Resource-coordinated optimization scheduling methods, devices, and electronic equipment

Abstract

This invention provides a resource-coordinated optimized scheduling method, apparatus, and electronic device, belonging to the field of power system technology. The method includes: constructing a computing network objective function based on computing tasks; solving the computing network objective function to obtain the optimized scheduling result of the computing tasks; determining the computing load demand of the computing tasks based on the optimized scheduling result; determining the target impact data of the power network objective function based on the comparison between the computing load demand and the maximum power supply of distributed generation in the power grid; constructing the power network objective function based on the target impact data; and solving the power network objective function to obtain the optimized scheduling result of the power network, thereby achieving resource-coordinated optimized scheduling. This invention addresses the problem that traditional optimized scheduling methods fail to effectively match the high energy consumption of computing centers with the volatility of the power system, affecting the overall stability and economy of the power system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power systems, in particular to a resource coordination-based optimal scheduling method, a resource coordination-based optimal scheduling device, an electronic device, a machine-readable storage medium and a computer program product. BACKGROUND

[0002] With the rapid development of artificial intelligence, big data and other technologies, the demand for computing power has increased dramatically, and higher requirements have been placed on the stability and sustainability of the power system. As the basic infrastructure for energy transmission, the power grid is the core platform for information processing. In traditional optimal scheduling methods, the optimal scheduling methods for the computing power grid and the power grid are mostly independent of each other, resulting in a mismatch between the high energy consumption of the computing power center and the volatility of the power system, affecting the stability and economy of the overall power system. SUMMARY

[0003] The purpose of the embodiments of the present application is to provide a resource coordination-based optimal scheduling method, device and electronic device to solve the problem that in traditional optimal scheduling methods, the optimal scheduling methods for the computing power grid and the power grid are mostly independent of each other, resulting in a mismatch between the high energy consumption of the computing power center and the volatility of the power system, affecting the stability and economy of the overall power system.

[0004] To achieve the above-mentioned purpose, the embodiments of the present application provide a resource coordination-based optimal scheduling method, comprising:

[0005] constructing a computing power grid target function based on the computing power grid; the computing power grid target function represents the minimization of the running cost of the computing power task;

[0006] solving the computing power grid target function to obtain the optimal scheduling result of the computing power task, and determining the computing power load demand of the computing power task based on the optimal scheduling result of the computing power task;

[0007] determining the target influence data of the power grid target function based on the comparison result of the computing power load demand and the maximum power supply of the distributed power source in the power grid; the power grid target function represents the minimization of the running cost of the power grid;

[0008] constructing a power grid target function based on the target influence data, and solving the power grid target function to obtain the optimal scheduling result of the power grid, so as to realize resource coordination-based optimal scheduling.

[0009] Optionally, the computing power task includes the power consumption of online tasks and the power consumption of offline tasks; the computing power grid target function is constructed based on the computing power grid, comprising:

[0010] constructing a computing power network objective function based on the electricity consumption of the online task, the electricity consumption of the offline task, and a time-of-use electricity price;

[0011] The time-of-use electricity price comprises at least a peak electricity price and a valley electricity price.

[0012] Optionally, the electricity consumption of the online task is calculated based on the number of online tasks and a task quantity electricity consumption coefficient of each time unit in a set time period.

[0013] Optionally, the number of online tasks of each time unit in the set time period is calculated based on the original number of online tasks, the number of transferred-in online tasks, and the number of transferred-out online tasks of each time unit in the set time period.

[0014] Optionally, the electricity consumption of the offline task is calculated based on the electricity consumption of each offline subtask of each time unit in a set time period and the task quantity execution state of each offline subtask.

[0015] Optionally, the solving of the computing power network objective function to obtain the optimized scheduling result of the computing power task and the determination of the computing power load demand quantity of the computing power task based on the optimized scheduling result of the computing power task comprise:

[0016] solving the computing power network objective function to obtain the optimized scheduling result of the electricity consumption of the online task and the optimized scheduling result of the electricity consumption of the offline task; the optimized scheduling result of the electricity consumption of the online task comprises the number of online tasks of each time unit in a set time period; and the optimized scheduling result of the electricity consumption of the offline task comprises the electricity consumption of each offline subtask of each time unit in a set time period.

[0017] calculating the load demand quantity of the online task based on the number of online tasks of each time unit in the set time period and a task quantity electricity consumption coefficient;

[0018] determining the load demand quantity of the offline task based on the electricity consumption of each offline subtask of each time unit in the set time period and the task quantity execution state of each offline subtask.

[0019] determining the computing power load demand quantity of the computing power task based on the load demand quantity of the online task and the load demand quantity of the offline task.

[0020] Optionally, the determination of the target influence data of the power network objective function based on the comparison result of the computing power load demand quantity and the maximum power supply quantity of the distributed power source in the power network comprises:

[0021] When the computing power load demand is greater than the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources and the power purchase cost of the main grid.

[0022] When the computing power load demand is less than or equal to the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources.

[0023] Optionally, the distributed power source includes at least one of wind turbines, photovoltaic generators, diesel generators, and gas turbines.

[0024] On the other hand, embodiments of the present invention also provide an optimized scheduling device based on resource coordination, comprising:

[0025] A construction module is used to construct a computing power network objective function based on computing power tasks; the computing power network objective function represents minimizing the operating cost of the computing power task;

[0026] The first solution module is used to solve the objective function of the computing power network, obtain the optimized scheduling result of the computing power task, and determine the computing power load requirement of the computing power task based on the optimized scheduling result of the computing power task.

[0027] The determination module is used to determine the target impact data of the power grid objective function based on the comparison between the computing power load demand and the maximum power supply of distributed power sources in the power grid; the power grid objective function represents minimizing the operating cost of the power grid.

[0028] The second solution module is used to construct a power grid objective function based on the target impact data, and to solve the power grid objective function to obtain the optimized scheduling result of the power grid, so as to achieve optimized scheduling of resources in a coordinated manner.

[0029] Optionally, the computing power task includes the power consumption of online tasks and the power consumption of offline tasks; the computing power task based on the computing power network constructs the computing power network objective function, including:

[0030] Based on the power consumption of the online tasks, the power consumption of the offline tasks, and the time-of-use electricity price, a computing power network objective function is constructed.

[0031] The time-of-use electricity price includes at least peak electricity price and off-peak electricity price.

[0032] Optionally, the power consumption of the online tasks is calculated based on the number of online tasks and the power consumption coefficient of the tasks in each time unit within a set time period.

[0033] Optionally, the number of online tasks in each time unit within the set time period is calculated based on the original number of online tasks, the number of online tasks transferred in, and the number of online tasks transferred out in each time unit within the set time period.

[0034] Optionally, the power consumption of the offline task is calculated based on the power consumption of the offline sub-tasks in each time unit within a set time period and the task execution status of each offline sub-task.

[0035] Optionally, solving the objective function of the computing power network to obtain the optimized scheduling result of the computing power task, and determining the computing power load requirement of the computing power task based on the optimized scheduling result of the computing power task, includes:

[0036] Solve the objective function of the computing power network to obtain the optimized scheduling results of the power consumption of the online tasks and the optimized scheduling results of the power consumption of the offline tasks; the optimized scheduling results of the power consumption of the online tasks include the number of online tasks in each time unit within a set time period; the optimized scheduling results of the power consumption of the offline tasks include the power consumption of the offline subtasks in each time unit within a set time period.

[0037] Based on the number of online tasks and the power consumption coefficient of each time unit within the set time period, the load demand of online tasks is calculated.

[0038] Based on the power consumption of offline subtasks in each time unit within the set time period and the task execution status of each offline subtask, the load requirement of offline tasks is determined.

[0039] The computing power load requirement of the computing power task is determined based on the load requirements of the online task and the offline task.

[0040] Optionally, determining the target impact data of the power grid objective function based on the comparison between the computing power load demand and the maximum power supply of distributed generation in the power grid includes:

[0041] When the computing power load demand is greater than the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources and the power purchase cost of the main grid.

[0042] When the computing power load demand is less than or equal to the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources.

[0043] Optionally, the distributed power source includes at least one of wind turbines, photovoltaic generators, diesel generators, and gas turbines.

[0044] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described resource-cooperative optimization scheduling method.

[0045] On the other hand, the present invention also provides a machine-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described resource-cooperative optimized scheduling method.

[0046] On the other hand, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-mentioned resource-cooperative optimization scheduling method.

[0047] Through the above technical solution, this embodiment of the invention obtains the optimized scheduling result of computing tasks corresponding to minimizing the operating cost of the computing tasks by solving the objective function of the computing power network. Then, based on the optimized scheduling result of the computing power tasks, it compares the computing load demand of the computing tasks with the maximum power supply of distributed power sources in the power grid to determine the target impact data of the power grid objective function. Based on the target impact data, it constructs and solves the power grid objective function to obtain the optimized scheduling result of the power grid. This embodiment of the invention forms a new model of resource interaction between the generation side and the load side. Through deep collaboration between computing power and electricity, it achieves optimal scheduling that minimizes the operating cost of computing tasks and the operating cost of the power grid, as well as achieving efficient resource utilization and stable system operation. Therefore, this embodiment of the invention effectively matches the high energy consumption of computing centers with the volatility of the power system, improving the overall stability and economy of the power system.

[0048] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0049] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0050] Figure 1 This is a flowchart illustrating the resource-coordinated optimization scheduling method provided by the present invention.

[0051] Figure 2 This is a schematic diagram of the distributed power source provided by the present invention;

[0052] Figure 3 This is a schematic diagram of the collaborative optimization of computing power network and power grid provided by the present invention;

[0053] Figure 4This is a schematic diagram of the structure of the resource-coordinated optimized scheduling device provided by the present invention;

[0054] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0055] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0056] The coordinated optimization of power grids and computing networks has become an important strategy for promoting socio-economic development. How to leverage the flexible adjustment capabilities of computing tasks and tap their potential for flexible electricity consumption to promote coordinated optimization between power grids and computing networks has become a focus of attention for many researchers. Traditional optimization methods for computing networks and power grids mostly operate independently, lacking a coordination mechanism, leading to resource waste and low operational efficiency. The high energy consumption of computing centers and the volatility of the power system are not effectively matched, affecting the overall system's stability and economy.

[0057] The present invention aims to provide an optimized scheduling method based on resource collaboration. Through deep collaboration between computing power and electricity, it achieves efficient resource utilization, stable system operation, and optimal scheduling, thereby effectively matching the high energy consumption of computing centers with the volatility of power systems and improving the overall stability and economy of the power system.

[0058] Method Implementation Examples

[0059] Please refer to Figure 1 This invention provides an optimized scheduling method based on resource collaboration, comprising:

[0060] Step 100: Construct the objective function of the computing power network based on the computing power network's computing power tasks.

[0061] Electronic devices construct a computing power network objective function based on computing power tasks. In one embodiment, the computing power tasks include the power consumption of online tasks and the power consumption of offline tasks. The computing power network objective function represents the operating cost of minimizing the power consumption of online and offline tasks. Based on the temporal flexibility model of offline tasks, this embodiment of the invention considers the spatial flexibility of online tasks to construct a computing power network objective function that integrates the spatiotemporal characteristics of computing power tasks. The objective of the computing power network objective function in this embodiment of the invention is to minimize the operating cost of the computing power network, integrating the spatial and temporal flexibility of online and offline tasks to change user power consumption behavior. That is, the overall optimization objective is to minimize the operating cost of the computing power network and achieve optimal scheduling.

[0062] Online tasks include real-time load monitoring and adjustment, distributed resource collaborative control, and other tasks requiring real-time processing. Taking real-time load monitoring and adjustment as an example, this involves collecting real-time load data from distribution transformers and lines via smart terminals to dynamically adjust load distribution and prevent overload or low voltage issues. Offline tasks include load forecasting and model training, network structure optimization and planning, and other tasks requiring long-term analysis. Taking load forecasting and model training as an example, this involves using historical load data to train a forecasting model to predict future load trends.

[0063] In one embodiment, step 100, constructing a computing network objective function based on computing network tasks, includes: constructing a computing network objective function based on the power consumption of the online tasks, the power consumption of the offline tasks, and the time-of-use electricity price.

[0064] The electronic device constructs a computing power network objective function based on the power consumption of the online tasks, the power consumption of the offline tasks, and the time-of-use electricity price. In one embodiment, the computing power network objective function C... s.min The calculation formula is as follows: C s.min =E s ·C fs (t)=(E lx +E zx )·C fs (t); where E s C represents the total power consumption of the computing network. fs (t) represents the time-of-use electricity price at time t, E zx E lx These represent the electricity consumption of the computing network executing online tasks and offline tasks at time t, respectively. The time-of-use (TOU) electricity price includes at least peak and off-peak electricity prices. In one embodiment, the TOU electricity price includes peak and off-peak prices. In another embodiment, the TOU electricity price includes peak, off-peak, and off-peak prices over a 24-hour period. Peak price refers to the electricity price during peak user electricity consumption periods, off-peak price refers to the electricity price during off-peak user electricity consumption periods, and off-peak price refers to the electricity price during the period between peak and off-peak user electricity consumption periods. This embodiment of the invention establishes a 24-hour TOU electricity price to guide users in selecting the most suitable computing task time. The TOU electricity price model is simulated as follows:

[0065]

[0066] Among them, C av For off-peak electricity pricing, C l For off-peak electricity pricing, C h For peak electricity pricing, λ1 and λ2 are electricity price coefficients, and C is the peak electricity price. l <C av <C h .

[0067] In one embodiment, the power consumption of online tasks is calculated based on the number of online tasks and the power consumption coefficient of the task volume in each time unit within a set time period. The electronic device calculates the power consumption of online tasks based on the number of online tasks and the power consumption coefficient of the task volume in each time unit within the set time period. Further, the number of online tasks in each time unit within the set time period is calculated based on the original number of online tasks, the number of online tasks transferred in, and the number of online tasks transferred out in each time unit within the set time period. Therefore, the power consumption E of the online task is calculated. zx The value is calculated using the following formula:

[0068]

[0069] Among them, E zx.(t) Y represents the power consumption of the computing network executing online tasks at time t. zx(t) y represents the number of online tasks processed by the computing network, where I is the power consumption coefficient for the task load. ini (t) represents the original number of online tasks submitted to the local machine for each time unit within the specified time period, i.e., the original number of online tasks for each time unit within the specified time period. in (t) represents the number of online tasks transferred in each time unit within a specified time period, y to (t) represents the number of online tasks transferred out in each time unit within a set time period. T is the time period, for example, T could be 24 hours. t is each time unit, which can be a second, so each time unit can be every second. In other embodiments, t can also be every minute or every hour.

[0070] This embodiment of the invention also requires setting online task volume constraints. For example, the online task volume constraints can be set as follows: the number of online tasks transferred out in each time unit within a set time period should not exceed the original number of online tasks originally allocated to local processing: y to (t)≤y ini (t); Considering limited processing capacity, the number of online tasks processed by the computing power network shall not exceed the set maximum value y. max Ensure the server does not become overloaded: Y zx(t) ≤y max .

[0071] Furthermore, the power consumption of offline tasks is calculated based on the power consumption of offline sub-tasks in each time unit within a set time period and the task execution status of each offline sub-task. For example, the power consumption E of an offline task... lx The formula is as follows:

[0072] Among them, E lx.(t)O represents the power consumption of the computing network executing offline tasks at time t. lx(t) Let t represent the power consumption of the offline subtask at time t. K represents the task execution status, a 0 / 1 variable, where 0 indicates the offline subtask is not executed during this period, and 1 indicates the offline subtask is executed. T is the time period, for example, T could be 24 hours. t represents each time unit, which can be a second, so each time unit can be a second.

[0073] In some embodiments, computing network constraints can also be set. For example, the computing network constraints can be set as follows: 1. Total power consumption of the computing network. Where P(t) is the power consumption of the computing network at time t; 2. Power constraint conditions Where P lx (t) represents the power of the offline task, P zx (t) represents the power of the online task, P 损耗 (t) represents the power loss of the computing network.

[0074] Step 200: Solve the objective function of the computing power network to obtain the optimized scheduling result of the computing power task, and determine the computing power load requirement of the computing power task based on the optimized scheduling result of the computing power task.

[0075] Electronic devices can use methods such as linear programming, genetic algorithms, and simulated degradation algorithms to solve the objective function of the computing power network and obtain the optimized scheduling result of the computing power tasks. The optimized scheduling result of the computing power tasks includes the optimized scheduling result of the power consumption of online tasks and the optimized scheduling result of the power consumption of offline tasks. In one embodiment, the optimized scheduling result of the power consumption of online tasks includes the number of online tasks in each time unit within a set time period, for example, the number of online tasks per second in 24 hours. The optimized scheduling result of the power consumption of offline tasks includes the power consumption of offline sub-tasks in each time unit within a set time period, for example, the power consumption of offline sub-tasks per second in 24 hours. Therefore, step 200, solving the objective function of the computing power network to obtain the optimized scheduling result of the computing power tasks, and determining the computing power load requirement of the computing power tasks based on the optimized scheduling result of the computing power tasks, includes:

[0076] Step 210: Solve the objective function of the computing network to obtain the optimized scheduling results of the power consumption of the online tasks and the optimized scheduling results of the power consumption of the offline tasks.

[0077] Electronic devices can use methods such as linear programming, genetic algorithms, and simulated degradation algorithms to solve the objective function of the computing power network, and obtain the number of online tasks and the power consumption of offline subtasks in each time unit within a set time period. For example, the number of online tasks per second and the power consumption of offline subtasks per second within 24 hours.

[0078] Step 220: Calculate the load demand of online tasks based on the number of online tasks and the power consumption coefficient of each time unit within the set time period.

[0079] The electronic device can calculate the load demand of online tasks by substituting the number of online tasks and the power consumption coefficient of each time unit within the set time period into formula (2).

[0080] Step 230: Determine the load requirement of the offline task based on the power consumption of the offline sub-task in each time unit within the set time period and the task execution status of each offline sub-task.

[0081] The electronic device can calculate the load requirement of the offline task by substituting the power consumption of the offline subtask in each time unit within the set time period and the task execution status of each offline subtask into formula (3).

[0082] Step 240: Determine the computing power load requirement of the computing power task based on the load requirements of the online task and the load requirements of the offline task.

[0083] In one embodiment, the electronic device can determine the computing power load requirement of the computing power task based on the sum of the load requirements of online tasks and offline tasks. In other embodiments, the electronic device can determine the computing power load requirement of the computing power task based on the weighted sum of the load requirements of online tasks and offline tasks. In this embodiment of the invention, the power supply of the power grid is determined by the computing power load requirement. Therefore, based on the calculated computing power load requirement, it is convenient to determine the power supply of the power grid and the type of power source.

[0084] This invention utilizes the data center's computing power network as a crucial schedulable load. By constructing and solving the objective function of the computing power network, it obtains optimized scheduling results for the power consumption of online tasks and offline tasks, minimizing the operating cost of computing tasks. This invention can follow time-of-use pricing, selecting on-site online computation during off-peak hours and offline task computation during peak hours. The power supply capacity of the power grid is determined by the computing power load demand. As the computing power network load curve changes with user behavior, the power supply requirements of the power grid also change accordingly, effectively "shaving peaks and filling valleys" to reduce the power supply pressure on the distribution network during peak hours.

[0085] Step 300: Based on the comparison between the computing power load demand and the maximum power supply of distributed power sources in the power grid, determine the target impact data of the power grid objective function.

[0086] This invention determines the power supply capacity of the power grid by measuring the computing load demand of the computing power network. In one embodiment, determining the target impact data of the power grid objective function based on a comparison between the computing load demand and the maximum power supply capacity of distributed generation sources in the power grid includes: when the computing load demand is greater than the maximum power supply capacity, determining the target impact data of the power grid objective function as the generation cost of distributed generation sources and the main grid's power purchase cost; when the computing load demand is less than or equal to the maximum power supply capacity, determining the target impact data of the power grid objective function as the generation cost of distributed generation sources.

[0087] When the computing power load demand exceeds the maximum supply capacity of distributed generation in the power grid, electricity will be purchased from the main grid to meet the demand side's electricity consumption. In this case, the objective impact data of the power grid's objective function is the generation cost of distributed generation and the main grid's electricity purchase cost. When the computing power load demand is less than or equal to the maximum supply capacity of distributed generation in the power grid, electricity will not be purchased from the main grid to meet the demand side's electricity consumption. In this case, the objective impact data of the power grid's objective function is the generation cost of distributed generation. The distributed generation includes at least one of wind turbines, photovoltaic generators, diesel generators, and gas turbines. To consider as many energy types as possible for power supply, please refer to... Figure 2 The distributed power sources used in the energy management system of the source-side distribution network include wind turbines, photovoltaic generators, diesel generators (or diesel generator sets), and gas turbines (or gas turbine sets), which represent new energy sources. Therefore, the power generation cost of distributed power sources includes the power generation costs of wind turbines, photovoltaic generators, diesel generators, and gas turbines.

[0088] The objective function C of the power grid is constructed based on the generation cost of distributed power sources. d.min The calculation formula is as follows:

[0089] C d.min =C WT +C PV +C DG +C GT +C L ;Formula (4)

[0090] Among them, C WT C represents the cost of generating electricity from a wind turbine. PV C represents the cost of electricity generated by a photovoltaic generator. DG C represents the cost of generating electricity using a diesel generator. GT C represents the cost of gas turbine power generation. L This indicates the cost of power grid losses.

[0091] The objective function C of the power grid is constructed based on the generation cost of distributed power sources and the power purchase cost of the main grid. d.min The calculation formula is as follows:

[0092] C d.min =C WT +C PV +C DG +C GT +C G +C L ;Formula (5)

[0093] Among them, C WT C represents the cost of generating electricity from a wind turbine. PV C represents the cost of electricity generated by a photovoltaic generator. DG C represents the cost of generating electricity using a diesel generator. GT C represents the cost of gas turbine power generation. G C represents the cost of purchasing electricity from the main grid. L This represents the grid loss cost. The embodiments of this invention are illustrated using formula (5). The power grid objective function represents minimizing the operating cost of the power grid. That is, the power grid optimization objective is to minimize the economic cost of generating power while meeting the demand for computing power load, thereby forming a new model of resource interaction between the new energy generation side and the load side.

[0094] Wind turbines are used on the source side of the power distribution network. When their own power supply is insufficient, they can be combined with other new energy sources to supplement the power supply demand. The wind turbine power generation cost C in formula (5) is... WT Including wind turbine maintenance costs C WT1 Management expenses C WT2 The cost of generating electricity from a wind turbine (C) WT Calculated using the following formula:

[0095]

[0096] Where k WT1 k is the operation and maintenance factor of the unit to which the wind turbine belongs. WT2 P represents the management cost coefficient for the unit owning the wind turbine generator. WT (t) represents the output power of the wind turbine at time t, and T is the period.

[0097] The photovoltaic generator power generation cost C in formula (5) PV Including photovoltaic generator maintenance costs C PV1 Management expenses C PV2 The cost of electricity generated by a photovoltaic generator (C) PV Calculated using the following formula:

[0098]

[0099] Where kPV1 k is the operation and maintenance factor of the unit to which the photovoltaic generator belongs. PV2 P represents the management cost coefficient for the unit owning the photovoltaic generator. PV (t) represents the output power of the photovoltaic generator at time t, and T is the period.

[0100] A diesel generator combines a diesel engine and an electric generator, using diesel fuel to drive the generator and produce electricity. During operation, the diesel generator incurs fuel costs and maintenance costs. The diesel generator's power generation cost C in formula (5) is... DG Including fuel costs C DG1 Maintenance fee C DG2 And the cost of treating pollutants C DG3 The cost of generating electricity using a diesel generator (C) DG Calculated using the following formula:

[0101]

[0102] Where a, b, and c are the fuel consumption cost coefficients of the diesel generator, and k DG2 Diesel generator maintenance factor, k DG3 P is the cost coefficient for pollutant treatment. DG (t) represents the actual power of the diesel generator at time t, and T is the period.

[0103] When the power generation of the gas turbine cannot meet the user's demand, the power grid will provide supplementary power to ensure that the unit can operate normally and reach full load conditions. The gas turbine incurs fuel costs and maintenance costs during operation. The gas turbine power generation cost C in formula (5) is... GT Including fuel costs C GT1 and maintenance fee C GT2 Gas turbine power generation cost C GT Calculated using the following formula:

[0104]

[0105] Where l, m, and n are the gas turbine fuel consumption cost coefficients, and k GT2 Gas turbine maintenance factor, P GT (t) represents the actual power of the gas turbine at time t, and T is the period.

[0106] The main grid electricity purchase cost C in formula (5) G Calculated using the following formula:

[0107]

[0108] Where C kG (t) represents the main grid electricity purchase price at time t, P G(t) represents the main grid operating power at time t, where T is the period.

[0109] The grid loss cost C in formula (5) L Calculated using the following formula:

[0110]

[0111] Where C kL (t) is the grid loss coefficient at time t, P L (t) represents the power loss (dissipation) of the power grid at time t, where T is the period.

[0112] This embodiment of the invention also requires setting power grid constraints. Power grid constraints include power balance constraints and generator unit operating power constraints. The formula for the power balance constraint is as follows:

[0113]

[0114] Where P(t) is the power consumption of the computing network at time t, and P1(t) is the power consumption of other electrical loads at time t. WT (t) represents the output power of the wind turbine at time t, P PV (t) represents the output power of the photovoltaic generator at time t, P DG (t) represents the output power of the diesel generator at time t, P GT (t) represents the gas turbine output power at time t, P L (t) represents the power loss at time t.

[0115] The power constraints for generator sets include power constraints for wind turbines, photovoltaic generators, diesel generators, and gas turbines. The formula for the power constraint of wind turbines is as follows: P WT (t)≤P WT.i (t). Where P WT.i (t) represents the predicted power output of the wind turbine.

[0116] The formula for the operating power constraint of a photovoltaic generator is as follows: P PV (t)≤P PV.i (t). Where P PV.i (t) represents the predicted output power of the photovoltaic generator.

[0117] The formula for the operating power constraint of a diesel generator is as follows: P DG (t) min ≤P DG (t)≤P DG (t) max , where P DG (t) min PDG (t) max These are the upper and lower limits of the output power of the diesel generator.

[0118] Gas turbine operating power constraint: P GT (t) min ≤P GT (t)≤P GT (t) max , where P GT (t) min P GT (t) max These are the upper and lower limits of the gas turbine's output power, respectively.

[0119] Step 400: Construct a power grid objective function based on the target impact data, and solve the power grid objective function to obtain the optimized scheduling result of the power grid, so as to achieve optimized scheduling of resources in a coordinated manner.

[0120] In the objective function of the power grid, the power grid needs to calculate the operating cost of the power grid based on the power generation and consumption information, fuel cost, line transmission and other information under the optimal operating state to reflect the effect of optimal resource allocation within the power grid. In the power grid, when new energy sources can generate power, the main focus should be on supplying economical and environmentally friendly electricity. However, it should be considered that the distributed energy sources on the demand side may not be able to fully meet the demand side's electricity consumption, so the distribution network should be fully utilized to purchase electricity from the main grid. The optimization objective of the power grid objective function is to minimize the power generation costs of wind turbine generators, photovoltaic generators, diesel generators, gas turbine generators and main grid purchase costs. Electronic equipment can refer to formula (5) to construct the power grid objective function based on the power generation costs of wind turbine generators, photovoltaic generators, diesel generators, gas turbine generators, main grid purchase costs and grid loss costs.

[0121] In this embodiment of the invention, the power supply capacity of the power grid is determined by the computing power load demand of the computing power network. An objective function for the power grid, aimed at minimizing economic costs, is constructed based on the interaction power of wind and solar generators, diesel and gas turbine units, and the distribution network. Under conditions of meeting power supply demand and stability, the generation costs of each power generation component in the power grid will also change according to real-time electricity prices.

[0122] Please refer to Figure 3This invention focuses on the collaborative optimization of computing power networks and power grids, establishing a two-layer optimization model (two-layer objective function). The lower-layer model consists of online and offline tasks of the computing power network, while the upper-layer model's power supply energy comprises wind power, photovoltaic power, diesel generators, gas turbines, and power purchased from the main grid. Energy interaction and allocation are handled by a smart distribution network, thereby optimizing the scheduling model. When the power grid's demand exceeds the maximum power provided by the distributed power sources, power will be purchased from the main grid to meet the demand side's electricity consumption. Therefore, the power grid optimization objective is to minimize the economic cost of generating enough electricity to meet the computing power network's demand, thus forming a new model of resource interaction between the new energy generation side and the load side. The computing power network objective function minimizes the computing power network's operating cost, integrating the spatial and temporal flexibility of online and offline tasks to change user electricity consumption behavior. The overall optimization objective is to minimize network operating costs and achieve optimal scheduling. In this embodiment of the invention, by constructing a dual objective function for bidirectional collaboration between the power grid and the computing power network, the collaboration between the two can not only improve energy utilization efficiency, but also promote the rapid development of information technology, achieve efficient allocation of energy and information resources, reduce operating costs, and improve the reliability and flexibility of the power system.

[0123] This invention provides an optimized scheduling result for computing tasks that minimizes their operating costs by solving the objective function of the computing power network. Based on this optimized scheduling result, the invention compares the computing load demand of the computing tasks with the maximum power supply of distributed generation sources in the power grid to determine the target impact data of the power grid objective function. This data is then used to construct and solve the power grid objective function, resulting in the optimized scheduling result of the power grid. This invention establishes a new model for resource interaction between the generation and load sides. Through deep collaboration between computing power and electricity, it achieves optimal scheduling that minimizes both the operating costs of computing tasks and the operating costs of the power grid, as well as efficient resource utilization and stable system operation. Therefore, this invention effectively matches the high energy consumption of computing centers with the volatility of the power system, improving the overall stability and economy of the power system.

[0124] Device Examples

[0125] Please refer to Figure 4 On the other hand, embodiments of the present invention also provide an optimized scheduling device based on resource coordination, comprising:

[0126] The construction module 401 is used to construct the computing power network objective function based on the computing power tasks of the computing power network; the computing power network objective function represents minimizing the operating cost of the computing power tasks;

[0127] The first solving module 402 is used to solve the objective function of the computing power network, obtain the optimized scheduling result of the computing power task, and determine the computing power load requirement of the computing power task based on the optimized scheduling result of the computing power task.

[0128] The determination module 403 is used to determine the target impact data of the power grid objective function based on the comparison results between the computing power load demand and the maximum power supply of distributed power sources in the power grid; the power grid objective function represents minimizing the operating cost of the power grid.

[0129] The second solution module 404 is used to construct a power grid objective function based on the target impact data, and to solve the power grid objective function to obtain the optimized scheduling result of the power grid, so as to achieve optimized scheduling of resources in a coordinated manner.

[0130] By solving the objective function of the computing power network, the optimal scheduling result of the computing power task corresponding to minimizing the operating cost of the computing power task is obtained. Then, based on the optimized scheduling result of the computing power task, the computing power load demand of the computing power task is calculated and compared with the maximum power supply of distributed generation in the power grid. The target impact data of the power grid objective function is determined. Based on the target impact data, the power grid objective function is constructed and solved to obtain the optimized scheduling result of the power grid. This embodiment of the invention forms a new mode of resource interaction between the generation side and the load side. Through deep collaboration between computing power and electricity, it achieves optimal scheduling that minimizes the operating cost of computing power tasks and the operating cost of the power grid, as well as achieving efficient resource utilization and stable system operation. Thus, this embodiment of the invention achieves an effective match between the high energy consumption of the computing power center and the volatility of the power system, improving the overall stability and economy of the power system.

[0131] Optionally, the computing power task includes the power consumption of online tasks and the power consumption of offline tasks; the computing power task based on the computing power network constructs the computing power network objective function, including:

[0132] Based on the power consumption of the online tasks, the power consumption of the offline tasks, and the time-of-use electricity price, a computing power network objective function is constructed.

[0133] The time-of-use electricity price includes at least peak electricity price and off-peak electricity price.

[0134] Optionally, the power consumption of the online tasks is calculated based on the number of online tasks and the power consumption coefficient of the tasks in each time unit within a set time period.

[0135] Optionally, the number of online tasks in each time unit within the set time period is calculated based on the original number of online tasks, the number of online tasks transferred in, and the number of online tasks transferred out in each time unit within the set time period.

[0136] Optionally, the power consumption of the offline task is calculated based on the power consumption of the offline sub-tasks in each time unit within a set time period and the task execution status of each offline sub-task.

[0137] Optionally, solving the objective function of the computing power network to obtain the optimized scheduling result of the computing power task, and determining the computing power load requirement of the computing power task based on the optimized scheduling result of the computing power task, includes:

[0138] Solve the objective function of the computing power network to obtain the optimized scheduling results of the power consumption of the online tasks and the optimized scheduling results of the power consumption of the offline tasks; the optimized scheduling results of the power consumption of the online tasks include the number of online tasks in each time unit within a set time period; the optimized scheduling results of the power consumption of the offline tasks include the power consumption of the offline subtasks in each time unit within a set time period.

[0139] Based on the number of online tasks and the power consumption coefficient of each time unit within the set time period, the load demand of online tasks is calculated.

[0140] Based on the power consumption of offline subtasks in each time unit within the set time period and the task execution status of each offline subtask, the load requirement of offline tasks is determined.

[0141] The computing power load requirement of the computing power task is determined based on the load requirements of the online task and the offline task.

[0142] Optionally, determining the target impact data of the power grid objective function based on the comparison between the computing power load demand and the maximum power supply of distributed generation in the power grid includes:

[0143] When the computing power load demand is greater than the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources and the power purchase cost of the main grid.

[0144] When the computing power load demand is less than or equal to the maximum power supply, the target impact data of the power grid objective function is determined to be the generation cost of distributed power sources.

[0145] Optionally, the distributed power source includes at least one of wind turbines, photovoltaic generators, diesel generators, and gas turbines.

[0146] The resource-coordinated optimization scheduling device includes a processor and a memory. The aforementioned construction module 401, first solution module 402, determination module 403, second solution module 404, etc., are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0147] A processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured.

[0148] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0149] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call logical instructions in the memory 530 to execute a resource-coordinated optimization scheduling method. This method includes: constructing a computing network objective function based on computing tasks within the computing network; the computing network objective function characterizes minimizing the operating cost of the computing tasks; solving the computing network objective function to obtain the optimized scheduling result of the computing tasks, and determining the computing load demand of the computing tasks based on the optimized scheduling result; determining the target impact data of the power grid objective function based on a comparison between the computing load demand and the maximum power supply of distributed power sources in the power grid; the power grid objective function characterizes minimizing the operating cost of the power grid; constructing the power grid objective function based on the target impact data, and solving the power grid objective function to obtain the optimized scheduling result of the power grid, thereby achieving resource-coordinated optimization scheduling.

[0150] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0151] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a machine-readable storage medium. When the computer program is executed by a processor, the computer can execute a resource-coordinated optimization scheduling method. This method includes: constructing a computing network objective function based on computing tasks of the computing network; the computing network objective function characterizes minimizing the operating cost of the computing tasks; solving the computing network objective function to obtain the optimized scheduling result of the computing tasks, and determining the computing load demand of the computing tasks based on the optimized scheduling result; determining the target impact data of the power network objective function based on a comparison between the computing load demand and the maximum power supply of distributed power sources in the power grid; the power network objective function characterizes minimizing the operating cost of the power grid; constructing the power network objective function based on the target impact data, and solving the power network objective function to obtain the optimized scheduling result of the power grid, thereby achieving resource-coordinated optimization scheduling.

[0152] In another aspect, the present invention also provides a machine-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform a resource-coordinated optimization scheduling method. This method includes: constructing a computing network objective function based on computing tasks of a computing network; the computing network objective function characterizing minimizing the operating cost of the computing tasks; solving the computing network objective function to obtain an optimized scheduling result for the computing tasks; and determining the computing load demand of the computing tasks based on the optimized scheduling result; determining target impact data for the power network objective function based on a comparison between the computing load demand and the maximum power supply of distributed power sources in the power grid; the power network objective function characterizing minimizing the operating cost of the power grid; constructing the power network objective function based on the target impact data; and solving the power network objective function to obtain an optimized scheduling result for the power grid, thereby achieving resource-coordinated optimization scheduling.

[0153] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0154] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing scheduling based on resource coordination, characterized in that, The method comprises the following steps: constructing a computing power network target function based on a computing power network; the computing power network target function represents the minimization of the operation cost of the computing power task; solving the computing power network target function to obtain the optimized scheduling result of the computing power task, and determining the computing power load demand of the computing power task based on the optimized scheduling result of the computing power task; determining the target influence data of the power grid target function based on the comparison result of the computing power load demand and the maximum power supply of the distributed power source in the power grid; the power grid target function represents the minimization of the operation cost of the power grid; constructing the power grid target function based on the target influence data, and solving the power grid target function to obtain the optimized scheduling result of the power grid, so as to realize the optimized scheduling of resource coordination; the computing power task includes the power consumption of online tasks and the power consumption of offline tasks; The computing power task construction computing power network target function based on the computing power network includes: constructing the computing power network target function based on the electricity consumption of the online task, the electricity consumption of the offline task and the time-of-use electricity price; wherein the time-of-use electricity price at least includes peak electricity price and valley electricity price; the computing power network target function C s.min The calculation formula is as follows: ; wherein is the total electricity consumption of the computing power network, is the time-of-use electricity price at time t, , respectively, the electricity consumption of the computing power network executing online task at time t and the electricity consumption of the offline task. determining the target influence data of the power grid target function based on the comparison result of the computing power load demand and the maximum power supply of the distributed power source in the power grid, comprising: in the case that the computing power load demand is greater than the maximum power supply, determining the target influence data of the power grid target function as the power generation cost of the distributed power source and the main grid power purchase cost; in the case that the computing power load demand is less than or equal to the maximum power supply, determining the target influence data of the power grid target function as the power generation cost of the distributed power source; A power grid objective function C constructed based on the generation cost of distributed power sources d.min The calculation formula is as follows: ; wherein, C WT represents the generation cost of a wind power generator, C PV represents the generation cost of a photovoltaic generator, C DG represents the generation cost of a diesel generator, C GT represents the generation cost of a gas turbine, C L represents the grid loss cost; The power grid target function C constructed based on the power generation cost of the distributed power supply and the power purchase cost of the main grid d.min The calculation formula is as follows: ; wherein, C WT The power generation cost of the wind turbine is represented by C PV The power generation cost of the photovoltaic generator is represented by C DG The power generation cost of the diesel generator is represented by C GT The power generation cost of the gas turbine is represented by C G The power purchase cost of the main grid is represented by C L The power grid loss cost is represented by 2. The method of claim 1, wherein, the power consumption of the online task is calculated based on the number of online tasks and the task amount power consumption coefficient of each time unit in a set time period. 3.The resource coordination based optimal scheduling method of claim 2, wherein, the number of online tasks of each time unit in the set time period is calculated based on the original number of online tasks, the number of transferred online tasks and the number of transferred offline tasks of each time unit in the set time period.

4. The method of claim 1, wherein, the power consumption of the offline task is calculated based on the power consumption of each offline subtask and the task amount execution state of each offline subtask in each time unit in a set time period.

5. The resource coordination-based optimized scheduling method according to claim 1, wherein: solving the computing power network target function to obtain the optimized scheduling result of the computing power task, and determining the computing power load demand of the computing power task based on the optimized scheduling result of the computing power task, comprising: solving the computing power network target function to obtain the optimized scheduling result of the power consumption of the online task and the optimized scheduling result of the power consumption of the offline task; the optimized scheduling result of the power consumption of the online task includes the number of online tasks of each time unit in a set time period; the optimized scheduling result of the power consumption of the offline task includes the power consumption of each offline subtask of each time unit in a set time period; calculating the load demand of the online task based on the number of online tasks of each time unit in the set time period and the task amount power consumption coefficient; determining the load demand of the offline task based on the power consumption of each offline subtask of each time unit in the set time period and the task amount execution state of each offline subtask. Determine a computing power load demand of the computing power task based on the load demand of the online task and the load demand of the offline task.

6. The method of claim 1-5, wherein, The distributed power source includes at least one of a wind power generator, a photovoltaic power generator, a diesel generator, and a gas turbine.

7. A resource coordination based optimal scheduling apparatus, characterized in that, Comprise: A construction module configured to construct a computing power network objective function based on a computing power task of the computing power network; The computing power network objective function represents minimization of operation cost of the computing power task; A first solving module configured to solve the computing power network objective function to obtain an optimized scheduling result of the computing power task, and determine a computing power load demand of the computing power task based on the optimized scheduling result of the computing power task; A determination module configured to determine target influence data of a power network objective function based on a comparison result of the computing power load demand and a maximum power supply of the distributed power source in the power network; The power network objective function represents minimization of operation cost of the power network; A second solving module configured to construct the power network objective function based on the target influence data, and solve the power network objective function to obtain an optimized scheduling result of the power network, so as to realize optimized scheduling of resource coordination; The computing power task includes power consumption of an online task and power consumption of an offline task; The computing power task construction computing power network target function based on the computing power network includes: constructing the computing power network target function based on the electricity consumption of the online task, the electricity consumption of the offline task, and the time-of-use electricity price; wherein the time-of-use electricity price at least includes peak electricity price and valley electricity price; the computing power network target function C s.min The calculation formula is as follows: ; wherein is the total electricity consumption of the computing power network, is the time-of-use electricity price at time t, , respectively, the electricity consumption of the computing power network for executing online tasks at time t and the electricity consumption of the offline task. The determination of the target influence data of the power network objective function based on the comparison result of the computing power load demand and the maximum power supply of the distributed power source in the power network comprises: In a case where the computing power load demand is greater than the maximum power supply, the target influence data of the power network objective function is determined as generation cost of the distributed power source and main grid power purchase cost; In a case where the computing power load demand is less than or equal to the maximum power supply, the target influence data of the power network objective function is determined as generation cost of the distributed power source; A power grid objective function C constructed based on the generation cost of distributed power sources d.min The calculation formula is as follows: ; wherein, C WT represents the generation cost of a wind power generator, C PV represents the generation cost of a photovoltaic generator, C DG represents the generation cost of a diesel generator, C GT represents the generation cost of a gas turbine, C L represents the grid loss cost; A power grid target function C constructed based on the power generation cost of the distributed power supply and the power purchase cost of the main grid d.min The calculation formula is as follows: ; wherein, C WT represents the power generation cost of the wind turbine, C PV represents the power generation cost of the photovoltaic generator, C DG represents the power generation cost of the diesel generator, C GT represents the power generation cost of the gas turbine, C G represents the power purchase cost of the main grid, C L represents the power grid loss cost. 8.The resource coordination based optimal scheduling apparatus of claim 7, wherein, The power consumption of the online task is calculated based on online task quantity and task quantity power consumption coefficient of each time unit in a set time period. 9.The resource coordination based optimal scheduling apparatus of claim 8, wherein, The online task quantity of each time unit in the set time period is calculated based on original online task quantity, transferred-in online task quantity, and transferred-out online task quantity of each time unit in the set time period. 10.The resource coordination based optimal scheduling apparatus according to claim 7, wherein, The power consumption of each offline subtask in each time unit in the set time period is calculated based on power consumption of each offline subtask and task quantity execution state of each offline subtask.

11. The resource coordination based optimized scheduling apparatus according to claim 7, wherein The solving of the computing power network objective function to obtain the optimized scheduling result of the computing power task, and the determination of the computing power load demand of the computing power task based on the optimized scheduling result of the computing power task comprise: The solving of the computing power network objective function to obtain the optimized scheduling result of the power consumption of the online task and the optimized scheduling result of the power consumption of the offline task; the optimized scheduling result of the power consumption of the online task includes online task quantity of each time unit in a set time period; and the optimized scheduling result of the power consumption of the offline task includes power consumption of each offline subtask in each time unit in the set time period. calculate a load demand amount of the online task based on the number of online tasks and the task amount power consumption coefficient of each time unit in the set time period; determine a load demand amount of the offline task based on the power consumption of each offline subtask and the task amount execution state of each offline subtask in each time unit in the set time period; determine the computing power load demand amount of the computing power task based on the load demand amount of the online task and the load demand amount of the offline task. 12.The resource coordination based optimal scheduling apparatus according to any one of claims 7 to 11, wherein, The distributed power supply includes at least one of a wind turbine, a photovoltaic generator, a diesel generator, and a gas turbine.

13. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the resource coordination-based optimal scheduling method in any one of claims 1 to 6.

14. A machine-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the resource coordination-based optimal scheduling method in any one of claims 1 to 6.

15. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the resource coordination-based optimal scheduling method in any one of claims 1 to 6. The computer program is executed by the processor to implement the resource coordination-based optimal scheduling method in any one of claims 1 to 6.

Images