Calculation power center scheduling method and device considering power-calculation power combination

Abstract

The invention relates to the technical field of computing collaboration, and particularly provides a computing power center scheduling method and device considering power-computing power combination, and the method comprises the steps: obtaining the information entropy of a computing power center and the energy entropy of a power system; substituting the information entropy of the computing power center and the energy entropy of the power system into a pre-constructed day-ahead power-computing collaborative optimization model and solving to obtain an optimization result; and performing day-ahead optimization scheduling on the computing power center based on the optimization result. According to the technical scheme provided by the invention, the schedulability of the calculation power task is quantified by defining the information entropy, the space-time supply and demand difference of the power system is represented by the energy entropy, the scheduling optimization model is constructed, the electric calculation cooperative scheduling is realized, the power-calculation power unified cooperative scheduling is realized, the energy consumption is reduced, and the calculation power utilization rate is improved.

Description

Technical Field

[0001] This invention relates to the field of computer collaborative technology, and specifically to a computing center scheduling method and apparatus that considers the combination of power and computing power. Background Technology

[0002] With the rapid development of artificial intelligence, cloud computing, and the Internet of Things (IoT), computing centers have become a core infrastructure supporting the digital economy. The high energy consumption problem brought about by the expansion of computing centers is becoming increasingly serious, with their electricity consumption accounting for a continuously rising proportion of global electricity consumption. At the same time, the increasing proportion of renewable energy and the intensified uncertainty on both the source and load sides of the new power system pose challenges to solving power system scheduling problems, including dimensionality explosion and difficulty in finding solutions. Power-computing power coordinated scheduling is a powerful method to achieve efficient energy use in data centers and the absorption of renewable energy. Existing scheduling methods mostly focus on local optimization, such as reducing the energy consumption of individual devices through Dynamic Voltage and Frequency Scaling (DVFS) or using real-time electricity prices to guide task load migration. Existing methods cannot fully coordinate the flexibility of power and computing power, and have the following shortcomings:

[0003] Existing technologies do not fully consider the spatiotemporal distribution characteristics of electricity supply and demand (such as regional electricity price differences, renewable energy volatility, and power grid capacity), resulting in low energy utilization.

[0004] Many computational tasks (such as offline training and batch data processing) have time-delay characteristics, but existing scheduling models lack quantitative representation of their flexibility and cannot achieve global optimization.

[0005] Power system regulation strategies under high-proportion renewable energy are not deeply coupled with computing power task scheduling, making it difficult to mitigate fluctuations in power supply and demand.

[0006] In summary, the fundamental problem is the lack of quantification of the coupling relationship between power and computing power, namely, the lack of a unified quantification standard for the energy distribution of the power system and the complexity of computing tasks, which makes it impossible to establish a joint optimization goal.

[0007] Therefore, there is an urgent need for a scheduling method that can jointly characterize the dynamic demands of computing power tasks and the spatiotemporal characteristics of the power system, so as to fully tap the synergistic potential between computing power center tasks and power system scheduling. Summary of the Invention

[0008] To overcome the above-mentioned shortcomings, this invention proposes a computing center scheduling method and device that considers the combination of power and computing power.

[0009] Firstly, a computing center scheduling method considering the combination of power and computing power is provided, the computing center scheduling method considering the combination of power and computing power includes:

[0010] Obtain the information entropy of the computing center and the energy entropy of the power system;

[0011] The information entropy of the computing center and the energy entropy of the power system are substituted into the pre-constructed day-ahead power-computing collaborative optimization model and solved to obtain the optimization results;

[0012] Based on the optimization results, the computing center is scheduled for day-ahead optimization.

[0013] The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

[0014] Preferably, the information entropy of the computing center is as follows:

[0015]

[0016] In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

[0017] Furthermore, the energy entropy of the power system of the computing center is as follows:

[0018]

[0019] In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ P P respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load(t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

[0020] Furthermore, the power generation of the computing center's energy system at time t is as follows:

[0021]

[0022] The load power of the computing center at time t is as follows:

[0023]

[0024] The charging and discharging power of the computing center energy storage system at time t is as follows:

[0025]

[0026] In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

[0027] Furthermore, the IT equipment energy consumption in region j at time t is as follows:

[0028]

[0029] In the above formula, M j Let J be the number of servers in region j. The idle load of the units in region j For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,tLet λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

[0030] Furthermore, the pre-constructed day-ahead computer-computer co-optimization model includes: the day-ahead computer-computer co-optimization objective function and its corresponding constraints.

[0031] Furthermore, the objective function for the current computer-computer collaborative optimization is as follows:

[0032]

[0033] In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

[0034] Furthermore, the constraints include: power balance constraints, power system generation capacity constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

[0035] Furthermore, the power balance constraints are as follows:

[0036] P IT (t)+P cool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t)

[0037] The power generation constraints of the power system are as follows:

[0038]

[0039] The power generation capacity ramp-up rate constraint is as follows:

[0040]

[0041] The constraints of the energy storage system are as follows:

[0042]

[0043] The temperature constraints for the computer room are as follows:

[0044]

[0045] The resource allocation constraints are as follows:

[0046]

[0047] The task deadline constraints are as follows:

[0048]

[0049] In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant. gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η charge For the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

[0050] Furthermore, the indoor temperature of region j at time t+1 is as follows:

[0051]

[0052] In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows:

[0053]

[0054] In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

[0055] Furthermore, after performing day-ahead optimization scheduling on the computing center based on the optimization results, the process includes:

[0056] Curve fitting is performed on the optimization results at each time point in the next 24 hours to obtain the optimization result curve;

[0057] Obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve;

[0058] Substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points;

[0059] Based on the intraday optimization results corresponding to the future 96 points, the computing center is scheduled for intraday optimization.

[0060] Furthermore, the pre-constructed intraday computer-computer collaborative optimization model is as follows:

[0061]

[0062] In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info (τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

[0063] Secondly, a computing center scheduling device that considers the combination of power and computing power is provided, the computing center scheduling device considering the combination of power and computing power includes:

[0064] The first acquisition module is used to acquire the information entropy of the computing center and the energy entropy of the power system.

[0065] The first analysis module is used to substitute the information entropy of the computing center and the energy entropy of the power system into the pre-constructed day-ahead power-computing collaborative optimization model and solve it to obtain the optimization results;

[0066] The first scheduling module is used to perform day-ahead optimization scheduling of the computing center based on the optimization results;

[0067] The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

[0068] Preferably, the information entropy of the computing center is as follows:

[0069]

[0070] In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

[0071] Furthermore, the energy entropy of the power system of the computing center is as follows:

[0072]

[0073] In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ PP respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load (t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

[0074] Furthermore, the power generation of the computing center's energy system at time t is as follows:

[0075]

[0076] The load power of the computing center at time t is as follows:

[0077]

[0078] The charging and discharging power of the computing center energy storage system at time t is as follows:

[0079]

[0080] In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

[0081] Furthermore, the IT equipment energy consumption in region j at time t is as follows:

[0082]

[0083] In the above formula, M j Let J be the number of servers in region j. The idle load of the units in region j For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,t Let λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

[0084] Furthermore, the pre-constructed day-ahead computer-computer co-optimization model includes: the day-ahead computer-computer co-optimization objective function and its corresponding constraints.

[0085] Furthermore, the objective function for the current computer-computer collaborative optimization is as follows:

[0086]

[0087] In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

[0088] Furthermore, the constraints include: power balance constraints, power system generation capacity constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

[0089] Furthermore, the power balance constraints are as follows:

[0090] P IT (t)+P cool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t)

[0091] The power generation constraints of the power system are as follows:

[0092]

[0093] The power generation capacity ramp-up rate constraint is as follows:

[0094]

[0095] The constraints of the energy storage system are as follows:

[0096]

[0097] The temperature constraints for the computer room are as follows:

[0098]

[0099] The resource allocation constraints are as follows:

[0100]

[0101] The task deadline constraints are as follows:

[0102]

[0103] In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant. gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η chargeFor the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

[0104] Furthermore, the indoor temperature of region j at time t+1 is as follows:

[0105]

[0106] In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows:

[0107]

[0108] In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

[0109] Furthermore, the device includes:

[0110] The fitting module is used to perform curve fitting on the optimization results corresponding to each time point in the next 24 hours to obtain the optimization result curve.

[0111] The second acquisition module is used to obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve;

[0112] The second analysis module is used to substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points.

[0113] The second scheduling module is used to perform intraday optimization scheduling of the computing center based on the intraday optimization results corresponding to the future 96 points.

[0114] Furthermore, the pre-constructed intraday computer-computer collaborative optimization model is as follows:

[0115]

[0116] In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info (τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

[0117] Thirdly, a computer device is provided, comprising: one or more processors;

[0118] The processor is used to store one or more programs;

[0119] When the one or more programs are executed by the one or more processors, the computing center scheduling method that takes into account the combination of power and computing power is implemented.

[0120] Fourthly, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the computer program is executed, the computing center scheduling method considering the combination of power and computing power is implemented.

[0121] The above-described technical solutions of the present invention have at least one or more of the following beneficial effects:

[0122] This invention provides a method and apparatus for scheduling computing centers that considers the integration of power and computing power. The method includes: acquiring the information entropy of the computing center and the energy entropy of the power system; substituting the information entropy of the computing center and the energy entropy of the power system into a pre-constructed day-ahead power-computing collaborative optimization model and solving it to obtain optimization results; and performing day-ahead optimized scheduling of the computing center based on the optimization results. The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage device charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, start time of computing tasks, total running time, end time, and start / stop status of the computing server. The technical solution provided by this invention quantifies the schedulability of computing tasks by defining information entropy and characterizes the spatiotemporal supply and demand differences of the power system by energy entropy. It constructs a scheduling optimization model to achieve power-computing collaborative scheduling, realizing unified collaborative scheduling of power and computing power, reducing energy consumption and improving computing power utilization. Attached Figure Description

[0123] Figure 1 This is a schematic diagram of the main steps of the computing center scheduling method considering the combination of power and computing power in an embodiment of the present invention;

[0124] Figure 2 This is a schematic diagram of the coordinated fluctuation of energy entropy and information entropy in an embodiment of the present invention. Detailed Implementation

[0125] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0126] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0127] As disclosed in the background section, with the rapid development of artificial intelligence, cloud computing, and the Internet of Things, computing centers have become a core infrastructure supporting the digital economy. The high energy consumption problem brought about by the expansion of computing centers is becoming increasingly serious, with their electricity consumption accounting for a continuously rising proportion of global electricity consumption. At the same time, the increasing proportion of renewable energy and the intensified uncertainty on both the source and load sides under the new power system pose challenges to solving power system scheduling problems, including dimensionality explosion and difficulty in finding solutions. Power-computing power coordinated scheduling is a powerful method to achieve efficient energy use in data centers and the absorption of new energy sources. Existing scheduling methods mostly focus on local optimization, such as reducing the energy consumption of a single device through dynamic voltage-frequency regulation (DVFS) or using real-time electricity prices to guide task load migration. Existing methods cannot fully coordinate the flexibility of power and computing power, and have the following shortcomings:

[0128] Existing technologies do not fully consider the spatiotemporal distribution characteristics of electricity supply and demand (such as regional electricity price differences, renewable energy volatility, and power grid capacity), resulting in low energy utilization.

[0129] Many computational tasks (such as offline training and batch data processing) have time-delay characteristics, but existing scheduling models lack quantitative representation of their flexibility and cannot achieve global optimization.

[0130] Power system regulation strategies under high-proportion renewable energy are not deeply coupled with computing power task scheduling, making it difficult to mitigate fluctuations in power supply and demand.

[0131] In summary, the fundamental problem is the lack of quantification of the coupling relationship between power and computing power, namely, the lack of a unified quantification standard for the energy distribution of the power system and the complexity of computing tasks, which makes it impossible to establish a joint optimization goal.

[0132] Therefore, there is an urgent need for a scheduling method that can jointly characterize the dynamic demands of computing power tasks and the spatiotemporal characteristics of the power system, so as to fully tap the synergistic potential between computing power center tasks and power system scheduling.

[0133] To address the aforementioned issues, this invention provides a computing center scheduling method and apparatus that considers the integration of power and computing power. The method includes: acquiring the information entropy of the computing center and the energy entropy of the power system; substituting the information entropy of the computing center and the energy entropy of the power system into a pre-constructed day-ahead power-computing collaborative optimization model and solving it to obtain optimization results; and performing day-ahead optimized scheduling of the computing center based on the optimization results. The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage device charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status of the computing center's energy system nodes. The technical solution provided by this invention quantifies the schedulability of computing tasks by defining information entropy, characterizes the spatiotemporal supply and demand differences of the power system by energy entropy, constructs a scheduling optimization model, realizes power-computing collaborative scheduling, achieves unified collaborative scheduling of power and computing power, reduces energy consumption, and improves computing power utilization.

[0134] The above plan will be explained in detail below.

[0135] Example 1

[0136] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of a computing center scheduling method considering the combination of power and computing power according to an embodiment of the present invention. Figure 1 As shown, the computing center scheduling method considering the combination of power and computing power in this embodiment of the invention mainly includes the following steps:

[0137] Step S101: Obtain the information entropy of the computing center and the energy entropy of the power system;

[0138] Step S102: Substitute the information entropy of the computing center and the energy entropy of the power system into the pre-constructed day-ahead power-computing collaborative optimization model and solve it to obtain the optimization result;

[0139] Step S103: Perform day-ahead optimization scheduling on the computing center based on the optimization results;

[0140] The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

[0141] In this embodiment, the smaller the information entropy value, the higher the system data processing efficiency and the more balanced the computing power distribution. The information entropy of the computing power center is as follows:

[0142]

[0143] In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

[0144] In one implementation, energy entropy is used to quantify the disorder of energy distribution in a power system, reflecting the dynamic balance characteristics of renewable energy output fluctuations, load demand changes, and energy storage regulation capabilities. The power system energy entropy of the computing center is as follows:

[0145]

[0146] In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ P P respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load (t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

[0147] In one specific implementation, a diagram illustrating the coordinated fluctuations of energy entropy and information entropy is shown below. Figure 2 As shown, specifically:

[0148] (1) The impact of computing power fluctuations on energy entropy

[0149] Scenario 1: Sudden high-latency sensitive task (information entropy increases): A → B

[0150] A surge in computing power load leads to increased server power consumption, which in turn leads to a sharp increase in electricity load. If renewable energy output is insufficient, energy entropy will rise.

[0151] Control strategies: Activate energy storage discharge or switch to backup server clusters to reduce energy entropy.

[0152] Scenario 2: Interruptible task migration (information entropy decreases): C→D

[0153] Tasks are transferred to low-load nodes → server load balancing → localized cooling energy consumption is reduced → energy entropy decreases.

[0154] (2) The reverse effect of energy fluctuations on information entropy

[0155] Scenario 1: Photovoltaic output drops sharply (energy entropy increases): E→F

[0156] Insufficient power supply → triggers server frequency reduction or task migration → increases task processing latency → increases information entropy.

[0157] Scenario 2: Efficient charging and discharging of energy storage systems (energy entropy ↓): G → H

[0158] Stable energy supply → supports dynamic allocation of computing power tasks → information entropy decreases due to task queue optimization.

[0159] In one implementation, the power generation of the computing center's energy system at time t is as follows:

[0160]

[0161] The load power of the computing center at time t is as follows:

[0162]

[0163] The charging and discharging power of the computing center energy storage system at time t is as follows:

[0164]

[0165] In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

[0166] In one implementation, the IT equipment energy consumption of region j at time t is as follows:

[0167]

[0168] In the above formula, M j Let J be the number of servers in region j. The idle load of the units in region j For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,t Let λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

[0169] In one implementation, the pre-constructed day-ahead computer-computer co-optimization model includes: a day-ahead computer-computer co-optimization objective function and its corresponding constraints.

[0170] In one implementation, the day-to-day computer-computer collaborative optimization objective function is as follows:

[0171]

[0172] In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

[0173] In one implementation, the constraints include: power balance constraints, power system generation power constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

[0174] In one implementation, the power balance constraint is as follows:

[0175] P IT (t)+Pcool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t)

[0176] The power generation constraints of the power system are as follows:

[0177]

[0178]

[0179] The power generation capacity ramp-up rate constraint is as follows:

[0180]

[0181] The constraints of the energy storage system are as follows:

[0182]

[0183] The temperature constraints for the computer room are as follows:

[0184]

[0185] The resource allocation constraints are as follows:

[0186]

[0187] The task deadline constraints are as follows:

[0188]

[0189] In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant.gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η charge For the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

[0190] In one implementation, the indoor temperature of region j at time t+1 is as follows:

[0191]

[0192] In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows:

[0193]

[0194] In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

[0195] In one implementation, after performing day-ahead optimization scheduling on the computing center based on the optimization results, the process includes:

[0196] Curve fitting is performed on the optimization results at each time point in the next 24 hours to obtain the optimization result curve;

[0197] Obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve;

[0198] Substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points;

[0199] Based on the intraday optimization results corresponding to the future 96 points, the computing center is scheduled for intraday optimization.

[0200] In one implementation, the pre-built intraday computer-computer collaborative optimization model is as follows:

[0201]

[0202] In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info (τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

[0203] Example 2

[0204] Based on the same inventive concept, the present invention also provides a computing power center scheduling device that considers the combination of power and computing power, the computing power center scheduling device considering the combination of power and computing power includes:

[0205] The first acquisition module is used to acquire the information entropy of the computing center and the energy entropy of the power system.

[0206] The first analysis module is used to substitute the information entropy of the computing center and the energy entropy of the power system into the pre-constructed day-ahead power-computing collaborative optimization model and solve it to obtain the optimization results;

[0207] The first scheduling module is used to perform day-ahead optimization scheduling of the computing center based on the optimization results;

[0208] The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

[0209] Preferably, the information entropy of the computing center is as follows:

[0210]

[0211] In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

[0212] Furthermore, the energy entropy of the power system of the computing center is as follows:

[0213]

[0214] In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ P P respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load (t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

[0215] Furthermore, the power generation of the computing center's energy system at time t is as follows:

[0216]

[0217] The load power of the computing center at time t is as follows:

[0218]

[0219] The charging and discharging power of the computing center energy storage system at time t is as follows:

[0220]

[0221] In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

[0222] Furthermore, the IT equipment energy consumption in region j at time t is as follows:

[0223]

[0224] In the above formula, M j Let J be the number of servers in region j. The idle load of the units in region j For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,t Let λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

[0225] Furthermore, the pre-constructed day-ahead computer-computer co-optimization model includes: the day-ahead computer-computer co-optimization objective function and its corresponding constraints.

[0226] Furthermore, the objective function for the current computer-computer collaborative optimization is as follows:

[0227]

[0228] In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

[0229] Furthermore, the constraints include: power balance constraints, power system generation capacity constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

[0230] Furthermore, the power balance constraints are as follows:

[0231] P IT (t)+P cool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t)

[0232] The power generation constraints of the power system are as follows:

[0233]

[0234]

[0235]

[0236] The power generation capacity ramp-up rate constraint is as follows:

[0237]

[0238] The constraints of the energy storage system are as follows:

[0239]

[0240] The temperature constraints for the computer room are as follows:

[0241]

[0242] The resource allocation constraints are as follows:

[0243]

[0244] The task deadline constraints are as follows:

[0245]

[0246] In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant. gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η charge For the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

[0247] Furthermore, the indoor temperature of region j at time t+1 is as follows:

[0248]

[0249] In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows:

[0250]

[0251] In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

[0252] Furthermore, the device includes:

[0253] The fitting module is used to perform curve fitting on the optimization results corresponding to each time point in the next 24 hours to obtain the optimization result curve.

[0254] The second acquisition module is used to obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve;

[0255] The second analysis module is used to substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points.

[0256] The second scheduling module is used to perform intraday optimization scheduling of the computing center based on the intraday optimization results corresponding to the future 96 points.

[0257] Furthermore, the pre-constructed intraday computer-computer collaborative optimization model is as follows:

[0258]

[0259] In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info(τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

[0260] Example 3

[0261] Based on the same inventive concept, this invention also provides a computer device, which includes a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to implement corresponding method flows or corresponding functions, thereby realizing the steps of a power-computing power combined computing center scheduling method in the above embodiments.

[0262] Example 4

[0263] Based on the same inventive concept, this invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of a power-computing power combined computing center scheduling method in the above embodiments.

[0264] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0265] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0266] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0267] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0268] 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 it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A computing center scheduling method considering the combination of power and computing power, characterized in that, The method includes: Obtain the information entropy of the computing center and the energy entropy of the power system; The information entropy of the computing center and the energy entropy of the power system are substituted into the pre-constructed day-ahead power-computing collaborative optimization model and solved to obtain the optimization results; Based on the optimization results, the computing center is scheduled for day-ahead optimization. The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

2. The method as described in claim 1, characterized in that, The information entropy of the computing center is as follows: In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

3. The method as described in claim 2, characterized in that, The energy entropy of the power system of the computing center is as follows: In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ P P respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load (t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

4. The method as described in claim 3, characterized in that, The power generation of the computing center's energy system at time t is as follows: The load power of the computing center at time t is as follows: The charging and discharging power of the computing center energy storage system at time t is as follows: In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

5. The method as described in claim 4, characterized in that, The IT equipment energy consumption in region j at time t is as follows: In the above formula, M j Let J be the number of servers in region j. For the idle load of units in region j, P j peak For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,t Let λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

6. The method as described in claim 5, characterized in that, The pre-constructed day-ahead computer-computer co-optimization model includes: the day-ahead computer-computer co-optimization objective function and its corresponding constraints.

7. The method as described in claim 6, characterized in that, The objective function for the current computer-computer collaborative optimization is as follows: In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

8. The method as described in claim 7, characterized in that, The constraints include: power balance constraints, power system generation power constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

9. The method as described in claim 8, characterized in that, The power balance constraints are as follows: P IT (t)+P cool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t) The power generation constraints of the power system are as follows: The power generation capacity ramp-up rate constraint is as follows: The constraints of the energy storage system are as follows: The temperature constraints for the computer room are as follows: The resource allocation constraints are as follows: The task deadline constraints are as follows: In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant. gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η charge For the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

10. The method as described in claim 9, characterized in that, The indoor temperature of region j at time t+1 is as follows: In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows: In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

11. The method as described in claim 10, characterized in that, After performing day-ahead optimization scheduling on the computing center based on the optimization results, the process includes: Curve fitting is performed on the optimization results at each time point in the next 24 hours to obtain the optimization result curve; Obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve; Substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points; Based on the intraday optimization results corresponding to the future 96 points, the computing center is scheduled for intraday optimization.

12. The method as described in claim 11, characterized in that, The pre-constructed intraday computer-computer collaborative optimization model is as follows: In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info (τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

13. A computing center scheduling device considering the integration of power and computing power, characterized in that, The device includes: The first acquisition module is used to acquire the information entropy of the computing center and the energy entropy of the power system. The first analysis module is used to substitute the information entropy of the computing center and the energy entropy of the power system into the pre-constructed day-ahead power-computing collaborative optimization model and solve it to obtain the optimization results; The first scheduling module is used to perform day-ahead optimization scheduling of the computing center based on the optimization results. The optimization results include at least one of the following: photovoltaic power generation, wind turbine power generation, self-owned power plant power generation, energy storage equipment charging and discharging power, cooling system power, grid-purchased electricity power of the computing center, computing task start time, total running time, end time, and computing server start / stop status.

14. The apparatus as claimed in claim 13, characterized in that, The information entropy of the computing center is as follows: In the above formula, H info Let M be the information entropy of the computing center, M be the total number of computing tasks, and p be the information entropy of the computing center. m Let q be the probability distribution of the type of computing task m, N be the total number of nodes in the computing center's energy system, and q be the probability distribution of the type of computing task m. i Let λ be the load percentage of node i in the computing center's energy system, λ1 be the first weighting coefficient, Var(L) be the variance of the load vector L, λ2 be the second weighting coefficient, and Avg(D) be the mean of the communication link delay vector D.

15. The apparatus as claimed in claim 14, characterized in that, The energy entropy of the power system of the computing center is as follows: In the above formula, H energy Let μ be the energy entropy of the power system of the computing center. P and σ P P respectively net The mean and standard deviation of (t), where T is the scheduling period, K is the total number of energy storage devices, and S is the standard deviation of (t). k (t) represents the state of charge of energy storage device k, S opt To preset the optimal state of charge, ΔS max P represents the allowable fluctuation range of the state of charge of energy storage devices. net (t) represents the net power of the computing center's energy system, including energy storage, at time t. net (t)=P gen (t)-P load (t)+P storage (t), P gen (t) represents the power generation of the computing center's energy system at time t, P load (t) represents the load power of the computing center at time t, P storage (t) represents the charging and discharging power of the computing center's energy storage system at time t, and γ is the third weighting coefficient.

16. The apparatus as claimed in claim 15, characterized in that, The power generation of the computing center's energy system at time t is as follows: The load power of the computing center at time t is as follows: The charging and discharging power of the computing center energy storage system at time t is as follows: In the above formula, Let be the photovoltaic power generation of node i in the energy system of the computing center at time t. Let t be the wind turbine power generation capacity of node i in the computing power center energy system. Let t be the power generation capacity of the self-owned power plant of node i in the energy system of the computing center. Let be the total energy consumption of the cooling system in region j at time t. Let P be the power consumption of IT equipment in region j at time t. other (t) represents the other energy consumption of the computing center at time t, η storage For energy storage charging and discharging efficiency, Let t be the charging and discharging power of energy storage device k in the computing center energy storage system at time t. Q IT The heat generated by IT equipment is represented by COP, which stands for coefficient of performance in cooling systems.

17. The apparatus as claimed in claim 16, characterized in that, The IT equipment energy consumption in region j at time t is as follows: In the above formula, M j Let J be the number of servers in region j. The idle load of the units in region j For the full load of the units in region j, W j,t Let be the total load of region j at time t. W represents the computational capacity of region j. j,t =μ j,t +∑ X λ x,j,t μ j,t Let λ be the load of the units in region j that are not subject to scheduling. x,j,t Let X be the load of unit x participating in the scheduling in region j, and let X be the total number of units participating in the scheduling in region j.

18. The apparatus as claimed in claim 17, characterized in that, The pre-constructed day-ahead computer-computer co-optimization model includes: the day-ahead computer-computer co-optimization objective function and its corresponding constraints.

19. The apparatus as claimed in claim 18, characterized in that, The objective function for the current computer-computer collaborative optimization is as follows: In the above formula, H energy H(t) represents the information entropy of the computing center at time t. info (t) represents the power system energy entropy of the computing center at time t, α is the information entropy weight, β is the energy entropy weight, and c grid (t) represents the electricity price at time t, P grid (t) represents the power consumption of the computing center at time t.

20. The apparatus as claimed in claim 19, characterized in that, The constraints include: power balance constraints, power system generation power constraints, generation capacity ramp-up rate constraints, energy storage system constraints, computer room temperature constraints, resource allocation constraints, and task deadline constraints.

21. The apparatus as claimed in claim 20, characterized in that, The power balance constraints are as follows: P IT (t)+P cool (t)≤P grid (t)+P PV (t)+P Wind (t)+P storage (t)+P gas (t) The power generation constraints of the power system are as follows: The power generation capacity ramp-up rate constraint is as follows: The constraints of the energy storage system are as follows: The temperature constraints for the computer room are as follows: The resource allocation constraints are as follows: The task deadline constraints are as follows: In the above formula, P grid (t) represents the power consumption of the computing center at time t, P. PV (t) represents the photovoltaic power generation of the computing center's energy system, P Wind (t) represents the wind turbine power generation capacity of the computing center's energy system. and These represent the minimum and maximum power generation values ​​of the self-owned power plant in the computing center's energy system, respectively. and These are the minimum and maximum photovoltaic power generation values ​​of the energy system in the computing center, respectively. and These represent the minimum and maximum power generation of the wind turbines in the computing center's energy system, P. gas (t) represents the power generation capacity of the self-owned power plant in the computing center's energy system at time t. This represents the maximum ramp rate for a self-owned power plant. P represents the maximum downhill ramp rate of the self-owned power plant. gas (t-1) represents the power generation capacity of the self-owned power plant of the computing center's energy system at time t-1, P IT (t) represents the energy consumption of the IT equipment at time t, P cool (t) represents the energy consumption of the cooling system at time t, SOC(t+1) represents the state of charge of the energy storage system at time t+1, and SOC(t) represents the state of charge of the energy storage system at time t. η charge For the charging efficiency of energy storage systems, P charge (t) represents the charging power of the computing center's energy storage system at time t, P discharge (t) represents the discharge power of the computing center's energy storage system at time t, and η discharge For the discharge efficiency of the energy storage system, Let T be the indoor temperature of region j at time t+1. max The highest operating temperature in the computer room, x m,n (t) represents the resource utilization rate of computing task m on computing node n in the computing center. To calculate the total computing capacity of node n in the computing center, Let m be the start time of the computing task m. Let m be the total runtime of the computing task. t is the deadline for computing task m.

22. The apparatus as claimed in claim 21, characterized in that, The indoor temperature of region j at time t+1 is as follows: In the above formula, Let Δt be the indoor temperature of region j at time t, and Δt be the temperature per unit time, C. j Let J be the heat capacity of the thermal channel in region j. The cooling capacity provided by the cooling system for region j at time t. For the server in region j at time t, heat is generated. For the heat transfer of region j on the outer wall at time t, Let be the solar radiation heat in region j at time t. For the illumination heat generation in region j at time t, K j Let c be the number of wind turbines in region j. a m is the amount of air. j,t Let be the wind speed of the fan in region j at time t. Let be the supply air temperature of region j at time t. The calculation formula is as follows: In the above formula, Let be the supply air temperature of region j at time t-1. Let J be the cold aisle heat capacity. The amount of cooling provided by the cooling system for region j at time t-1.

23. The apparatus as claimed in claim 22, characterized in that, The device includes: The fitting module is used to perform curve fitting on the optimization results corresponding to each time point in the next 24 hours to obtain the optimization result curve. The second acquisition module is used to obtain the reference values ​​of the optimization results for the next 96 points from the optimization result curve; The second analysis module is used to substitute the optimization result reference values ​​corresponding to the future 96 points into the pre-constructed intraday computer-computer collaborative optimization model and solve it to obtain the intraday optimization results corresponding to the future 96 points. The second scheduling module is used to perform intraday optimization scheduling of the computing center based on the intraday optimization results corresponding to the future 96 points.

24. The apparatus as claimed in claim 23, characterized in that, The pre-constructed intraday computer-computer collaborative optimization model is as follows: In the above formula, H energy (τ) represents the information entropy of the computing center at time τ, H info (τ) represents the power system energy entropy of the computing center at time τ, w is the deviation weight, and u(τ) is the intraday optimization result corresponding to time τ. rf(τ) This is the reference value for the optimization result at time τ.

25. A computer device, characterized in that, include: one or more processors; The processor is used to execute one or more programs; When the one or more programs are executed by the one or more processors, the computing center scheduling method considering the combination of power and computing power as described in any one of claims 1 to 12 is implemented.

26. A computer-readable storage medium, characterized in that, It contains a computer program, which, when executed, implements the computing center scheduling method that considers the combination of power and computing power as described in any one of claims 1 to 12.

Images