Power distribution network energy system multi-objective optimization method and device based on cloud side-end cooperation

Abstract

The invention provides a power distribution network energy system multi-objective optimization method and device based on cloud side-end cooperation, and relates to an intelligent power distribution network control technology, and the method comprises the steps: determining at least one target energy system in each energy system and a target scheduling demand of each target energy system in response to an energy scheduling demand, receiving at least one computing task request sent by the end-side equipment in each energy system through the corresponding side-end cluster control unit, and determining a computing task to be allocated; based on a target scheduling demand of each target energy system, transmission delay time between each end-side device and different side-end cluster control units, and transmission delay time between the different side-end cluster control units and a cloud device, the target energy system is scheduled. And calling a target model constructed by taking the minimum total power supply cost of the single energy system and the minimum total delay time of all the energy systems as targets, and carrying out optimization solution on the target model to obtain an optimal scheduling scheme. The economical efficiency and the response speed of the system can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent distribution network control technology, and specifically to a multi-objective optimization method for a distribution network energy system based on cloud-edge-end collaboration, a multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration, a machine-readable storage medium and a terminal device. Background Art

[0002] With the continuous advancement of the energy revolution, as an important physical carrier of the energy Internet, concepts such as energy systems have emerged in this context. Energy systems integrate multiple energy forms such as electricity, heat, and gas in the region by utilizing advanced physical information technology and innovative management models, and achieve coordinated planning, coordinated operation, interactive response, and mutual complementarity among multiple heterogeneous energy subsystems. While meeting the diversified energy needs within the system, it effectively improves the comprehensive utilization efficiency of energy and promotes sustainable energy development.

[0003] With the development of new generation information technology, distributed power supply plays an important role in power grid operation. Under the framework of "power supply-grid-load-energy storage", providing efficient and reliable information processing methods through cloud-edge collaborative communication and computing networks is the key to ensuring the long-term stable operation of the system. At this stage, the optimization and scheduling problems of integrated energy systems are mainly focused on single or multiple energy systems, ignoring their coordinated operation. With the continuous development and progress of energy systems and the gradual maturity of operating modes, users' demand for multi-energy loads is getting higher and higher. The economic efficiency of a single optimized integrated energy system can no longer meet users' multi-energy load needs and the direction of social development. At the same time, the existing energy system scheduling cannot jointly optimize the total power supply cost and the overall system delay. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a multi-objective optimization method for a distribution network energy system based on cloud-edge-end collaboration, a multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration, a machine-readable storage medium and a terminal device to solve the above-mentioned problems.

[0005] In order to achieve the above-mentioned objectives, the first aspect of the present application provides a multi-objective optimization method for a distribution network energy system based on cloud-edge-end collaboration, the method comprising:

[0006] In response to the energy scheduling requirements, determine at least one target energy system in each energy system and the target scheduling requirements of each target energy system, and receive at least one computing task request sent by the end-side device in each energy system via the corresponding edge-end cluster control unit to determine the computing task to be assigned;

[0007] Based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, a target model constructed with the goal of minimizing the total power supply cost of a single energy system and minimizing the total delay time of all energy systems is called, and the target model is optimized and solved to obtain an optimal scheduling solution, which includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task to be assigned;

[0008] The optimal output power of each target energy system is sent to the edge cluster control unit corresponding to each target energy system, and each computing task request is allocated to the corresponding optimal edge cluster control unit.

[0009] Optionally, determining at least one target energy system among the energy systems includes:

[0010] Obtain dispatchable capacity and energy quotes for each energy system;

[0011] The dispatchable capacity of each energy system is cumulatively calculated in order of the energy quotations of each energy system from low to high, until the sum obtained by the current cumulative calculation operation is not less than the energy dispatch demand, and the energy system corresponding to the current cumulative calculation operation is determined as the target energy system.

[0012] Optionally, determine the target scheduling requirements of each target energy system, including:

[0013] Sort each target energy system in order of energy price from low to high;

[0014] The target dispatching requirements of the first N-1 target energy systems are determined as the dispatchable capacity of each target energy system, and the difference between the energy dispatching requirement and the sum of the dispatchable capacities of the first N-1 target energy systems is determined as the target dispatching requirement of the Nth target energy system.

[0015] Optionally, the target model includes:

[0016] Energy system dispatch model and system delay model;

[0017] The energy system dispatch model is:

[0018] f1=min(f om +f buy -f sell )

[0019] Among them, f1 is the total electricity supply cost of the energy system, f om represents the operation and maintenance cost of the energy system, f buyrepresents the cost of the energy system purchasing electricity from the grid, f sell It represents the revenue that the energy system receives from selling excess electricity to the grid;

[0020] The system delay model is:

[0021]

[0022] Among them, f2 is the total system delay time, β j,i represents the weight coefficient of each energy system, n represents the number of weight coefficients of the energy system, T j,i It is the total delay time served by the i-th end-side device uploading the computing task request to the cloud device, and then sending it to the j-th edge cluster control unit through the cloud device.

[0023] Optionally, optimizing and solving the target model to obtain an optimal scheduling solution for each target energy system includes:

[0024] Based on the system delay model, with the goal of minimizing the total system delay time, the optimal edge cluster control unit corresponding to each computing task to be assigned is determined through a preset computing task allocation algorithm;

[0025] The computing task allocation algorithm comprises:

[0026] S10, obtaining the current resource utilization rate of the edge cluster control unit corresponding to each target energy system, and determining that the operation state of the edge cluster control unit whose current resource utilization rate is lower than the resource utilization rate threshold is a low-load state;

[0027] S11, sorting edge cluster control units in a low-load state from low to high according to resource utilization rates to build an edge cluster sequence;

[0028] S12, determining the task set length based on the number of computing tasks to be assigned and the number of edge cluster control units in the edge cluster sequence;

[0029] S13, sorting the computing tasks to be assigned from early to late according to the computing task request time of each computing task to be assigned, establishing a computing task sequence, and dividing the computing tasks in the computing task sequence into multiple task sets according to the length of the task set;

[0030] S14. Allocate the computing tasks in the first task set in the computing task sequence to the first edge cluster control unit of the edge cluster sequence, delete the first task set in the computing task sequence from the computing task sequence, update the computing task sequence, and determine whether there are any unallocated task sets in the computing task sequence. If so, return to step S10; otherwise, complete the allocation of computing tasks.

[0031] Optionally include:

[0032] Taking the total power supply cost of each target energy system as the minimum as the goal, the energy system scheduling model is optimized and solved by a preset optimization algorithm to obtain the optimal output power that meets the target scheduling requirements of each target energy system;

[0033] The preset optimization algorithm includes:

[0034] S20, initializing the particle swarm, determining the number of particles in the particle swarm and the initial position of each particle, mapping the output power of each target energy system to the position of each particle, and determining the dimension of each particle based on the number of target energy systems;

[0035] S21. Determine the individual extreme value and the group extreme value of each particle based on the energy system scheduling model;

[0036] S22, determining the current simulation temperature based on the group extreme value;

[0037] S23, determining the annealing fitness value of each particle at the current simulation temperature based on the individual extreme value and the group extreme value of each particle;

[0038] S24, according to the annealing fitness value of each particle, determine the candidate population extreme value from all particles through the roulette algorithm, and update the current population extreme value with the candidate population extreme value;

[0039] S25, updating the speed and position of each particle based on the updated group extreme value;

[0040] S26. Determine the individual extreme value and the group extreme value of each particle after update based on the energy system scheduling model;

[0041] S27, lower the current simulation temperature. If the convergence condition is not met, return to step S23. If the convergence condition is met, stop searching and output the current group extreme value as the optimal scheduling solution.

[0042] Optionally, determining the individual extreme value and the group extreme value of each particle based on the energy system scheduling model includes:

[0043] Calculating the fitness value of each particle based on the energy system scheduling model, and determining the individual extreme value and the group extreme value of each particle according to the fitness value of each particle;

[0044] Based on the individual extreme value and group extreme value of each particle, the annealing fitness value of each particle at the current simulation temperature is determined, including:

[0045] The annealing fitness value of each particle at the current simulation temperature is determined by the following formula:

[0046]

[0047] Among them, f(P i ) represents the fitness value of each particle, f(P g ) represents the optimal fitness value among all particles, N represents the number of particles, and t represents the current simulation temperature.

[0048] Optionally, update the velocity and position of each particle based on the updated population extremum, including:

[0049] Update the velocity and position of each particle using the following formula:

[0050]

[0051] x i,j (t+1)=x i,j (t)+v i,j (t+1),j=1,2,...,d

[0052] in, C = c1 + c2, c1 and c2 are learning factors, r1 and r2 are random numbers between (0, 1), P g ' ,j is the updated population extreme value, v i,j is the velocity of the i-th particle in the j-th dimension, P i,j is the fitness value of the jth particle in the jth dimension, x i,j The position of the i-th particle in the j-th dimension.

[0053] Optionally, determining the individual extreme value and the group extreme value of each particle after update based on the energy system scheduling model includes:

[0054] If the fitness value of the current particle after the update is better than the fitness value before the update, the current position of the current particle is updated with the updated position of the current particle; otherwise, an update probability is determined, and based on the update probability, the current position of the current particle is updated with the updated position of the current particle.

[0055] Optionally, determining an update probability includes:

[0056] The update probability is determined by the following updates:

[0057]

[0058] Among them, T i represents the current temperature, Δf represents the difference between the fitness value of the current particle after the update and the fitness value of the current particle before the update, and e is a natural constant.

[0059] Optionally, determining a current simulation temperature based on a population extreme value includes:

[0060] The current simulated temperature is determined by the following formula:

[0061]

[0062] Lower the current simulation temperature, including:

[0063] Reduce the current simulated temperature by the following formula:

[0064] T k-1 =αT k

[0065] Among them, T k-1 represents the simulated temperature after reduction, α is the cooling coefficient, T k represents the simulated temperature before the decrease, T0 represents the current simulated temperature, f(P g ) represents the optimal fitness value among all particles.

[0066] In a second aspect of the present application, a multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration is provided, the device comprising:

[0067] A demand response module is configured to determine at least one target energy system in each energy system and a target scheduling requirement of each target energy system in response to an energy scheduling requirement, and receive at least one computing task request sent by an edge-side device in each energy system via a corresponding edge-side cluster control unit, and determine a computing task to be assigned;

[0068] The scheduling optimization module is configured to call a target model constructed with the goal of minimizing the total power supply cost of a single energy system and the total delay time of all energy systems as the goal based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, and optimize and solve the target model to obtain an optimal scheduling solution, wherein the optimal scheduling solution includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task request;

[0069] The scheduling module is configured to send the optimal output power of each target energy system to the edge cluster control unit corresponding to each target energy system, and allocate each computing task request to the corresponding optimal edge cluster control unit.

[0070] According to a third aspect of the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor is configured to execute the method as described above.

[0071] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0072] This application can reduce the overall delay of multiple energy systems while ensuring the cost efficiency of power supply of a single energy system, effectively improving the economy and system response speed of the energy system.

[0073] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0075] Figure 1 A method flow chart of a multi-objective optimization method for a distribution network energy system based on cloud-edge-end collaboration provided in a preferred embodiment of the present application;

[0076] Figure 2 A schematic diagram of scheduling control provided for a preferred embodiment of the present application;

[0077] Figure 3 A schematic diagram of a cloud-edge-end collaboration model for an energy system provided for a preferred embodiment of the present application;

[0078] Figure 4 Flow chart of optimization algorithm provided for the preferred embodiment of the present application;

[0079] Figure 5 A schematic block diagram of a multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration provided in a preferred embodiment of the present application;

[0080] Figure 6 A schematic diagram of a terminal device provided for a preferred embodiment of the present application.

[0081] Description of Reference Numerals

[0082] 10-terminal device, 100-processor, 101-memory, 102-computer program. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0084] It should be noted that the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0085] like Figure 1 As shown, the first aspect of the present application provides a multi-objective optimization method for a distribution network energy system based on cloud-edge-end collaboration, the method comprising:

[0086] S100, in response to energy scheduling requirements, determining at least one target energy system in each energy system and a target scheduling requirement of each target energy system, and receiving at least one computing task request sent by an edge-side device in each energy system via a corresponding edge-end cluster control unit, and determining a computing task to be assigned;

[0087] S200, based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, call a target model constructed with the goal of minimizing the total power supply cost of a single energy system and minimizing the total delay time of all energy systems, optimize and solve the target model to obtain an optimal scheduling solution, the optimal scheduling solution includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task request;

[0088] S300. Send the optimal output power of each target energy system to the edge cluster control unit corresponding to each target energy system, and allocate each computing task to be allocated to the corresponding optimal edge cluster control unit.

[0089] In this way, the present application can reduce the overall delay of the multi-energy system while ensuring the cost efficiency of the power supply of a single energy system, effectively improving the economy and system response speed of the energy system.

[0090] It is understandable that an energy system may include a variety of distributed resources, such as power supply units, battery energy storage units, electric vehicles, gas turbines, electric boilers, electric refrigeration units, trigeneration units and other controllable loads. Figure 2 and Figure 3As shown, the energy system cloud-edge-end collaborative model constructed by the present application includes the distributed management of energy system units by edge clusters, the management and collaborative control of edge clusters by the cloud, and the coordination of the scheduling and allocation of overall analysis and computing tasks. Among them, each distributed resource is distributed on the end side, and resources are aggregated through a multi-energy energy station. For example, loads such as electric vehicles and electric boilers can be replenished through a multi-energy energy station, and power supply units and battery energy storage units provide power to the multi-energy station. At the same time, local control and data collection can be performed through the multi-energy energy station, for example, collecting power supply power data of the power supply unit, charging and discharging power data of the battery energy storage unit, and power consumption data of users. One or more multi-energy energy stations constitute a resource cluster, wherein the energy system edge cluster is the information management platform of the energy system control unit, which has the functions of information collection and reporting, control decomposition, monitoring and alarm, etc., provides secure access to various distributed resources, realizes the monitoring, prediction, analysis and control functions of the energy system control unit, and transmits the control commands of the cloud control application to the energy system control unit; the energy system control unit is responsible for the aggregation and management of distributed resources, and accesses the edge cluster, aggregates the distributed resources of the agent according to the scheduling requirements, and sends the operation information of the distributed resources to the energy system edge cluster as the minimum control unit, the minimum reporting unit and the minimum clearing unit, and participates in the cloud control application, wherein the edge cluster can be a physical edge cluster or a virtual edge cluster. It can be understood that in this application, the multi-energy energy station is used as the energy system control unit, and the edge cluster control unit on the side exchanges data with one or more multi-energy energy stations. For example, the edge cluster control unit receives resource parameters such as power supply, load power demand and distributed resource topology from the multi-energy energy station, and uploads the received information to the cloud. The cloud processes the received information through the deployed cloud collaborative management and control platform to make global scheduling and control decisions. For example, the cloud can determine which energy systems should increase or decrease the power supply according to the status and demand of each energy system, and generate corresponding energy scheduling instructions, decompose the energy scheduling instructions into multiple computing tasks and send them to the edge cluster control unit for subsequent processing. In this application, a distribution network energy system model is first constructed, including an energy output model, a distributed energy power supply model, and an energy storage system model. The distribution network energy system relies on distributed multi-energy to supply the power demand of the local power grid. The distributed energy power supply unit has the characteristics of small capacity and small scale, so that it presents the characteristics of decentralized and modular layout near the user. For example, energy output refers to renewable natural resources that can be converted into electrical energy, such as wind power generation, etc. Distributed energy can be distributed photovoltaic units, etc. Energy storage devices usually operate in three modes: charging, discharging and standby. It can be understood that in order to ensure the reliable operation of energy storage devices, energy storage devices cannot perform power supply and discharge operations at the same time.

[0091] In step S100, the cloud first determines the target energy system according to the electricity market clearing result or the dispatching control demand. Specifically, determining at least one target energy system in each energy system includes: obtaining the dispatchable capacity and energy quotation of each energy system; performing cumulative calculation operations on the dispatchable capacity of each energy system in order from low to high according to the energy quotation of each energy system, until the sum obtained by the current cumulative calculation operation is not less than the energy dispatching demand, and determining that the energy system corresponding to the current cumulative calculation operation is the target energy system. For example, taking four energy systems A, B, C and D as an example, each energy system has a different adjustable capacity and quotation. For example, the quotations of energy systems A, B, C and D are 0.1 / kWh, 0.1 / kWh, 0.2 / kWh, 0.3 / kWh and 0.4 / kWh, respectively, and the corresponding adjustable capacities are 10kWh, 20kWh, 30kWh and 40kWh, respectively. If the power dispatching control center issues a demand of 100kWh, the cloud first checks the quotation of the energy system, and then calls the adjustable capacity from the lowest quotation until the demand is met. Determine the target dispatching demand of each target energy system, including: sorting each target energy system in order from low to high according to the energy quotation; determining the target dispatching demand of the first N-1 target energy systems as the dispatchable capacity of each target energy system, and determining the difference between the energy dispatching demand and the sum of the dispatchable capacity of the first N-1 target energy systems as the target dispatching demand of the Nth target energy system. For example, taking the 80kWh demand issued by the power dispatching control center as an example, since the quotation of energy system A is the lowest, the cloud first calls all 10kWh of A's adjustable capacity, and the remaining demand is 70kWh, and then calls all 20kWh of B's ​​adjustable capacity, and the remaining demand is 50kWh, and continues to call all 30kWh of C's adjustable capacity, and the remaining demand is 20kWh, and finally calls D's 20kWh adjustable capacity. That is, the target dispatching demands of energy system A, energy system B, energy system C, and energy system D are 10kWh, 20kWh, 30kWh, and 20kWh, respectively.

[0092] In step S200, the present application is based on multi-objective optimization scheduling that collaboratively considers energy and information. In terms of energy, the economic efficiency of a single energy system is considered, and a scheduling model is constructed with the goal of minimizing the power supply cost of a single energy system. In terms of information, the cloud-edge-end synergy of computing resources in the multi-energy system scheduling process is considered, and a scheduling model is constructed with the goal of minimizing the total delay of the cloud-edge-end system of the multi-energy system. It can be understood that when the output of distributed energy fluctuates greatly, an energy storage system can be used for power compensation while adapting to changes in load demand. If the output of distributed energy is not enough to meet the load demand, the distribution network operator will purchase electricity from the main power grid. In the case of oversupply of electricity, the operator will sell the excess electricity back to the power grid. For example, according to the electricity price during the scheduling period, the optimal power purchase and sales strategy for a single energy system is determined to ensure optimal economic efficiency. In this application, the target model includes: an energy system scheduling model and a system delay model; wherein the energy system scheduling model is:

[0093] f1=min(f om +f buy -f sell )

[0094] Among them, f1 is the total electricity supply cost of the energy system, f om represents the operation and maintenance cost of the energy system, f buy represents the cost of the energy system purchasing electricity from the grid, f sell It represents the revenue that the energy system receives from selling excess electricity to the grid;

[0095] The system delay model is:

[0096]

[0097] Among them, f2 is the total system delay time, β j,i represents the weight coefficient of each energy system, n represents the number of weight coefficients of the energy system, T j,i It is the total delay time served by the i-th end-side device uploading the computing task request to the cloud device, and then sending it to the j-th edge cluster control unit through the cloud device.

[0098] It is understandable that the operation and maintenance costs of the energy system are directly related to the output power of the energy system. For example, the mapping relationship between the output power of each energy system and the operation cost of the energy system, i.e., the operation and maintenance costs, can be predetermined. For example, let f om=f×α, where f is the output power of the energy system and α is the operating cost conversion coefficient. The cost of the energy system purchasing electricity from the power grid or the income of the energy system selling excess electricity to the power grid can be calculated based on the real-time electricity price during the scheduling period. In the present application, the constraints of the energy system scheduling model may include: the output power of the current energy system is not greater than the maximum output power of the current energy system, and the output power of the current energy system is not less than the target scheduling demand power of the current energy system.

[0099] In step S200, the target model is optimized and solved to obtain the optimal scheduling solution for each target energy system, including:

[0100] Based on the energy system scheduling model, with the goal of minimizing the total system delay time, the optimal edge cluster control unit corresponding to each computing task request is determined through a preset computing task allocation algorithm;

[0101] The computing task allocation algorithm comprises:

[0102] S10. Obtain the current resource utilization rate of the edge cluster control unit corresponding to each target energy system, and determine that the operating state of the edge cluster control unit whose current resource utilization rate is lower than the resource utilization rate threshold is a low-load state. For example, the processor utilization rate, memory utilization rate and other parameters of the edge cluster control unit can be used as the current resource utilization rate of the edge cluster control unit, and the current resource utilization rate of the edge cluster control unit is compared with the preset resource utilization rate threshold, and the edge cluster control unit whose current resource utilization rate is lower than the resource utilization rate threshold is determined to be in a low-load state. The obtained sequence of edge cluster control units in a low-load state is represented as D, A i =C i <θ i , where A i represents the decision principle of the ith edge cluster, C i represents the computing resource utilization rate of the i-th edge cluster, θ i Indicates the computing resource usage threshold of the i-th edge cluster.

[0103] S11. Sort the edge cluster control units in low-load state from low to high according to resource utilization rate to construct an edge cluster sequence. Let the number of edge cluster control units contained in sequence D be N, and sort the edge cluster control units in sequence D from low to high according to resource utilization rate to form a sequence D'={S1,S2,...,S N}, where S1,S2,...,S NRepresents the sorted edge cluster control units. It is understandable that each energy system control unit can correspond to multiple edge cluster control units, and the edge cluster control unit corresponding to each target energy system can be any edge cluster control unit corresponding to each target energy system, or all edge cluster control units corresponding to each target energy system. For example, the edge cluster control unit in the low-load state among all edge cluster control units corresponding to each target energy system is selected to construct an edge cluster sequence.

[0104] S12: Determine the task set length based on the number of computing tasks to be assigned and the number of edge cluster control units in the edge cluster sequence. The task set length N0 is expressed as N0=[N t / N], where N t The number of computing tasks to be assigned.

[0105] S13, sorting the computing tasks to be assigned from earliest to latest according to the computing task request time of each computing task to be assigned, establishing a computing task sequence, and dividing the computing tasks in the computing task sequence into multiple task sets according to the length of the task set. Sort the current computing tasks to be assigned, for example, sort them according to the time sequence of the computing task requests, starting from the computing task with the earliest request time, and selecting a fixed number of task sets N0.

[0106] S14, assign the computing tasks in the first task set in the computing task sequence to the first edge cluster control unit of the edge cluster sequence, delete the first task set in the computing task sequence from the computing task sequence, update the computing task sequence, and determine whether there are any unassigned task sets in the computing task sequence. If so, return to step S10, otherwise, complete the assignment of computing tasks. Specifically, convert the original task sequence into a new sequence composed of task sets T a , T a The task set in D is assigned to the first edge cluster in D'. After completion, the assigned task set is removed from T a Exit, where the sequence T a It is expressed as:

[0107]

[0108] in, Represents the computing task to be assigned.

[0109] It can be understood that the transmission delay between each end-side device and the edge cluster control unit, and the transmission delay between each edge cluster control unit and the cloud device can be predetermined, and it varies according to the transmission distance and transmission environment between the devices. During scheduling, based on the above-mentioned system delay model, with the minimum total system delay time as the goal, the optimal allocation strategy for each computing task to be allocated can be obtained through the preset computing task allocation algorithm, that is, the cloud device allocates each computing task to be allocated to the corresponding optimal edge cluster control unit according to the minimum total system delay time based on the preset computing task allocation algorithm. Among them, existing algorithms such as particle swarm algorithm, genetic algorithm, etc. can be used to optimize and solve the system delay model.

[0110] In step S200, the target model is optimized and solved to obtain the optimal scheduling plan for each target energy system, including: taking the total power supply cost of each target energy system as the minimum, optimizing and solving the energy system scheduling model through a preset optimization algorithm to obtain the optimal output power that meets the target scheduling requirements of each target energy system; the particle swarm optimization algorithm based on simulated annealing in this application is used to optimize and solve the energy system scheduling model, and the local optimal solution of the algorithm is jumped out with a certain probability to obtain the global optimal solution, that is, the optimal output power of the energy system. Then, if Figure 4 As shown, the preset optimization algorithm includes:

[0111] S20, initialize the particle swarm, determine the number of particles in the particle swarm and the initial position of each particle, map the output power of each target energy system to the position of each particle, for example, the position of the particle represents a solution, that is, the output power of the power supply of the distribution network energy system; at the same time, determine the dimension of each particle based on the number of target energy systems, for example, if the number of target energy systems determined according to the scheduling requirements is 4, then the particle dimension is 4.

[0112] S21. Determine the individual extreme value and group extreme value of each particle based on the energy system scheduling model, including: calculating the fitness value of each particle based on the energy system scheduling model, and determining the individual extreme value and group extreme value of each particle according to the fitness value of each particle. Determine the fitness function based on the target model, for example, take f1 as the fitness function, take the value of f1 as the fitness value, calculate the individual extreme value and group extreme value of each particle respectively, and calculate the individual extreme value and group extreme value of each particle respectively. Among them, the individual extreme value is the optimal fitness value of the current particle, and the group extreme value is the optimal fitness value among all particles, wherein the smaller the fitness value, the better the fitness value.

[0113] S22. Determine the current simulation temperature based on the group extreme value. Specifically, the current simulation temperature is determined by the following formula:

[0114]

[0115] Where T0 represents the current simulation temperature, f(P g ) represents the extreme value of the group.

[0116] S23, determining the annealing fitness value of each particle at the current simulation temperature based on the individual extreme value and the group extreme value of each particle, including:

[0117] The annealing fitness value of each particle at the current simulation temperature is determined by the following formula:

[0118]

[0119] Among them, f(P i ) represents the fitness value of each particle, f(P g ) represents the optimal fitness value among all particles, N represents the number of particles, and t represents the current simulation temperature.

[0120] S24. According to the annealing fitness value of each particle, determine the candidate group extreme value from all particles through the roulette algorithm, and update the current group extreme value with the candidate group extreme value. Among them, the roulette algorithm selects a particle's own optimal solution to replace the group optimal solution through cumulative probability. The higher the fitness value of the particle, the greater the probability of being selected. For example, generate a random number between 0 and 1, and calculate the cumulative probability according to the annealing fitness value of each particle: According to the cumulative probability, the individual optimal position of the rth particle that meets the conditions is used as the candidate population extreme value P′ to replace the global optimal position, where the condition is: comfit(r-1)<f(P b )<comfit(r), where f(P b ) represents the individual extreme value, that is, the individual optimal position.

[0121] S25. Update the speed and position of each particle based on the updated group extreme value, including:

[0122] Update the velocity and position of each particle using the following formula:

[0123]

[0124] x i,j (t+1)=x i,j (t)+v i,j (t+1),j=1,2,...,d

[0125] in, C = c1 + c2, c1 and c2 are learning factors, r1 and r2 are random numbers between (0, 1), P g ' ,j is the updated population extreme value, vi,j is the velocity of the i-th particle in the j-th dimension, P i,j is the fitness value of the jth particle in the jth dimension, x i,j The position of the i-th particle in the j-th dimension.

[0126] S26. Determine the individual extreme value and group extreme value of each particle after the update based on the energy system scheduling model. If the fitness value of the current particle after the update is better than the fitness value before the update, update the current position of the current particle with the updated position of the current particle; otherwise, determine the update probability, and based on the update probability, update the current position of the current particle with the updated position of the current particle. Specifically, calculate the update target value of each particle, and calculate the difference between the fitness value of the new particle and the fitness value of the previous generation example Δf=f(P t+1 )-f(P t ), if Δf < 0, the position of the new particle is accepted, otherwise the position is accepted according to a certain probability. This application determines the update probability through the following update:

[0127]

[0128] Among them, T i represents the current temperature, Δf represents the difference between the fitness value of the current particle after the update and the fitness value of the current particle before the update, and e is a natural constant.

[0129] S27. Lower the current simulation temperature. If the convergence condition is not met, return to step S23. If the convergence condition is met, stop searching and output the current group extreme value as the optimal scheduling solution, that is, the optimal output power of each energy system. In this way, the present application optimizes the output power of a single energy system during scheduling to optimize its economy. At the same time, the computing task allocation of the entire system is optimized to minimize the total delay time of the system, that is, to minimize f1+f2 during scheduling.

[0130] This application reduces the current simulated temperature by the following formula:

[0131] T k-1 =αT k

[0132] Among them, T k-1 represents the simulated temperature after reduction, α is the cooling coefficient, T k represents the simulated temperature before the decrease, T0 represents the current simulated temperature, f(P g ) represents the optimal fitness value among all particles.

[0133] like Figure 5 As shown, in a second aspect of the present application, a multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration is provided, and the method includes:

[0134] A demand response module is configured to determine at least one target energy system in each energy system and a target scheduling requirement of each target energy system in response to an energy scheduling requirement, and receive at least one computing task request sent by an edge-side device in each energy system via a corresponding edge-side cluster control unit, and determine a computing task to be assigned;

[0135] The scheduling optimization module is configured to call a target model constructed with the goal of minimizing the total power supply cost of a single energy system and the total delay time of all energy systems as the goal based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, and optimize and solve the target model to obtain an optimal scheduling solution, wherein the optimal scheduling solution includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task request;

[0136] The scheduling module is configured to send the optimal output power of each target energy system to the edge cluster control unit corresponding to each target energy system, and allocate each computing task request to the corresponding optimal edge cluster control unit.

[0137] It is understandable that the technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0138] According to a third aspect of the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor is configured to execute the method as described above.

[0139] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0140] like Figure 6 FIG. 1 is a schematic diagram of a terminal device provided in an embodiment of the present application. Figure 6 As shown, the terminal device 10 of this embodiment includes: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the steps in the above method embodiment are implemented. Alternatively, when the processor 100 executes the computer program 102, the functions of each module / unit in the above device embodiments are implemented.

[0141] Exemplarily, the computer program 102 may be divided into one or more modules / units, one or more modules / units are stored in the memory 101, and are executed by the processor 100 to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 102 in the terminal device 10.

[0142] The terminal device 10 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will appreciate that Figure 6 It is only an example of the terminal device 10 and does not constitute a limitation of the terminal device 10. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0143] The processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0144] The memory 101 may be an internal storage unit of the terminal device 10, such as a hard disk or memory of the terminal device 10. The memory 101 may also be an external storage device of the terminal device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 10. Further, the memory 101 may also include both an internal storage unit of the terminal device 10 and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the terminal device 10. The memory 101 may also be used to temporarily store data that has been output or is to be output.

[0145] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0146] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0147] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration, characterized in that: The method comprises: In response to the energy scheduling requirements, determine at least one target energy system in each energy system and the target scheduling requirements of each target energy system, and receive at least one computing task request sent by the end-side device in each energy system via the corresponding edge-end cluster control unit to determine the computing task to be assigned; Based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, a target model constructed with the goal of minimizing the total power supply cost of a single energy system and minimizing the total delay time of all energy systems is called, and the target model is optimized and solved to obtain an optimal scheduling solution, which includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task to be assigned; The optimal output power of each target energy system is sent to the edge cluster control unit corresponding to each target energy system, and each computing task to be assigned is assigned to the corresponding optimal edge cluster control unit.

2. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 1 is characterized in that: Determining at least one target energy system among the energy systems, including: Obtain dispatchable capacity and energy quotes for each energy system; The dispatchable capacity of each energy system is cumulatively calculated in order of the energy quotations of each energy system from low to high, until the sum obtained by the current cumulative calculation operation is not less than the energy dispatch demand, and the energy system corresponding to the current cumulative calculation operation is determined as the target energy system.

3. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 2 is characterized in that: Determine the target dispatch requirements for each target energy system, including: Sort each target energy system in order of energy price from low to high; The target dispatching requirements of the first N-1 target energy systems are determined as the dispatchable capacity of each target energy system, and the difference between the energy dispatching requirement and the sum of the dispatchable capacities of the first N-1 target energy systems is determined as the target dispatching requirement of the Nth target energy system.

4. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 1 is characterized in that: The target model comprises: Energy system dispatch model and system delay model; The energy system dispatch model is: f1=min(f om +f buy -f sell ) Among them, f1 is the total electricity supply cost of the energy system, f om represents the operation and maintenance cost of the energy system, f buy represents the cost of the energy system purchasing electricity from the grid, f sell It represents the revenue that the energy system receives from selling excess electricity to the grid; The system delay model is: Among them, f2 is the total system delay time, β j,i represents the weight coefficient of each energy system, n represents the number of weight coefficients of the energy system, T j,i It is the total delay time served by the i-th end-side device uploading the computing task request to the cloud device, and then sending it to the j-th edge cluster control unit through the cloud device.

5. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 4 is characterized in that: The target model is optimized and solved to obtain the optimal scheduling scheme for each target energy system, including: Based on the system delay model, with the goal of minimizing the total system delay time, the optimal edge cluster control unit corresponding to each computing task to be assigned is determined through a preset computing task allocation algorithm; The computing task allocation algorithm comprises: S10, obtaining the current resource utilization rate of the edge cluster control unit corresponding to each target energy system, and determining that the operation state of the edge cluster control unit whose current resource utilization rate is lower than the resource utilization rate threshold is a low-load state; S11, sorting edge cluster control units in a low-load state from low to high according to resource utilization rates to build an edge cluster sequence; S12, determining the task set length based on the number of computing tasks to be assigned and the number of edge cluster control units in the edge cluster sequence; S13, sorting the computing tasks to be assigned from early to late according to the computing task request time of each computing task to be assigned, establishing a computing task sequence, and dividing the computing tasks in the computing task sequence into multiple task sets according to the length of the task set; S14. Allocate the computing tasks in the first task set in the computing task sequence to the first edge cluster control unit of the edge cluster sequence, delete the first task set in the computing task sequence from the computing task sequence, update the computing task sequence, and determine whether there are any unallocated task sets in the computing task sequence. If so, return to step S10; otherwise, complete the allocation of computing tasks.

6. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 4 is characterized in that: The target model is optimized and solved to obtain the optimal scheduling scheme for each target energy system, including: Taking the total power supply cost of each target energy system as the minimum as the goal, the energy system scheduling model is optimized and solved by a preset optimization algorithm to obtain the optimal output power that meets the target scheduling requirements of each target energy system; The preset optimization algorithm includes: S20, initializing the particle swarm, determining the number of particles in the particle swarm and the initial position of each particle, mapping the output power of each target energy system to the position of each particle, and determining the dimension of each particle based on the number of target energy systems; S21, determining the individual extreme value and the group extreme value of each particle based on the energy system scheduling model; S22, determining the current simulation temperature based on the group extreme value; S23, determining the annealing fitness value of each particle at the current simulation temperature based on the individual extreme value and the group extreme value of each particle; S24, according to the annealing fitness value of each particle, determine the candidate population extreme value from all particles through the roulette algorithm, and update the current population extreme value with the candidate population extreme value; S25, updating the speed and position of each particle based on the updated group extreme value; S26. Determine the individual extreme value and the group extreme value of each particle after update based on the energy system scheduling model; S27, lower the current simulation temperature. If the convergence condition is not met, return to step S23. If the convergence condition is met, stop searching and output the current group extreme value as the optimal scheduling solution.

7. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 6 is characterized in that: Determining individual extreme values ​​and group extreme values ​​of each particle based on the energy system scheduling model includes: Calculating the fitness value of each particle based on the energy system scheduling model, and determining the individual extreme value and the group extreme value of each particle according to the fitness value of each particle; Based on the individual extreme value and group extreme value of each particle, the annealing fitness value of each particle at the current simulation temperature is determined, including: The annealing fitness value of each particle at the current simulation temperature is determined by the following formula: Among them, f(P i ) represents the fitness value of each particle, f(P g ) represents the optimal fitness value among all particles, N represents the number of particles, and t represents the current simulation temperature.

8. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 6 is characterized in that: Update the velocity and position of each particle based on the updated population extremum, including: Update the velocity and position of each particle using the following formula: x i,j (t+1)=x i,j (t)+v i,j (t+1),j=1,2,...,d in, C = c1 + c2, c1 and c2 are learning factors, r1 and r2 are random numbers between (0, 1), P g ' ,j is the updated population extreme value, v i,j is the velocity of the i-th particle in the j-th dimension, P i,j is the fitness value of the jth particle in the jth dimension, x i,j The position of the i-th particle in the j-th dimension.

9. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 6 is characterized in that: Determining the individual extreme value and the group extreme value of each particle after update based on the energy system scheduling model includes: If the fitness value of the current particle after the update is better than the fitness value before the update, the current position of the current particle is updated with the updated position of the current particle; otherwise, an update probability is determined, and based on the update probability, the current position of the current particle is updated with the updated position of the current particle.

10. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 9 is characterized in that: Determine the probability of an update, including: The update probability is determined by the following updates: Among them, T i represents the current temperature, Δf represents the difference between the fitness value of the current particle after the update and the fitness value of the current particle before the update, and e is a natural constant.

11. The multi-objective optimization method for distribution network energy system based on cloud-edge-end collaboration according to claim 6 is characterized in that: Determine the current simulation temperature based on the population extreme value, including: The current simulated temperature is determined by the following formula: Lower the current simulation temperature, including: Reduce the current simulated temperature by the following formula: T k-1 =αT k Among them, T k-1 represents the simulated temperature after reduction, α is the cooling coefficient, T k represents the simulated temperature before the decrease, T0 represents the current simulated temperature, f(P g ) represents the optimal fitness value among all particles.

12. A multi-objective optimization device for a distribution network energy system based on cloud-edge-end collaboration, characterized in that: The device comprises: A demand response module is configured to determine at least one target energy system in each energy system and a target scheduling requirement of each target energy system in response to an energy scheduling requirement, and receive at least one computing task request sent by an edge-side device in each energy system via a corresponding edge-side cluster control unit, and determine a computing task to be assigned; The scheduling optimization module is configured to call a target model constructed with the goal of minimizing the total power supply cost of a single energy system and the total delay time of all energy systems as the goal based on the target scheduling requirements of each target energy system, the transmission delay time between each end-side device and different edge cluster control units, and the transmission delay time between different edge cluster control units and cloud devices, and optimize and solve the target model to obtain an optimal scheduling solution, wherein the optimal scheduling solution includes the optimal output power of each target energy system and the optimal edge cluster control unit corresponding to each computing task request; The scheduling module is configured to send the optimal output power of each target energy system to the edge cluster control unit corresponding to each target energy system, and allocate each computing task request to the corresponding optimal edge cluster control unit.

13. A machine-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor is configured to perform the method according to any one of claims 1 to 11.

14. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 11 are implemented.

Images