Data center demand response optimization method, device, equipment and storage medium

Abstract

The application provides a data center demand response optimization method, device and equipment and a storage medium, and relates to the electric power economic field.The data center demand response optimization method comprises the following steps: acquiring the characteristics of a data center network load, the heat exchange coupling relationship of devices in the data center and the operation characteristics of IT devices and air conditioning devices in the data center; classifying the network load according to the characteristics of the data center network load; constructing a data center thermodynamic model according to the heat exchange coupling relationship of the devices in the data center; establishing an energy consumption model of the IT devices and the air conditioning devices; constructing comprehensive constraint conditions according to the classification of the network load, the data center thermodynamic model and the energy consumption model of the IT devices and the air conditioning devices; and realizing a good balance between energy saving and consumption reduction and economic benefits of the data center based on the adjustable potential of the data center and in combination with the electrothermal coupling relationship between the IT devices and the air conditioning devices in the data center.

Description

Technical Field

[0001] This invention relates to the field of power economics technology, specifically to a method, apparatus, equipment, and storage medium for optimizing data center demand response. Background Technology

[0002] With the rapid development of next-generation information technologies such as 5G and cloud computing, data centers, as a core component of cyber-physical systems, are experiencing increasing energy consumption. According to relevant statistics, in 2020, data centers accounted for 2.7% of the total electricity consumption of the entire society, with IT equipment and air conditioning / cooling equipment accounting for 85% of the total energy consumption of data centers. Meanwhile, some network loads in data centers have latency-tolerant characteristics, which allows data centers to participate in power grid regulation as a large-scale, new type of demand response entity.

[0003] Numerous studies, both domestically and internationally, have explored demand response mechanisms and energy conservation in data centers. However, significant challenges remain in addressing demand response within the context of the coupling relationships between various data center devices. In existing technologies, M. Ghamkhari proposed an IT network load management method to handle latency-tolerant jobs with equal completion deadlines while maintaining their Quality of Service (QoS). However, this optimization process did not consider the power consumption and associated costs of the cooling system. Dr. Moor used computational fluid dynamics (CFD) to simulate and predict heat distribution in data centers; however, such simulations are computationally expensive and unsuitable for real-time control applications. Therefore, a precise model is needed to explain the electrothermal coupling between data center IT equipment and air conditioning equipment, enabling the full utilization of the data center's demand response potential while ensuring user service quality. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a data center demand response optimization method, apparatus, equipment, and storage medium. Based on the adjustable potential of data centers and combined with the electrothermal coupling relationship between data center IT equipment and air conditioning equipment, a good balance is achieved between energy saving and consumption reduction and economic benefits in data centers.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] On the one hand, a data center demand response optimization method is provided, the method including:

[0009] To obtain the characteristics of data center network load, the heat exchange coupling relationship of equipment within the data center, and the operating characteristics of data center IT equipment and air conditioning equipment;

[0010] Classify network load based on the characteristics of data center network load;

[0011] Based on the heat transfer coupling relationship of equipment within the data center, a thermodynamic model of the data center is constructed.

[0012] Based on the operating characteristics of data center IT equipment and air conditioning equipment, establish energy consumption models for IT equipment and air conditioning equipment;

[0013] Based on the classification of network load, the thermodynamic model of data centers, and the energy consumption models of IT equipment and air conditioning equipment, comprehensive constraints are constructed.

[0014] Based on comprehensive constraints and with the objective of minimizing the sum of data center operating costs and customer service quality penalty costs, the data center demand response optimization problem is linearized into a mixed-integer linear programming problem.

[0015] The mixed-integer linear programming problem is solved to obtain the optimized demand response strategy for the data center.

[0016] Preferably, the comprehensive constraints include network load maximum response time constraints, air conditioning equipment power constraints, IT equipment temperature constraints, and inlet airflow temperature constraints.

[0017] Preferably, the data center network load is divided into interactive load and batch processing load;

[0018] The interactive load has the highest priority;

[0019] The batch processing workload includes DT1 type workload and DT2 type workload;

[0020] Both the DT1 and DT2 loads have time-transfer characteristics, and the DT2 load also has a scalability feature.

[0021] Preferably, the heat exchange principle of the equipment in the data center is as follows:

[0022] Data center server rooms use cold or hot aisle isolation for heat dissipation;

[0023] The variable-speed fans at the bottom of the server room compress cold air into the sealed cold aisle through the air intake louvers of the raised floor. The filters, humidifiers and coolers installed in the cold aisle ensure that the air parameters remain constant. The fans of the IT equipment draw in cold air to cool the equipment, and after forming hot air, it is discharged from the rear of the rack to the hot aisle.

[0024] Preferably, the convective heat transfer process of the data center IT equipment can be described using an equivalent thermal resistance model, satisfying the following relationship:

[0025]

[0026] C IT =c p,IT n racks (n servers m servers +m rack )

[0027] In the formula, R in C represents the equivalent thermal resistance of convective heat transfer at the IT equipment inlet. IT For the heat capacity of IT equipment, This represents the temperature of the IT equipment at time t. This represents the power of the IT equipment at time t. c represents the air temperature in the equipment rack at time t. p,IT n represents the average specific heat capacity of IT equipment. racks n is the number of racks. servers m is the number of IT devices servers For the quality of a single IT device, m rack The mass of a single empty rack;

[0028] The air convection heat transfer process in the rack, the air convection heat transfer process in the cold and hot aisles, and the heat transfer process between the cold and hot aisles and the wall surfaces are modeled separately.

[0029]

[0030] C Rack =V freeSpace n racks ρ air c p,air

[0031] In the formula, C Rack For the heat capacity of the air in the rack, m air Where κ is the airflow rate, κ is the heat transfer coefficient, and c is the airflow rate. p,air The specific heat capacity of air, V represents the air temperature in the cold aisle at time t. freeSpace ρ represents the usable space volume within the computer room. air Let be the density of air, and Δt be the time step interval;

[0032]

[0033] C cAisle =V cAisle ρ air c p,air

[0034] In the formula, C cAisle R is the heat capacity of the air in the cold aisle. cold The equivalent thermal resistance for convective heat transfer between the cold aisle and its wall surface. Let V be the inlet airflow temperature of the IT equipment at time t. cAisle This refers to the volume of the cold aisle space.

[0035]

[0036] C hAisle =V hAisle ρ air c p,air

[0037] In the formula, C hAisle R is the heat capacity of the air in the hot passage. hot The equivalent thermal resistance for convective heat transfer between the heat channel and its wall surface. V represents the air temperature in the hot passage at time t. hAisle For the volume of the thermal aisle space;

[0038]

[0039]

[0040] In the formula, Let t be the temperature of the cold aisle wall. Let t be the temperature of the hot aisle wall. Let t be the outdoor temperature at time t.

[0041] Preferably, the establishment of the energy consumption model for IT equipment and air conditioning equipment specifically involves:

[0042] IT equipment power consumption It can be modeled as IT equipment resource utilization. and inlet airflow temperature The function satisfies the following relation:

[0043]

[0044] In the formula, P idle The static power of the IT equipment is given by a0, b0, and c0, which are constants.

[0045] Flexibility in demand response is provided by adjusting the power consumption of air conditioning equipment. It can be modeled as the air temperature in the hot passage. and inlet airflow temperature The function satisfies the following relation:

[0046]

[0047]

[0048] In the formula, Let COP be the power of the air conditioning equipment at time t. HVAC This refers to the coefficient of performance (COP) of air conditioning equipment. A higher COP indicates better energy efficiency. Let t be the air supply temperature of the air conditioner.

[0049] Preferably, the solution to the mixed-integer linear programming problem specifically involves:

[0050] The objective is to minimize the sum of data center operating costs and customer service quality penalty costs, specifically including:

[0051] min(C el +C QoS )

[0052]

[0053]

[0054] In the formula, C el For the total operating cost of the data center, C QoS Customer service quality penalty cost, where T is the length of the entire scheduling cycle, f(t) is the electricity price at time t, and c1 is the penalty coefficient for data center load reduction. Let c1 be the data center load reduction at time t, and c2 be the penalty coefficient for data center load time migration. Let t be the data center load migration amount;

[0055] The comprehensive constraints for constructing a data center include:

[0056] Maximum response time constraint for interactive network load:

[0057]

[0058]

[0059] In the formula, Let t be the time during which the network load arrives and waits to be processed. Let t be the time required to process a unit of network load at time t, μ be the number of network loads that the IT equipment can process per unit time, and D be the maximum latency time to meet the user's quality of service requirements.

[0060] Maximum response time constraint for batch network load:

[0061]

[0062]

[0063] In the formula, Let be the length of the batch load queue of type l at time t. For batch workloads of type l arriving at time t, Let λ be the batch processing load of type l being processed at time t. l This is the processing deadline for batch workloads of type l.

[0064] Air conditioning supply air temperature constraints:

[0065]

[0066] In the formula, These are the upper and lower limits of the air conditioner's air supply temperature and output power, respectively.

[0067] IT equipment temperature constraints:

[0068]

[0069] In the formula, This refers to the upper temperature limit for the normal operation of IT equipment.

[0070] IT equipment inlet airflow temperature range constraints:

[0071]

[0072] In the formula, The lower limit of the import temperature for IT equipment. This is the upper limit for the import temperature of IT equipment.

[0073] Furthermore, a data center demand response optimization device is provided, the device comprising:

[0074] The acquisition module is used to acquire the characteristics of the data center network load, the heat exchange coupling relationship of the equipment in the data center, and the operating characteristics of the data center IT equipment and air conditioning equipment.

[0075] The classification module is used to classify network load based on the characteristics of data center network load.

[0076] The modeling module is used to obtain the data center thermodynamic model and the energy consumption model of the IT equipment and air conditioning equipment based on the heat exchange coupling relationship of the equipment in the data center and the operating characteristics of the data center IT equipment and air conditioning equipment.

[0077] The constraint module is used to construct comprehensive constraints based on the classification of network load, the data center thermodynamic model, and the energy consumption models of IT equipment and air conditioning equipment; wherein, the comprehensive constraints include the maximum response time constraint of network load, the power constraint of air conditioning equipment, the temperature constraint of IT equipment, and the inlet airflow temperature constraint.

[0078] The linearization module is used to linearize the data center demand response optimization problem into a mixed-integer linear programming problem based on comprehensive constraints and with the goal of minimizing the sum of data center operating costs and customer service quality penalty costs.

[0079] The solution module is used to solve mixed-integer linear programming problems to obtain the optimized demand response strategy for the data center.

[0080] In another aspect, an apparatus is provided, including a processor and a memory for storing a processor-executable program, characterized in that when the processor executes the program stored in the memory, it implements the aforementioned data center demand response optimization method.

[0081] In another aspect, a storage medium is provided that stores a program, characterized in that, when the program is executed by a processor, it implements the aforementioned data center demand response optimization method.

[0082] (III) Beneficial Effects

[0083] The beneficial effects of this invention are that, compared with the prior art, it aims to minimize the sum of data center operating costs and customer service quality penalty costs. By establishing a heat exchange coupling relationship between IT equipment and air conditioning equipment, it constructs an overall energy consumption model of the data center and uses optimization methods to achieve a good balance between data center costs and customer service quality. This not only ensures energy saving and consumption reduction under uninterrupted operation of the data center, but also fully utilizes the potential of data center demand response when handling different types of network loads. Attached Figure Description

[0084] Figure 1 This is a flowchart illustrating a data center demand response optimization method according to the present invention.

[0085] Figure 2 This is a schematic diagram of the equipment heat exchange principle of a data center demand response optimization method according to the present invention;

[0086] Figure 3 This is a schematic diagram illustrating the predicted network load in a data center according to an embodiment of the present invention.

[0087] Figure 4 This is a schematic diagram showing the power consumption and supply air temperature of a data center air conditioning device according to an embodiment of the present invention;

[0088] Figure 5 This is a schematic diagram of the power consumption and total power consumption of a data center IT device according to an embodiment of the present invention. Detailed Implementation

[0089] The technical solutions in the embodiments of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0090] Example

[0091] like Figure 1-2 As shown in the figure, an embodiment of the present invention provides a data center demand response optimization method, the method comprising:

[0092] To obtain the characteristics of data center network load, the heat exchange coupling relationship of equipment within the data center, and the operating characteristics of data center IT equipment and air conditioning equipment;

[0093] Classify network load based on the characteristics of data center network load;

[0094] Based on the heat transfer coupling relationship of equipment within the data center, a thermodynamic model of the data center is constructed.

[0095] Based on the operating characteristics of data center IT equipment and air conditioning equipment, establish energy consumption models for IT equipment and air conditioning equipment;

[0096] Based on the classification of network load, the thermodynamic model of data centers, and the energy consumption models of IT equipment and air conditioning equipment, comprehensive constraints are constructed.

[0097] Based on comprehensive constraints and with the objective of minimizing the sum of data center operating costs and customer service quality penalty costs, the data center demand response optimization problem is linearized into a mixed-integer linear programming problem.

[0098] The mixed-integer linear programming problem is solved to obtain the optimized demand response strategy for the data center.

[0099] Furthermore, the comprehensive constraints include network load maximum response time constraints, air conditioning equipment power constraints, IT equipment temperature constraints, and inlet airflow temperature constraints.

[0100] Furthermore, the data center network load is divided into interactive load and batch processing load;

[0101] The interactive load has the highest priority;

[0102] The batch processing workload includes DT1 type workload and DT2 type workload;

[0103] Both the DT1 and DT2 loads have time-transfer characteristics, and the DT2 load also has a scalability feature.

[0104] Furthermore, the heat exchange principle of the equipment within the data center is as follows:

[0105] Data center server rooms use cold or hot aisle isolation for heat dissipation;

[0106] The variable-speed fans at the bottom of the server room compress cold air into the sealed cold aisle through the air intake louvers of the raised floor. The filters, humidifiers and coolers installed in the cold aisle ensure that the air parameters remain constant. The fans of the IT equipment draw in cold air to cool the equipment, and after forming hot air, it is discharged from the rear of the rack to the hot aisle.

[0107] Furthermore, the convective heat transfer process of the data center IT equipment can be described using an equivalent thermal resistance model, satisfying the following relationship:

[0108]

[0109] C IT =c p,IT n racks (n servers m servers +m rack )

[0110] In the formula, R in C represents the equivalent thermal resistance of convective heat transfer at the IT equipment inlet. IT For the heat capacity of IT equipment, This represents the temperature of the IT equipment at time t. This represents the power of the IT equipment at time t. c represents the air temperature in the equipment rack at time t. p,IT n represents the average specific heat capacity of IT equipment. racks n is the number of racks. servers m is the number of IT devices servers For the quality of a single IT device, m rack The mass of a single empty rack;

[0111] The air convection heat transfer process in the rack, the air convection heat transfer process in the cold and hot aisles, and the heat transfer process between the cold and hot aisles and the wall surfaces are modeled separately.

[0112]

[0113] C Rack =V freeSpace nracks ρ air c p,air

[0114] In the formula, C Rack For the heat capacity of the air in the rack, m air Where κ is the airflow rate, κ is the heat transfer coefficient, and c is the airflow rate. p,air The specific heat capacity of air, V represents the air temperature in the cold aisle at time t. freeSpace ρ represents the usable space volume within the computer room. air Let be the density of air, and Δt be the time step interval;

[0115]

[0116] C cAisle =V cAisle ρ air c p,air

[0117] In the formula, C cAisle R is the heat capacity of the air in the cold aisle. cold The equivalent thermal resistance for convective heat transfer between the cold aisle and its wall surface. Let V be the inlet airflow temperature of the IT equipment at time t. cAisle This refers to the volume of the cold aisle space.

[0118]

[0119] C hAisle =V hAisle ρ air c p,air

[0120] In the formula, C hAisle R is the heat capacity of the air in the hot passage. hot The equivalent thermal resistance for convective heat transfer between the heat channel and its wall surface. V represents the air temperature in the hot passage at time t. hAisle For the volume of the thermal aisle space;

[0121]

[0122]

[0123] In the formula, Let t be the temperature of the cold aisle wall. Let t be the temperature of the hot aisle wall. Let t be the outdoor temperature at time t.

[0124] Furthermore, the establishment of the energy consumption model for IT equipment and air conditioning equipment specifically involves:

[0125] IT equipment power consumption It can be modeled as IT equipment resource utilization. and inlet airflow temperature The function satisfies the following relation:

[0126]

[0127] In the formula, P idle The static power of the IT equipment is given by a0, b0, and c0, which are constants.

[0128] Flexibility in demand response is provided by adjusting the power consumption of air conditioning equipment. It can be modeled as the air temperature in the hot passage. and inlet airflow temperature The function satisfies the following relation:

[0129]

[0130]

[0131] In the formula, Let COP be the power of the air conditioning equipment at time t. HVAC This refers to the coefficient of performance (COP) of air conditioning equipment. A higher COP indicates better energy efficiency. Let t be the air supply temperature of the air conditioner.

[0132] Furthermore, the solution to the mixed-integer linear programming problem specifically involves:

[0133] The objective is to minimize the sum of data center operating costs and customer service quality penalty costs, specifically including:

[0134] min(C el +C QoS )

[0135]

[0136]

[0137] In the formula, C el For the total operating cost of the data center, C QoS Customer service quality penalty cost, where T is the length of the entire scheduling cycle, f(t) is the electricity price at time t, and c1 is the penalty coefficient for data center load reduction. Let c1 be the data center load reduction at time t, and c2 be the penalty coefficient for data center load time migration. Let t be the data center load migration amount;

[0138] The comprehensive constraints for constructing a data center include:

[0139] Maximum response time constraint for interactive network load:

[0140]

[0141]

[0142] In the formula, Let t be the time during which the network load arrives and waits to be processed. Let t be the time required to process a unit of network load at time t, μ be the number of network loads that the IT equipment can process per unit time, and D be the maximum latency time to meet the user's quality of service requirements.

[0143] Maximum response time constraint for batch network load:

[0144]

[0145]

[0146] In the formula, Let be the length of the batch load queue of type l at time t. For batch workloads of type l arriving at time t, Let λ be the batch processing load of type l being processed at time t. l This is the processing deadline for batch workloads of type l.

[0147] Air conditioning supply air temperature constraints:

[0148]

[0149] In the formula, These are the upper and lower limits of the air conditioner's air supply temperature and output power, respectively.

[0150] IT equipment temperature constraints:

[0151]

[0152] In the formula, This refers to the upper temperature limit for the normal operation of IT equipment.

[0153] IT equipment inlet airflow temperature range constraints:

[0154]

[0155] In the formula, The lower limit of the import temperature for IT equipment. This is the upper limit for the import temperature of IT equipment.

[0156] As a specific embodiment of the present invention, a small data center in northern China is selected as an example. The data center is equipped with 10 42U server racks, with each rack housing 8 IT devices with a rated power of 300W. The upper limit of the normal operating temperature of the IT devices is 80℃, and the upper and lower limits of the inlet airflow temperature are respectively... and The computer room is also equipped with 5 air conditioners, with upper and lower limits for the air supply temperature of which are respectively... and

[0157] Specifically, such as Figure 3 As shown, the optimization period is 24, i.e., Δt = 1h. Figure 3 This includes the RT, DT1, and DT2 network loads that the data center needs to handle in each time period. The maximum allowed latency for the RT load is 0.2 seconds, for the DT1 load it is 2 hours, and for the DT2 load it is 4 hours. The DT2 load is also scalable.

[0158] Specifically, such as Figure 4 As shown, Figure 4 The curves comparing the power consumption and supply air temperature of air conditioning equipment participating in and not participating in demand response in the data center are presented. It can be seen that during the 9-11 period, which is the period of heavy network load, the power consumption of air conditioning equipment is reduced and the supply air temperature of air conditioning equipment is increased by shifting the load to the low load period or reducing the load. During the 12-15 period, as the outdoor temperature gradually rises, the supply air temperature of air conditioning equipment decreases more slowly due to the cold storage characteristics of the computer room walls. During the 20-22 period, considering the effect of electricity price, the delayed allowable load is migrated to the period of lower electricity price.

[0159] Specifically, such as Figure 5 As shown, by Figure 5 It is evident that the data center effectively reduced the peak-to-valley difference in the power load curve after participating in demand response. By adjusting server resource utilization within an adjustable range, the data center's power load curve became even more stable. At this point, the data center can achieve peak shaving and valley filling of the power load by flexibly adjusting the DT1 and DT2 network loads.

[0160] As shown in Table 1, Table 1 illustrates the data center operating costs, customer service quality penalty costs, and data center power utilization efficiency (PUE) under three scenarios. The specific details of the three scenarios are as follows:

[0161] (1) The data center takes into account the network load latency and reduction characteristics, and the air conditioning supply temperature varies within the specified range;

[0162] (2) The data center does not consider network load latency and reduction characteristics, but the air conditioning supply temperature varies within the specified range;

[0163] (3) The data center takes into account the network load latency and reduction characteristics, but the air conditioning supply temperature is a constant value;

[0164] (4) The data center does not consider network load latency and reduction characteristics, and the air conditioning supply temperature is a constant value.

[0165] Table 1

[0166]

[0167] Table 1 shows that the total cost of the data center under the optimized model in Scenario 1 is the lowest, and the data center's power utilization efficiency is also the best. This demonstrates that data center participation in demand response can effectively reduce data center operating costs and improve power efficiency. Comparing Scenario 1 and Scenario 2, it can be seen that utilizing the network load transfer and reduction characteristics, as well as the building's cooling capacity, allows the air conditioning supply temperature to operate at a higher level, reducing air conditioning energy consumption and improving overall power efficiency. In Scenario 3, the air conditioning supply temperature is set at 22℃. If the air conditioning supply temperature is set too low, it will increase air conditioning energy consumption, increase electricity costs, and reduce power efficiency. Conversely, if the supply temperature is set too high, localized hotspots may occur. Therefore, it is necessary to control the air conditioning supply temperature within an appropriate range.

[0168] The beneficial effects of this invention are that, compared with the prior art, it aims to minimize the sum of data center operating costs and customer service quality penalty costs. By establishing a heat exchange coupling relationship between IT equipment and air conditioning equipment, it constructs an overall energy consumption model of the data center and uses optimization methods to achieve a good balance between data center costs and customer service quality. This not only ensures energy saving and consumption reduction under uninterrupted operation of the data center, but also fully utilizes the potential of data center demand response when handling different types of network loads.

[0169] As another embodiment of the present invention, a data center demand response optimization device is provided, the device comprising:

[0170] The acquisition module is used to acquire the characteristics of the data center network load, the heat exchange coupling relationship of the equipment in the data center, and the operating characteristics of the data center IT equipment and air conditioning equipment.

[0171] The classification module is used to classify network load based on the characteristics of data center network load.

[0172] The modeling module is used to obtain the data center thermodynamic model and the energy consumption model of the IT equipment and air conditioning equipment based on the heat exchange coupling relationship of the equipment in the data center and the operating characteristics of the data center IT equipment and air conditioning equipment.

[0173] The constraint module is used to construct comprehensive constraints based on the classification of network load, the data center thermodynamic model, and the energy consumption models of IT equipment and air conditioning equipment; wherein, the comprehensive constraints include the maximum response time constraint of network load, the power constraint of air conditioning equipment, the temperature constraint of IT equipment, and the inlet airflow temperature constraint.

[0174] The linearization module is used to linearize the data center demand response optimization problem into a mixed-integer linear programming problem based on comprehensive constraints and with the goal of minimizing the sum of data center operating costs and customer service quality penalty costs.

[0175] The solution module is used to solve mixed-integer linear programming problems to obtain the optimized demand response strategy for the data center.

[0176] As another embodiment of the present invention, an apparatus is provided, including a processor and a memory for storing a processor-executable program, characterized in that when the processor executes the program stored in the memory, it implements the above-described data center demand response optimization method.

[0177] In another embodiment of the present invention, a storage medium is provided, storing a program, characterized in that, when the program is executed by a processor, it implements the aforementioned data center demand response optimization method.

[0178] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A data center demand response optimization method, characterized in that, The method includes: To obtain the characteristics of data center network load, the heat exchange coupling relationship of equipment within the data center, and the operating characteristics of data center IT equipment and air conditioning equipment; Classify network load based on the characteristics of data center network load; Based on the heat transfer coupling relationship of equipment within the data center, a thermodynamic model of the data center is constructed. Based on the operating characteristics of data center IT equipment and air conditioning equipment, establish energy consumption models for IT equipment and air conditioning equipment; Based on the classification of network load, the thermodynamic model of data centers, and the energy consumption models of IT equipment and air conditioning equipment, comprehensive constraints are constructed. Based on comprehensive constraints and with the objective of minimizing the sum of data center operating costs and customer service quality penalty costs, the data center demand response optimization problem is linearized into a mixed-integer linear programming problem. Solve the mixed-integer linear programming problem to obtain the optimal demand response strategy for the data center; The convective heat transfer process of the data center IT equipment can be described using an equivalent thermal resistance model, satisfying the following relationship: In the formula, The equivalent thermal resistance for convective heat transfer at the IT equipment inlet. For the heat capacity of IT equipment, express t The temperature of IT equipment at all times express t The power of IT equipment at all times express t The air temperature inside the equipment rack at all times. The average specific heat capacity of IT equipment. For the number of racks, For the number of IT devices, For the quality of a single IT device, The mass of a single empty rack; The air convection heat transfer process in the rack, the air convection heat transfer process in the cold and hot aisles, and the heat transfer process between the cold and hot aisles and the wall surfaces are modeled separately. In the formula, The heat capacity of the air in the frame. For airflow, The heat transfer coefficient is... The specific heat capacity of air, for t The air temperature in the cold aisle at all times. This refers to the available space volume within the computer room. For the density of air, The time step interval; In the formula, The heat capacity of the air in the cold aisle. The equivalent thermal resistance for convective heat transfer between the cold aisle and its wall surface. for t The inlet airflow temperature of the IT equipment at all times. This refers to the volume of the cold aisle space. In the formula, This is the heat capacity of the air in the hot passage. The equivalent thermal resistance for convective heat transfer between the heat channel and its wall surface. for t The air temperature in the hot aisle at all times. For the volume of the thermal aisle space; In the formula, for t Constantly monitor the temperature of the cold aisle walls. for t Constant temperature of the hot aisle wall for t outdoor temperature at all times; The specific steps for establishing the energy consumption model for IT equipment and air conditioning equipment are as follows: IT equipment power consumption It can be modeled as IT equipment resource utilization. and inlet airflow temperature The function satisfies the following relation: In the formula, For IT equipment static power, It is a constant; Flexibility in demand response is provided by adjusting the power consumption of air conditioning equipment. It can be modeled as the air temperature in the hot passage. and inlet airflow temperature The function satisfies the following relation: In the formula, for t Air conditioning equipment power at all times This refers to the coefficient of performance (COP) of air conditioning equipment. A higher COP indicates better energy efficiency. for t Set the air conditioner's air supply temperature at all times; The specific steps for solving the mixed-integer linear programming problem are as follows: The objective is to minimize the sum of data center operating costs and customer service quality penalty costs, specifically including: In the formula, For the total operating cost of the data center, Serving customers with poor service incurs penalties. The length of the entire scheduling cycle. for t Electricity price at any time This is the penalty factor for reducing data center load. for t Real-time data center load reduction This is the penalty coefficient for data center load time migration. for t Real-time data center load migration volume; The comprehensive constraints for constructing a data center include: Maximum response time constraint for interactive network load: In the formula, for t The time it takes for network load to arrive at a given moment to be processed. for t The time required to process a unit of network load at any given moment. This refers to the amount of network load that an IT device can handle per unit of time. The maximum delay time required to meet user service quality requirements; Maximum response time constraint for batch network load: In the formula, for t time l The length of the batch load queue of the type, for t The time has arrived l Type of batch workload, for t Currently being processed l Type of batch workload, for l The processing deadline for batch workloads of this type; Air conditioning supply air temperature constraints: In the formula, , These are the upper and lower limits of the air conditioner's supply air temperature and output power, respectively. IT equipment temperature constraints: In the formula, This refers to the upper temperature limit for the normal operation of IT equipment. IT equipment inlet airflow temperature range constraints: In the formula, The lower limit of the import temperature for IT equipment. This is the upper limit for the import temperature of IT equipment.

2. The data center demand response optimization method according to claim 1, characterized in that, The comprehensive constraints include the maximum response time constraint for network load, the power constraint for air conditioning equipment, the temperature constraint for IT equipment, and the temperature constraint for inlet airflow.

3. The data center demand response optimization method according to claim 1, characterized in that, The data center network load is divided into interactive load and batch processing load; The interactive load has the highest priority; The batch processing workload includes DT1 type workload and DT2 type workload; Both the DT1 and DT2 loads have time-transfer characteristics, and the DT2 load also has a scalability feature.

4. The data center demand response optimization method according to claim 1, characterized in that, The heat exchange principle of the equipment in the data center is as follows: Data center server rooms use cold or hot aisle isolation for heat dissipation; The variable-speed fans at the bottom of the server room compress cold air into the sealed cold aisle through the air intake louvers of the raised floor. The filters, humidifiers and coolers installed in the cold aisle ensure that the air parameters remain constant. The fans of the IT equipment draw in cold air to cool the equipment, and after forming hot air, it is discharged from the rear of the rack to the hot aisle.

5. A data center demand response optimization device, characterized in that, The apparatus for performing the data center demand response optimization method according to any one of claims 1-4, the apparatus comprising: The acquisition module is used to acquire the characteristics of the data center network load, the heat exchange coupling relationship of the equipment in the data center, and the operating characteristics of the data center IT equipment and air conditioning equipment. The classification module is used to classify network load based on the characteristics of data center network load. The modeling module is used to obtain the data center thermodynamic model and the energy consumption model of the IT equipment and air conditioning equipment based on the heat exchange coupling relationship of the equipment in the data center and the operating characteristics of the data center IT equipment and air conditioning equipment. The constraint module is used to construct comprehensive constraints based on the classification of network load, the data center thermodynamic model, and the energy consumption models of IT equipment and air conditioning equipment; wherein, the comprehensive constraints include the maximum response time constraint of network load, the power constraint of air conditioning equipment, the temperature constraint of IT equipment, and the inlet airflow temperature constraint. The linearization module is used to linearize the data center demand response optimization problem into a mixed-integer linear programming problem based on comprehensive constraints and with the goal of minimizing the sum of data center operating costs and customer service quality penalty costs. The solution module is used to solve mixed-integer linear programming problems to obtain the optimized demand response strategy for the data center.

6. An apparatus comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the data center demand response optimization method according to any one of claims 1-4.

7. A storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the data center demand response optimization method according to any one of claims 1-4.

Images