coordinated intraday dispatching method based on the orderly power consumption scenario provided by the present invention fully considers the regional power consumption level and builds a provincial and local coordinated orderly power consumption framework; at the same time, it particularly considers the uncertainty of conscious response and energy-saving awareness of load-side users, establishes an interactive model between regional power grids and demand response users, and proposes a user refinement screening strategy; the present invention proposes a coordinated control and order recovery strategy for orderly power consumption loads in provincial and local power grids based on the intraday supply and demand level and the execution of demand response. The present invention helps to improve the efficiency of orderly power consumption and realize the coordinated intraday dispatching of provincial and local power grid loads to maintain supply and demand balance and ensure stable operation of the power grid, and has high engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings in the specification, which constitute a part of this application, are used to provide a further understanding of the present invention. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0038] Figure 1 The present invention is a flow chart of a method for coordinated daytime scheduling of provincial and local loads based on an orderly electricity consumption scenario. DETAILED DESCRIPTION

[0039] In order to deepen the understanding of the present invention, the present invention will be further described in detail below with reference to embodiments and drawings.

[0040] like Figure 1 As shown, this embodiment includes the following steps:

[0041] Considering the regional electricity consumption level, a framework for coordinated and orderly electricity consumption between provinces and regions is established;

[0042] The framework includes three parts: provincial power grid, regional power grid, and load-side users. The provincial power grid will combine the load benchmark curves of each node in the power system to predict the power consumption level of each region, and send power consumption level information and demand response instructions to each regional power grid; assuming that the demand response at node j will be carried out during the p period of a certain day and last for h hours, the values ​​are 16 and 4 respectively, then the power consumption level at node j is:

[0043]

[0044] Among them, E OC is the predicted electricity consumption level of node j before the tth round of demand response; is the load baseline value of node j at time l in the tth round of demand response day, which is the measured data; and They are electricity consumption level thresholds, which are set by the provincial power grid according to the historical electricity consumption data of the province, and are 3000MW and 6000MW respectively.

[0045] After receiving the demand response instructions and local electricity consumption level information from the provincial power grid, the regional power grid will select users who have registered to participate in the demand response, send demand response information (must include the demand response start time, end time, and the amount of load that the user is required to reduce), and will send the local electricity consumption level to the users.

[0046] After receiving the demand response information and local electricity consumption levels, the selected users will choose whether to participate in the demand response. If they choose to participate, when the demand response begins, the regional power grid will use advanced smart metering infrastructure to remotely control the user's smart home temperature control system to adjust the air-conditioning system load and achieve load reduction.

[0047] Assuming that orderly electricity consumption is carried out T times, for the regional power grid at the node, its objective function is to select the user group with the largest expected reward to maximize the long-term benefits:

[0048]

[0049] Among them, in the tth round of demand response, S j，t is the set of users selected at node j; ΔP i，t is the load reduction required for user i; i，t ∈{0, 1} is the final response status of user i. If it is 1, it means that user i participates in demand response, and if it is 0, it means that user i exits demand response.

[0050] Since the regional power grid budget is limited, the constraints for the objective function are:

[0051]

[0052] Among them, in the tth round of demand response, C i，t is the incentive that user i gets after the response is completed; B t Represents the budget for this round of demand response.

[0053] The regional power grid calculates the current demand response execution status every 15 minutes and reports it to the provincial power grid. The provincial power grid generates a rolling strategy for orderly power load coordinated control and sequential restoration for the provincial and local power grids in the next four hours every 15 minutes.

[0054] Combining the uncertainty of voluntary response and energy-saving awareness of users on the load side, an interactive model between the regional power grid and demand response users is established, and a user refinement screening strategy is proposed;

[0055] The decision of the user selected to participate in the demand response is highly uncertain. For any user i, its final response state is subject to the parameter p i，t The Bernoulli distribution of the proposed method is used to construct a contextual multi-armed bandits model (CMAB) to simulate the dynamic interaction between the regional power grid and registered demand response users.

[0056] In the CMAB model, each registered demand response user is equivalent to an arm, and the regional power grid is the intelligent agent. The local electricity consumption level sent by the regional power grid is the context information. In each round of demand response, the regional power grid performs the action of selecting users and obtains rewards from each selected user. For any user i, its reward r i，t for:

[0057] r i，t =ΔP i，t z i，t

[0058] According to the context information and rewards obtained in each round, the agent will learn and characterize the user response behavior based on a data-driven approach, and use the learning results to improve the user's refined screening strategy. Considering the variability of context information, the linear upper confidence bound (LinUCB) algorithm is used to optimize the model. Specifically, in the tth round of demand response, the estimated response probability of any user i at node j is:

[0059]

[0060] in:

[0061]

[0062] In the first round of demand response, all users have not been selected. For any user i, we have:

[0063] A i，0 =1, b i，0 =0

[0064] After the tth round of demand response, by updating user i’s A i and b i To achieve the learning of user response behavior, for any user i selected at node j, we have:

[0065] A i，t+1 =A i，t +(δ j，t )2

[0066] b i，t+1 =b i，t +δ j，t r i，t

[0067] If user i is not selected, then A i and b i constant.

[0068] Based on the confidence upper bound of user response probability, a user refined screening strategy is proposed, that is, in the tth round of demand response, the regional power grid solves the following optimization problem to determine the optimal user participation combination:

[0069]

[0070]

[0071] Among them, α is the exploration parameter of the LinUCB algorithm, which is set to 1 to balance the efficiency of exploration and utilization.

[0072] The user's real response probability and final response state are randomly generated by the simulation. The regional power grid using the LinUCB algorithm is compared with the demand response user interaction model and the random user selection method. The participation rate is significantly improved, as shown in Table 1.

[0073] Table 1 Average participation rate of demand response under different algorithms

[0074] Randomly select users LinUCB Average participation rate 47.45％ 55.58％

[0075] The model results show that the model and strategy can effectively improve the participation rate of demand response and the execution effect of orderly power consumption, which is of great significance to the coordinated daily dispatching of provincial and local power grid loads.

[0076] Furthermore, according to the supply and demand levels and the execution of demand response, a strategy for coordinated control of provincial and local power grid loads and sequential restoration can be proposed.

[0077] The above are only specific embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios, characterized in that: The following steps are involved: S1, the provincial power grid combines the load benchmark curves of each node in the power system to predict the power consumption level of each area, and sends the power consumption level information and demand response instructions to the regional power grid; S2, after receiving the power consumption level information and demand response instructions sent by the provincial power grid, the regional power grid selects users who have registered to participate in the demand response and sends them the power consumption level information and demand response instructions; S3, after receiving the electricity consumption level information and demand response instructions, the selected user who has registered to participate in the demand response chooses whether to participate in the demand response. If the user chooses to participate, when the demand response starts, the regional power grid will remotely control the user's smart home temperature control system to reduce the electricity load; S4: The regional power grid calculates the execution status of the current demand response every 15 minutes and reports it to the provincial power grid. The provincial power grid generates a strategy for coordinated control and orderly restoration of the provincial and local power grid's orderly power load for the next 4 hours every 15 minutes.

2. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios as claimed in claim 1 is characterized in that: The calculation method of the power consumption level in S1 is: Assuming that the demand response at node j will be carried out at time p on a certain day and last for h hours, the electricity consumption level at node j is: in, is the load baseline value of node j at time l in the tth round of demand response day; and They are respectively the electricity consumption level thresholds, which are set by the provincial power grid based on the historical electricity consumption data of the province.

3. The method for coordinated daytime dispatching of provincial and local loads based on orderly electricity consumption scenarios as claimed in claim 1 is characterized in that: The demand response information in S1 includes the demand response start time, end time and the load amount that the user is required to reduce.

4. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios as claimed in claim 1 is characterized in that: The method used by the regional power grid in S2 to select users who have registered to participate in demand response is: Assuming that orderly electricity consumption is carried out T times, for the regional power grid at node j, its objective function is to select the user group with the largest expected reward to maximize long-term benefits: Among them, in the tth round of demand response, S j,t is the set of users selected at node j; ΔP i,t is the load reduction required for user i; i,t ∈{0,1} is the final response status of user i. If it is 1, it means that user i participates in demand response, and if it is 0, it means that user i exits demand response; C i,t is the incentive that user i gets after the response is completed; B t Represents the budget for this round of demand response.

5. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios as claimed in claim 4 is characterized in that: The final response status is determined based on the total load reduced in the demand response. During the demand response process, the selected user has the right to participate in or exit the demand response at any time. However, as long as the load reduced reaches or exceeds the load required to be reduced, it is deemed that the user has participated in the demand response, otherwise it is deemed to have exited the demand response.

6. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios as claimed in claim 1, characterized in that: The method for coordinated intraday dispatching of provincial and local loads based on orderly power consumption scenarios is characterized in that the method in which the regional power grid in S2 selects users who have registered to participate in demand response is: In the tth round of demand response, the regional power grid solves the following optimization problem to determine the optimal user participation combination: Among them, in the tth round of demand response, the estimated response probability of any user i at node j is: In the first round of demand response, all users have not been selected. For any user i, A i,0 =1,b i,0 =0 After the tth round of demand response, by updating user i’s A i and b i To achieve the learning of user response behavior, for any user i selected at node j, there is A i,t+1 =A i,t +(d j,t )2 b i,t+1 =b i,t +d j,t r i,t Among them, r i,t is the reward of user i. If user i is not selected, then A i and b i constant.

7. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios according to claim 1, characterized in that: The orderly coordinated control and order restoration strategy of the provincial and local power grids in S4 is to conduct coordinated control and downward adjustment of the provincial and local power grids when demand response is not fully implemented and the load is greater than the power supply capacity; in the tight balance of power during the day, there is sufficient power supply capacity to restore the orderly power load order.

8. The method for coordinated daytime dispatching of provincial and local loads based on orderly power consumption scenarios as claimed in claim 7, characterized in that: The optimization goal of coordinated control and sequential restoration of provincial and local power grids is to minimize the operation and maintenance costs of thermal power units and the economic losses of power outages at system nodes: Among them, q is a certain period of time within a day; the total cost is C q ; a j 、b j and c j is a constant. j,q and Respectively represent the power generation and load of node j; Ω G and Ω fh They are the node set where thermal generators are installed and the load node set. fh is the load outage cost coefficient of node j; represents the power supply status of node j as a binary variable, where Indicates power supply, Indicates a power outage. The constraints of the objective function include thermal power unit output constraints, ramp constraints, system flow constraints, and branch flow safety constraints: in, Respectively represent the set of thermal power units and renewable energy units at node j; Ω j is the set of all nodes connected to node j. ij,q is the active power flow from node i to node j in period q. and are the minimum and maximum output limits of the generator at node j, respectively. and is the minimum and maximum ramp rate of the thermal generator at node j. j,q is the phase angle of node j in period q. ij is the reactance of the branch between nodes i and j. is the aggregated load of the regional power grid at node j during period q.

Images