Green electric power computing power fusion scheduling model construction method and scheduling platform

Abstract

The invention discloses a green power computing power fusion scheduling model construction method and a scheduling platform, and relates to the technical field of green computing power. The green power computing power fusion scheduling model construction method comprises the steps of determining an adjustable reference load, determining a peak clipping and valley filling optimization load, determining a clean energy output load, determining a computing power center computing power task execution optimal load curve, monitoring a computing power center fusion scheduling task execution process and the like. A green electric power computing power fusion scheduling platform is constructed, and fusion scheduling task execution responsivity monitoring, computing power center electricity charge economy analysis, computing power center carbon emission reduction analysis and the like of a computing power center are carried out. The beneficial effects of the invention are that the system employs the technologies of big data, cloud computing and the like, builds a green power calculation power scheduling platform, monitors and analyzes the clean energy supply and demand conditions of the calculation power centers of the whole province, provides calculation power task scheduling services for the calculation power centers of the whole province, integrates clean energy consumption and green power calculation power, and improves the efficiency of the whole province. And the calculation power load active matching clean energy supply is realized.

Description

Technical Field

[0001] The present invention relates to the field of green computing power technology, and in particular to a method for constructing a green power computing power fusion scheduling model and a scheduling platform. Background Art

[0002] Computing power is an important support for the development of the digital economy. It not only provides support for the transformation and upgrading of traditional industries, but also can generate new economic growth points. The processing and digital application of massive data have continuously put forward new requirements for the development of computing power from the demand side. The improvement of computing power and the increase of algorithm complexity will also provide a continuous new impetus for the development of the digital economy from the supply side.

[0003] In recent years, with the rapid development of the digital economy and the widespread application of information technology, the total scale of my country's computing power industry has grown rapidly at an average annual rate of nearly 30%, ranking second in the world. At the same time, the development of traditional computing power faces problems such as high energy consumption and high emissions. Promoting the green and low-carbon development of computing power has become an industry consensus.

[0004] Qinghai Province is rich in clean energy resources, but the regulating power supply is insufficient, resulting in a mismatch in the peak-shaving demand of the power system, making it difficult to consume daytime clean energy on a large scale. How to integrate clean energy consumption and green power computing power is the focus of this invention. Summary of the invention

[0005] In order to solve the above-mentioned technical problems of clean energy consumption and green power computing power integration, the present invention provides a green power computing power integration scheduling model construction method and scheduling platform, uses big data, cloud computing and other technologies to build a green power computing power scheduling platform, monitors and analyzes the clean energy supply and demand of computing power centers across the province, and provides computing power task scheduling services for all computing power centers in the province, so as to realize the active matching of computing power load with clean energy supply.

[0006] The present invention provides a method for constructing a green power computing power fusion scheduling model, comprising the following steps:

[0007] Step 1: Analyze the computing center load and determine the adjustable benchmark load;

[0008] Step 2: Determine the peak-to-valley load optimization based on the peak-to-valley electricity price period and the next day's baseline load;

[0009] Step 3: Analyze the clean energy output and determine the clean energy output load;

[0010] Step 4: The computing center calculates the optimal load curve for computing tasks;

[0011] Step 5: Monitor the execution process of the computing power center's fusion scheduling tasks.

[0012] The first step is to analyze the computing center load and determine the adjustable reference load, specifically:

[0013] Based on the power load of the computing center, analyze the load change pattern. Determine the baseline load at the corresponding time point of the next day based on the 7-day load of the computing center. Figure 1 shown.

[0014] The specific calculation formula is: i,可调节基准负荷 =P i,k,min +(P i,k,max -P i,k,min )×γ

[0015] Where P i,k,min , P i,k,max They represent the minimum load and maximum load at a certain time point in the past 7 days respectively. γ is an adjustable coefficient (which can be adjusted dynamically according to the actual load, usually γ∈[0,1]), representing 96 time points.

[0016] The second step is to determine the peak load shaving and valley filling optimization load based on the peak and valley electricity price period, specifically:

[0017] Electricity price distribution diagram Figure 2 As shown in the figure, the electricity price distribution chart shows that the electricity price is at its trough period from 9:00 to 17:00, and electricity consumption should be encouraged to save electricity costs. The electricity price is at its peak period from 7:00 to 9:00 and from 17:00 to 23:00, and electricity consumption should be reduced at this time.

[0018] According to the different electricity prices in peak, valley and normal periods, for cost considerations, dynamic adjustments are made while meeting the computing power requirements; electricity consumption during the peak period is dispatched to the valley period, and then "peak shaving and valley filling" is performed to meet the computing power requirements; the specific calculation formula is:

[0019]

[0020] β represents the number that maps the electricity price to [-1, 1]. The specific formula is: Where E0 represents the current electricity price, E min Indicates the minimum electricity price, E max Indicates the maximum electricity price. i,峰 It represents the peak power load corresponding to the peak electricity price period. The specific formula is P i,峰 =P i,可调节基准负荷 -P 历史负荷均值 , n1 and n2 represent the total number of time points in the peak and valley electricity price periods, respectively, and t represents the load duration period.

[0021] The third step is to load analyze the clean energy output and determine the clean energy output load. In order to absorb more clean energy, electricity consumption should be encouraged during the peak period of clean energy output and electricity consumption should be reduced during the low period of output. The clean energy output reasoning diagram is as follows: Figure 3 As shown; the clean energy output calculation formula is: Pi,清洁能源出力 =P i *(α j,1 -α j,2 ), where α j,1 is the proportion of clean energy in this period, and the calculation formula is: W j,清洁 is the total amount of clean energy during this period, W j,总 is the total energy in this period; α j,2 is the average proportion of clean energy in the past 7 days during this period, and the calculation formula is: H j,清洁 is the total clean energy in the past 7 days during this period, H j,总 is the total energy of the historical 7 days in this period, j=1,2,3 represent the peak electricity price period, wave electricity price period and flat electricity price period respectively.

[0022] The fourth step is to determine the ideal load curve for the next day, specifically: summarizing and superimposing the above-mentioned various types of dispatching loads to obtain the ideal load curve; the calculation formula is:

[0023] P i,调度 =P i,可调节基准负荷 +P 削峰填谷优化 +P i,清洁能源出力 ; i represents 96 time points;

[0024] In order to absorb more clean energy the next day, electricity consumption should be encouraged during the peak period of clean energy output, and electricity consumption should be reduced during the low output period.

[0025] The step five is to monitor the execution process of the fusion scheduling task of the computing power center, including the responsiveness monitoring of the fusion scheduling task execution of the computing power center, the economic analysis of the electricity cost of the computing power center, and the carbon emission reduction analysis of the computing power center.

[0026] The computing power center fusion scheduling responsiveness monitoring is specifically: based on the green power computing power fusion scheduling platform, online monitoring of the actual response of the fusion scheduling tasks of each computing power center, the specific responsiveness calculation process is:

[0027] Responsiveness Among them, j=1, 2, and 3 represent the peak electricity price period, wave electricity price period, and flat electricity price period respectively.

[0028] Where N j is the set of loads that respond in the corresponding time period, M j is the load set for the corresponding period.

[0029] Peak electricity price period:

[0030] N j ={X∈M j And X<P i}, where P iIndicates the scheduling load at the corresponding time.

[0031] Off-peak electricity price period:

[0032] N j ={X∈M j And X>P i}

[0033] Flat electricity price period:

[0034] (1)α j,1 -α j,2 >0, where α j,1 is the proportion of clean energy in this period, α j,2 The average proportion of clean energy in the past 7 days

[0035] N j ={X∈M j And X>P i}

[0036] (2)α j,1 -α j,2 <0

[0037] N j ={X∈M j And X<P i},

[0038] The economic analysis of electricity cost of the computing power center is specifically as follows: according to the fusion scheduling response rate, the valley electricity usage of each computing power center is counted, and the specific calculation formula is:

[0039] Response power j=1, 2, 3 represent the peak electricity price period, valley electricity price period, and flat electricity price period respectively, and i represents 96 time points.

[0040] The carbon emission reduction analysis of the computing power center is specifically as follows: according to the fusion scheduling response rate, it is calculated that the use of valley electricity is increased in the process of "computing follows source movement", the use of thermal power during peak hours is reduced, and the carbon emissions of thermal power are reduced.

[0041] A green power computing power fusion scheduling platform includes a computing power equipment monitoring and analysis module, a power computing power fusion scheduling module, and a green power transaction monitoring module; the computing power equipment monitoring and analysis module includes a computing power resource monitoring sub-module, a computing power power consumption monitoring sub-module, and a computing power center three-dimensional scene monitoring sub-module; the green power transaction monitoring module includes a green power intra-provincial transaction sub-module and a green power inter-provincial transaction sub-module.

[0042] The functions of the computing power equipment monitoring and analysis module include: the number of general computing servers in the unified computing power center; the total capacity of all server storage devices in the unified computing power center; the number of GPUs of all servers in the unified computing power center; the number of intelligent computing (such as AI training, reasoning, etc.) servers in the unified computing power center; the total capacity of the memory of all servers in the unified computing power center; the total computing power value of the unified computing power center; the real-time computing power value of the unified computing power center every 15 minutes, including the real-time computing power value of GPU and CPU; the total number of racks in the unified computing power center; the total PUE of the unified computing power center; the total computing power value of all servers in the unified computing power center Rated power consumption, including CPU and memory power consumption; rated power consumption of all computing servers in the unified computing power center, including GPU power consumption; rated power consumption of storage in the unified computing power center; rated power consumption of the network in the unified computing power center; real-time power consumption of the unified computing power center every 15 minutes; thermal map of power trends and clean energy power generation in the province at 330KV and above; personalized customization of overall three-dimensional scenes for different computing power centers; three-dimensional modeling of the floors where the computer rooms of different computing power centers are located and monitoring of the relevant computing power information of the floors; three-dimensional modeling of the cabinets and equipment in the cabinets in the computer rooms of different computing power centers and monitoring of the relevant computing power information of the floors.

[0043] The functions of the electric power computing power fusion scheduling module include: predicting the planned output of various energy sources in the province; monitoring the historical load of the computing power center in the past seven days; calculating the peak, valley and flat time-sharing electricity prices in the province; calculating the adjustable variables of the computing power center's load shaving and valley filling; calculating the output of clean energy in the province and the curve after load adjustment; generating the computing power task scheduling load curve through fusion analysis; monitoring and analyzing the responsiveness of the computing power center in the past week; monitoring and analyzing the electricity consumption during peak electricity prices, valley electricity prices and flat electricity prices.

[0044] The functions of the green electricity transaction monitoring module include: monitoring the green electricity transaction volume within the province in the past 12 months; the green electricity transaction details within the province include the seller's name, the buyer's name, the transaction volume and other data; monitoring the inter-provincial green electricity transaction volume in the past 12 months; the green electricity inter-provincial transaction details include the seller's name, the buyer's name, the transaction volume and other data.

[0045] The beneficial effects of the present invention are:

[0046] Using technologies such as big data and cloud computing, we will build a green power computing power dispatching platform to monitor and analyze the supply and demand of clean energy in computing power centers across the province, provide computing power task dispatching services for computing power centers across the province, actively match computing power load with clean energy supply, help the province's green computing power resources to efficient use and reasonable allocation, and promote the innovative development of the integration of clean energy and green computing power.

[0047] Based on the company's cloud platform, data center, artificial intelligence platform, real-time measurement center and other platforms, we will build a green power computing power dispatching center that integrates clean energy dispatching, green computing power balance, green electricity value realization, and energy and carbon management in the province. The center provides services including green power load monitoring, centralized power forecasting, power computing power dispatching, green power and green certificate trading monitoring to governments, computing power data centers, power generation and electricity-consuming enterprises.

[0048] Using big data models, cloud computing and other technologies, we conduct research on the integrated scheduling of electric power computing power, monitor the supply and demand of electric power computing power in real time, and actively match the computing power load with the supply of clean energy by transferring and allocating computing power tasks. This will help the province's green computing power resources to be efficiently utilized and reasonably allocated, and achieve "computing follows the source", providing efficient, stable and reliable control services to promote the innovative development of the integration of clean energy and green computing power. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A schematic diagram of the baseline load at the corresponding time point of the next day determined by the present invention;

[0050] Figure 2 This is a schematic diagram of electricity price distribution according to the present invention;

[0051] Figure 3 Schematic diagram of the reasoning for clean energy output. DETAILED DESCRIPTION

[0052] Embodiment 1, the present invention provides a method for constructing a green power computing power fusion scheduling model, comprising the following steps:

[0053] Step 1: Analyze the computing center load and determine the adjustable benchmark load;

[0054] Step 2: Determine the peak-to-valley load optimization based on the peak-to-valley electricity price period and the next day's baseline load;

[0055] Step 3: Analyze the clean energy output and determine the clean energy output load;

[0056] Step 4: The computing center calculates the optimal load curve for computing tasks;

[0057] Step 5: Monitor the execution process of the computing power center's fusion scheduling tasks.

[0058] The first step is to analyze the computing center load and determine the adjustable reference load, specifically:

[0059] Based on the power load of the computing center, analyze the load change pattern. Determine the baseline load at the corresponding time point of the next day based on the 7-day load of the computing center. Figure 1 shown.

[0060] The specific calculation formula is:

[0061] P i,可调节基准负荷 =P i,k,min +(P i,k,max -P i,k,min )×γ

[0062] Where P i,k,min , P i,k,max They represent the minimum load and maximum load at a certain time point in the past 7 days respectively. γ is an adjustable coefficient (which can be adjusted dynamically according to the actual load, usually γ∈[0,1]), representing 96 time points.

[0063] The second step is to determine the peak load shaving and valley filling optimization load based on the peak and valley electricity price period, specifically:

[0064] Electricity price distribution diagram Figure 2 As shown in the figure, the electricity price distribution chart shows that the electricity price is at its trough period from 9:00 to 17:00, and electricity consumption should be encouraged to save electricity costs. The electricity price is at its peak period from 7:00 to 9:00 and from 17:00 to 23:00, and electricity consumption should be reduced at this time.

[0065] According to the different electricity prices in peak, valley and normal periods, for cost considerations, dynamic adjustments are made while meeting the computing power requirements; electricity consumption during the peak period is dispatched to the valley period, and then "peak shaving and valley filling" is performed to meet the computing power requirements; the specific calculation formula is:

[0066]

[0067] β represents the number that maps the electricity price to [-1, 1]. The specific formula is: Where E0 represents the current electricity price, E min Indicates the minimum electricity price, E max Indicates the maximum electricity price. i,峰 It represents the peak power load corresponding to the peak electricity price period. The specific formula is P i,峰 =P i,可调节基准负荷 -P 历史负荷均值 , n1 and n2 represent the total number of time points in the peak and valley electricity price periods, respectively, and t represents the load duration period.

[0068] The step three, load analysis of clean energy output, determines the clean energy output load, specifically:

[0069] In order to absorb more clean energy, electricity consumption should be encouraged during peak clean energy output periods and electricity consumption should be reduced during low output periods. Figure 3 As shown; the specific calculation formula is: P i,清洁能源出力 =P i *(α j,1 -α j,2 ), where α j,1is the proportion of clean energy in this period, and the calculation formula is: W j,清洁 is the total amount of clean energy during this period, W j,总 is the total amount of energy in this period. j,2 is the average proportion of clean energy in the past 7 days during this period, and the calculation formula is: H j,清洁 is the total clean energy in the past 7 days during this period, H j,总 is the total energy of the historical 7 days in this period, j=1,2,3 represent the peak electricity price period, wave electricity price period and flat electricity price period respectively.

[0070] The fourth step is to determine the ideal load curve for the next day, specifically: summarizing and superimposing the above-mentioned various types of dispatching loads to obtain the ideal load curve. The calculation formula is:

[0071] P i,调度 =P i,可调节基准负荷 +P 削峰填谷优化 +P i,清洁能源出力 , i represents 96 time points;

[0072] In order to absorb more clean energy the next day, electricity consumption should be encouraged during the peak period of clean energy output, and electricity consumption should be reduced during the low period of output. Figure 3 shown.

[0073] The step five is to monitor the execution process of the integrated scheduling tasks of the computing power center, specifically: based on the green power computing power integration scheduling platform, carry out the responsiveness monitoring of the integrated scheduling tasks of each computing power center, the economic analysis of the electricity cost of the computing power center, and the carbon emission reduction analysis of the computing power center, to help the province's clean energy resources to be efficiently utilized and reasonably allocated, and promote the innovative development of the integration of clean energy and green computing power.

[0074] Monitoring scenario 1: Monitoring the responsiveness of the fusion scheduling of the computing power center. Based on the green power computing power fusion scheduling platform, the actual response of the fusion scheduling tasks of each computing power center is monitored online. The specific responsiveness calculation process is as follows:

[0075] Responsiveness Among them, j=1, 2, and 3 represent the peak electricity price period, wave electricity price period, and flat electricity price period respectively.

[0076] Where N j is the set of loads that respond in the corresponding time period, M j is the load set for the corresponding period.

[0077] Peak electricity price period:

[0078] N j ={X∈M j And X<P i}, where P iIndicates the scheduling load at the corresponding time.

[0079] Off-peak electricity price period:

[0080] N j ={X∈M j And X>P i}

[0081] Flat electricity price period:

[0082] (1)α j,1 -α j,2 >0, where α j,1 is the proportion of clean energy in this period, α j,2 7 for this period

[0083] Average share of clean energy per day

[0084] N j ={X∈M j And X>P i}

[0085] (2)α j,1 -α j,2 <0

[0086] N j ={X∈M j And X<P i},

[0087] Monitoring scenario 2: Economic analysis of electricity costs in computing centers. Based on the fusion scheduling response rate, the valley power usage of each computing center is counted. The specific calculation formula is as follows:

[0088] Response power j=1, 2, 3 represent the peak electricity price period, valley electricity price period, and flat electricity price period respectively, and i represents 96 time points.

[0089] Monitoring scenario 3: Carbon emission reduction analysis of computing power center. According to the fusion dispatch response rate, it is calculated that the use of valley electricity is increased in the process of "computing follows the source", the use of thermal power during peak hours is reduced, and the carbon emissions of thermal power are reduced.

[0090] Embodiment 2, a green power computing power fusion scheduling platform, including a computing power equipment monitoring and analysis module, a power computing power fusion scheduling module, and a green electricity transaction monitoring module; the computing power equipment monitoring and analysis module includes a computing power resource monitoring sub-module, a computing power power consumption monitoring sub-module, and a computing power center three-dimensional scene monitoring sub-module; the green electricity transaction monitoring module includes a green electricity intra-provincial transaction sub-module and a green electricity inter-provincial transaction sub-module.

[0091] The functions of the computing power equipment monitoring and analysis module include: the number of general computing servers in the unified computing power center; the total capacity of all server storage devices in the unified computing power center; the number of GPUs of all servers in the unified computing power center; the number of intelligent computing (such as AI training, reasoning, etc.) servers in the unified computing power center; the total capacity of the memory of all servers in the unified computing power center; the total computing power value of the unified computing power center; the real-time computing power value of the unified computing power center every 15 minutes, including the real-time computing power value of GPU and CPU; the total number of racks in the unified computing power center; the total PUE of the unified computing power center; the total computing power value of all servers in the unified computing power center Rated power consumption, including CPU and memory power consumption; rated power consumption of all computing servers in the unified computing power center, including GPU power consumption; rated power consumption of storage in the unified computing power center; rated power consumption of the network in the unified computing power center; real-time power consumption of the unified computing power center every 15 minutes; thermal map of power trends and clean energy power generation in the province at 330KV and above; personalized customization of overall three-dimensional scenes for different computing power centers; three-dimensional modeling of the floors where the computer rooms of different computing power centers are located and monitoring of the relevant computing power information of the floors; three-dimensional modeling of the cabinets and equipment in the cabinets in the computer rooms of different computing power centers and monitoring of the relevant computing power information of the floors.

[0092] The functions of the electric power computing power fusion scheduling module include: predicting the planned output of various energy sources in the province; monitoring the historical load of the computing power center in the past seven days; calculating the peak, valley and flat time-sharing electricity prices in the province; calculating the adjustable variables of the computing power center's load shaving and valley filling; calculating the output of clean energy in the province and the curve after load adjustment; generating the computing power task scheduling load curve through fusion analysis; monitoring and analyzing the responsiveness of the computing power center in the past week; monitoring and analyzing the electricity consumption during peak electricity prices, valley electricity prices and flat electricity prices.

[0093] The functions of the green electricity transaction monitoring module include: monitoring the green electricity transaction volume within the province in the past 12 months; the green electricity transaction details within the province include the seller's name, the buyer's name, the transaction volume and other data; monitoring the inter-provincial green electricity transaction volume in the past 12 months; the green electricity inter-provincial transaction details include the seller's name, the buyer's name, the transaction volume and other data.

[0094] Monitoring scenario 1: Monitoring the responsiveness of the fusion scheduling of the computing power center. Based on the green power computing power fusion scheduling platform, the actual response of the fusion scheduling tasks of each computing power center is monitored online. The specific responsiveness calculation process is as follows:

[0095] Responsiveness Among them, j=1, 2, and 3 represent the peak electricity price period, wave electricity price period, and flat electricity price period respectively.

[0096] Where N j is the set of loads that respond in the corresponding time period, M j is the load set for the corresponding period.

[0097] Peak electricity price period:

[0098] N j ={X∈M j And X<P i}, where P i Indicates the scheduling load at the corresponding time.

[0099] Off-peak electricity price period:

[0100] N j ={X∈M j And X>P i}

[0101] Flat electricity price period:

[0102] (1)α j,1 -α j,2 >0, where α j,1 is the proportion of clean energy in this period, α j,2 7 for this period

[0103] Average share of clean energy per day

[0104] N j ={X∈M j And X>P i}

[0105] (2)α j,1 -α j,2 <0

[0106] N j ={X∈M j And X<P i},

[0107] Monitoring scenario 2: Economic analysis of electricity costs in computing centers. Based on the fusion scheduling response rate, the valley power usage of each computing center is counted. The specific calculation formula is as follows:

[0108] Response power j=1, 2, 3 represent the peak electricity price period, valley electricity price period, and flat electricity price period respectively, and i represents 96 time points.

[0109] Monitoring scenario 3: Carbon emission reduction analysis of computing power center. According to the fusion dispatch response rate, it is calculated that the use of valley electricity is increased in the process of "computing follows the source", the use of thermal power during peak hours is reduced, and the carbon emissions of thermal power are reduced.

Claims

1. A method for constructing a green power computing power fusion scheduling model, characterized in that: The following steps are involved: Step 1: Analyze the computing center load and determine the adjustable benchmark load; Step 2: Determine the peak-to-valley load optimization based on the peak-to-valley electricity price period and the next day's baseline load; Step 3: Analyze the clean energy output and determine the clean energy output load; Step 4: The computing center calculates the optimal load curve for computing tasks; Step 5: Monitor the execution process of the computing power center's fusion scheduling tasks.

2. A method for constructing a green power computing power fusion scheduling model according to claim 1, characterized in that: The first step is to analyze the computing center load and determine the adjustable reference load, specifically: Based on the power load of the computing center, analyze the load change pattern; determine the baseline load at the corresponding time point of the next day based on the 7-day load of the computing center; The specific calculation formula is: P i,可调节基准负荷 =P i,k,min +(P i,k,max -P i,k,min )×γ Where P i,k,min , P i,k,max They represent the minimum load and maximum load at a certain time point in the past 7 days respectively. γ is an adjustable coefficient (which can be adjusted dynamically according to the actual load, usually γ∈[0,1]), representing 96 time points.

3. A method for constructing a green power computing power fusion scheduling model according to claim 2, characterized in that: The second step is to determine the peak load shaving and valley filling optimization load based on the peak and valley electricity price period, specifically: According to the different electricity prices in peak, valley and normal periods, for cost considerations, dynamic adjustments are made while meeting the computing power requirements; electricity consumption during the peak period is dispatched to the valley period, and then "peak shaving and valley filling" is performed to meet the computing power requirements; the specific calculation formula is: β represents the number that maps the electricity price to [-1, 1]. The specific formula is: Where E0 represents the current electricity price, E min Indicates the minimum electricity price, E max represents the maximum electricity price; P i,峰 It represents the peak power load corresponding to the peak electricity price period. The specific formula is P i,峰 =P i,可调节基准负荷 -P 历史负荷均值 , n1 and n2 represent the total number of time points in the peak and valley electricity price periods, respectively, and t represents the load duration period.

4. A method for constructing a green power computing power fusion scheduling model according to claim 3, characterized in that: In step 3, the clean energy output is analyzed by load to determine the clean energy output load. The clean energy output calculation formula is: i,清洁能源出力 =P i *(α j,1 -α j,2 ), where α j,1 is the proportion of clean energy in this period, and the calculation formula is: W j,清洁 is the total amount of clean energy during this period, W j,总 is the total energy in this period; α j,2 is the average proportion of clean energy in the past 7 days during this period, and the calculation formula is: H j,清洁 is the total clean energy in the past 7 days during this period, H j,总 is the total energy of the historical 7 days in this period, j=1,2,3 represent the peak electricity price period, wave electricity price period and flat electricity price period respectively.

5. A method for constructing a green power computing power fusion scheduling model according to claim 4, characterized in that: The step 4 is to determine the ideal load curve for the next day, specifically: summarizing and superimposing the above-mentioned various types of dispatching loads to obtain the ideal load curve; The calculation formula is: P i,调度 =P i,可调节基准负荷 +P 削峰填谷优化 +P i,清洁能源出力 ; i represents 96 time points.

6. A method for constructing a green power computing power fusion scheduling model according to claim 5, characterized in that: The step five is to monitor the execution process of the fusion scheduling task of the computing power center, including the responsiveness monitoring of the fusion scheduling task execution of the computing power center, the economic analysis of the electricity cost of the computing power center, and the carbon emission reduction analysis of the computing power center.

7. A method for constructing a green power computing power fusion scheduling model according to claim 6, characterized in that: The computing power center fusion scheduling responsiveness monitoring is specifically: based on the green power computing power fusion scheduling platform, online monitoring of the actual response of the fusion scheduling tasks of each computing power center, the specific responsiveness calculation process is: Responsiveness Where j = 1, 2, and 3 represent the peak electricity price period, wave electricity price period, and flat electricity price period respectively; Where N j is the set of loads that respond in the corresponding time period, M j is the load set of the corresponding period; Peak electricity price period: N j ={X∈M j And X<P i }, where P i Indicates the dispatch load at the corresponding time; Off-peak electricity price period: N j ={X∈M j And X>P i } Flat electricity price period: (1)α j,1 -α j,2 >0, where α j,1 is the proportion of clean energy in this period, α j,2 The average proportion of clean energy in the past 7 days N j ={X∈M j And X>P i } (2)α j,1 -α j,2 ＜0 N j ={X∈M j And X<P i }.

8. A method for constructing a green power computing power fusion scheduling model according to claim 6, characterized in that: The economic analysis of electricity cost of the computing power center is specifically as follows: according to the fusion scheduling response rate, the valley electricity usage of each computing power center is counted, and the specific calculation formula is: Response power j=1, 2, 3 represent the peak electricity price period, valley electricity price period, and flat electricity price period respectively, and i represents 96 time points.

9. A green power computing power fusion dispatching platform, characterized in that: It includes a computing power equipment monitoring and analysis module, an electric power computing power fusion scheduling module, and a green electricity transaction monitoring module; the computing power equipment monitoring and analysis module includes a computing power resource monitoring sub-module, a computing power power consumption monitoring sub-module, and a computing power center three-dimensional scene monitoring sub-module; the green electricity transaction monitoring module includes a green electricity intra-provincial transaction sub-module and a green electricity inter-provincial transaction sub-module.

Images