An efficient global optimization control method for computer room

By building energy consumption models and timing prediction and optimizing the control parameter combination, the problem that traditional computer room group control cannot optimize energy consumption as a whole is solved, and the efficient operation and energy saving of the refrigeration machine room are achieved.

CN117930659BActive Publication Date: 2024-11-22WUXI CHAOS ENERGY TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410099529.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-24
Publication Date
2024-11-22
Estimated Expiration
2044-01-24

AI Technical Summary

Technical Problem

Traditional computer room group control cannot consider control strategies from the perspective of overall system energy efficiency, resulting in poor energy consumption control of refrigeration machine room, and individual adjustment of equipment parameters may increase the energy consumption of other equipment, which will increase the overall system energy consumption.

Method used

Using the global optimization control method of efficient computer rooms, the energy consumption model is constructed based on historical operation data, the model parameters are regularly corrected, the cooling load demand is predicted through time sequence, the control parameter combination is traversed, the system energy efficiency value is calculated, and the control parameter set corresponding to the highest energy efficiency value is found. , and the optimal operating state is output through policy security filtering.

Benefits of technology

It has achieved the highest overall energy-efficient operation state of the refrigeration room equipment, optimized energy use and operation costs, ensured that the system operates on a safe and reliable basis, and achieved the goal of energy conservation and carbon reduction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117930659B_ABST
    Figure CN117930659B_ABST
Patent Text Reader

Abstract

The present invention discloses a high-efficiency global optimization control method for a computer room, and belongs to the technical field of energy-saving for refrigeration computer rooms. The method comprises the following steps: constructing a chiller energy consumption model, a chilled water pump group energy consumption model, a cooling water pump group energy consumption model, a cooling tower group energy consumption model, and a cooling tower thermal model based on operating data, and discretizing each target control parameter at equal intervals to form a Cartesian product combination of control parameters. Within a fixed control cycle, the cold load value, outdoor dry-bulb temperature, and outdoor wet-bulb temperature obtained according to load prediction are used as input quantities, and the system energy efficiency value corresponding to the Cartesian product combination of each control parameter is obtained by collaborative calculation in combination with the above-mentioned models, and the control parameter set corresponding to the highest energy efficiency value of the system that meets the calculation constraints is found. After filtering and processing through the policy security rules, the optimal operating state and parameters of each electromechanical equipment in this control cycle are output to achieve safe and efficient operation of the refrigeration system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a high-efficiency global optimization control method for a computer room, and belongs to the technical field of energy saving in a computer room. Background Art

[0002] In public buildings and industrial plants, the energy consumption of central air-conditioning systems accounts for 30-60% of the total energy consumption, and the energy consumption of refrigeration rooms accounts for 65-75% of the central air-conditioning energy consumption. Therefore, exploring efficient refrigeration room operation methods and consuming less energy is of great significance to achieving energy conservation and emission reduction goals.

[0003] The refrigeration system is a highly coupled system with complex parameter dependencies between devices. Adjusting the state or parameters of a device alone to reduce the energy consumption of the device may increase the energy consumption of other devices in the computer room, and ultimately increase the overall energy consumption of the system. Traditional computer room group control only implements device-level control of chillers, water pumps, and cooling towers and simple control strategies, such as adjusting the frequency of water pumps according to temperature / pressure differences or adjusting the number and frequency of cooling tower fans according to the degree of proximity. It is unable to consider the control strategy from the perspective of overall system energy efficiency, which will inevitably reduce the effect of its energy consumption control.

[0004] The core issue of efficient operation control of refrigeration room is to achieve the overall energy-saving and efficient operation of electromechanical equipment with the highest overall energy efficiency of all equipment in the refrigeration room under the premise of meeting the cooling load demand. The relevant control target parameters include the number of chillers, the number of chilled water pumps, the number of cooling water pumps, the number of cooling towers, the set value of chilled water outlet temperature, the frequency of chilled water pumps, the frequency of cooling water pumps, the frequency of cooling towers, etc. Summary of the invention

[0005] In order to solve the above problems, the present invention discloses an efficient global optimization control method for a computer room, and its specific technical solution is as follows:

[0006] An efficient global optimization control method for a computer room comprises the following steps:

[0007] Step S1, based on historical operation data, constructing a chiller energy consumption model, a chilled water pump group energy consumption model, a cooling water pump group energy consumption model, a cooling tower group energy consumption model and a cooling tower thermal model;

[0008] Step S2, as the operation data accumulates, the model parameters of step S1 are regularly corrected based on the difference between the measured value and the model prediction value;

[0009] Step S3: determine the control time period T, and obtain the cooling load demand Q for the next control period based on time series prediction. c ;

[0010] Step S4, discretizing each target control parameter at equal intervals to form a Cartesian product combination of control parameters;

[0011] Step S5: according to the cooling load demand Q c , outdoor dry-bulb temperature, and outdoor wet-bulb temperature as inputs, traverse each Cartesian product combination of control parameters, and combine the chiller energy consumption model, chilled water pump energy consumption model, cooling water pump energy consumption model, cooling tower group energy consumption model, and cooling tower thermal model in step S1 to collaboratively calculate the system energy efficiency value EER corresponding to each combination;

[0012] Step S6, searching for a control parameter set corresponding to the highest energy efficiency value of the system under the premise of satisfying the calculation constraint conditions;

[0013] Step S7, after the control parameter set obtained in step S6 is processed by policy security filtering, the optimal operating state and parameters of each electromechanical device are output;

[0014] Step S8, every fixed control time period T, the calculation of steps S3-S7 is executed once in a loop.

[0015] Furthermore, the energy consumption model expression of the chiller in step S1 is:

[0016]

[0017] Among them, COP i is the energy efficiency value of the i-th unit, PLR is the single unit load rate of the chiller, dT * is the percentage difference between the outlet water temperatures of the two chillers, T * dev is the percentage difference of the outlet water temperature of the evaporator of the chiller, M i is the rated cooling capacity of the i-th unit, M is the sum of the rated cooling capacities of all units, W iche is the energy consumption of the i-th chiller;

[0018]

[0019]

[0020] Furthermore, the energy consumption model of the chiller in step S1 is used to calculate the energy consumption value W of the chiller group under the current traversal operating conditions. che ,

[0021]

[0022] Among them, {a, b, c, …, l} are the target parameters of the chiller energy consumption prediction model.

[0023] Furthermore, the energy consumption model parameters {a, b, c, ..., l} of the chiller are obtained based on the measured operation data of the chiller, and PLR, dT *, T * dev , and then obtained through regression fitting.

[0024] Furthermore, in the energy consumption model of the chilled water pump group in step S1, when a frequency conversion device is installed, the energy consumption of the chilled water pump group is expressed as:

[0025] W chwe =N chw *f chw (Q chw )*f chwc (Q chw ,N chw ,ΔP)

[0026] Where: W chwe Indicates the power of the chilled water pump group; N chw Expressed as the number of chilled water pumps in operation; f chw Expressed as the power of a single chilled water pump; Q chw Expressed as the chilled water flow ratio; f chwx It is represented as the power correction model of the chilled water pump group; ΔP is represented as the chilled water control pressure difference.

[0027] Furthermore, the energy consumption model of the chilled water pump group in step S1 is used to calculate the energy consumption of the chilled water pump group under the current ergodic operating condition; when a frequency conversion device is installed, the energy consumption of the chilled water pump group is expressed as:

[0028]

[0029] Where: {C 1 ,C 2 ,C 3 ,C 4} is the target parameter of the energy consumption prediction model of the chilled water pump group,

[0030] The energy consumption model parameters of the chilled water pump group {C 1 ,C 2 ,C 3 ,C 4}Obtained through regression fitting from the measured operating data of a single water pump.

[0031] Furthermore, the power correction model f of the chilled water pump group chwc ,Based on the measured data of the chilled water pump group as the input sample, the power correction model of the chilled water pump group is trained and constructed. The training tool adopts BP artificial neural network, in which the input layer is Q chw、 N chw、 ΔP, the output layer is the power correction factor of the chilled water pump group;

[0032] The power predicted by the energy consumption model of the chilled water pump group is compared with the actual collected power, and the correction coefficient of the measured chilled water pump group is compared with the predicted value obtained by the power correction model of the chilled water pump group. When the relative error exceeds the predetermined value, the power correction model of the chilled water pump group is corrected.

[0033] Furthermore, in the energy consumption model of the cooling water pump group in step S1, when a frequency conversion device is installed, the energy consumption of the cooling water pump group is expressed as:

[0034] W cwe =N cw *f cw (Q cw )*f cwc (Q cw ,N cw )

[0035] Where: W cwe Indicates the cooling water pump power; Q cw Expressed as cooling water flow ratio, N cw Expressed as the number of cooling water pumps in operation; f cwc It is expressed as the power correction model of the cooling water pump group.

[0036] Furthermore, the energy consumption model of the cooling water pump group in step S1 is used to calculate the energy consumption of the cooling water pump group under the current traversal operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling water pump group is expressed as:

[0037]

[0038] Where: {C 1 ,C 2 ,C 3 ,C 4} is the target parameter of the cooling water pump group energy consumption prediction model,

[0039] The cooling water pump group energy consumption model parameters {C 1 ,C 2 ,C 3 ,C 4}Obtained through regression fitting from the measured operating data of a single cooling water pump.

[0040] Furthermore, the cooling water pump group power correction model f in step S1 cwc Based on the measured data of the cooling water pump group as the input sample, the power correction model of the cooling water pump group is trained and constructed. The training tool adopts BP artificial neural network, in which the input layer is Q cw、 N cwThe output layer is the power correction coefficient of the cooling water pump group. The power predicted by the cooling water pump group energy consumption model is compared with the actual collected power, and the measured cooling water pump group correction coefficient is compared with the predicted value obtained by the cooling water pump group power correction model. When the relative error exceeds the preset value, the cooling water pump group power correction model is corrected.

[0041] Furthermore, in the energy consumption model of the cooling tower group in step S1, when a frequency conversion device is installed, the energy consumption of the cooling tower group is expressed as:

[0042] W cte =N ct *f ct (Q ct )

[0043] Where: W cte Indicates the power of cooling tower group; N ct Expressed as the number of cooling towers in operation; Q ct Expressed as the cooling tower fan flow ratio; f ct (Q cw ) is expressed as the power of a single cooling tower.

[0044] Furthermore, the energy consumption model of the cooling tower group in step S1 is used to calculate the energy consumption of the cooling tower group under the current traversal operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling tower group is expressed as:

[0045]

[0046] Where: {C 1 ,C 2 ,C 3 ,C 4} is the target parameter of the cooling tower group energy consumption prediction model;

[0047] The cooling tower group energy consumption model parameters {C 1 ,C 2 ,C 3 ,C 4}Obtained by regression fitting from the measured operating data of the cooling tower.

[0048] Furthermore, in step S1, the power predicted by the cooling tower group energy consumption model is compared with the actually collected power, and the measured cooling tower group power is compared with the predicted value obtained by the cooling tower energy consumption model. When the relative error exceeds a predetermined value, the target parameters of the cooling tower group energy consumption model are corrected.

[0049] Furthermore, the thermal model of the cooling tower in step S1, when a frequency conversion device is installed, is expressed as:

[0050] Act =f ct (T wb ,T r ,LGRatio)

[0051] Among them, A ct Expressed as cooling tower approximation; T wb is the outdoor wet bulb temperature; T r is the cooling water temperature difference; LGRatio is the water-gas ratio expressed as:

[0052]

[0053] Furthermore, the thermal model of the cooling tower group in step S1 is used to calculate the approximation degree A under the current traversal operating condition. ct , expressed as:

[0054]

[0055] The cooling tower thermal model parameters {Coeff(1), ...Coeff(27)} are obtained by regression analysis of the actual operating data of the cooling tower.

[0056] Furthermore, the predicted approximation of the cooling tower thermal model is compared with the actually collected approximation, and when the relative error exceeds a predetermined value, the parameters of the cooling tower thermal model are corrected.

[0057] Furthermore, in step S3, every time a control cycle T is passed, a traversal optimization calculation and control strategy generation are performed, and the load forecasting is to use the time series forecasting algorithm to obtain the cooling load demand Q of the next control cycle T. c , as an input value for system energy efficiency optimization calculation.

[0058] Furthermore, the target control parameters in step S4 include the number of chillers, the number of chilled water pumps, the number of cooling water pumps, the number of cooling towers, the chilled water outlet temperature setting value, the chilled water pump frequency, the cooling water pump frequency and the cooling tower frequency;

[0059] According to the predicted cooling load demand Q c , traverse each control parameter Cartesian product combination, and calculate W by the chiller energy consumption model, the chilled water pump group energy consumption model, the cooling water pump group energy consumption model, and the cooling tower group energy consumption model. che , W chwe , W cwe , W cte , and then combined with the cooling load Q c The energy efficiency value of the refrigeration room is obtained as EER = Q c / (W che +W cwe +W cte+W chwe ), traverse each combination and calculate to generate the energy efficiency value result set EER of the refrigeration room s ={EER 1 , EER 2 , …, EER m}.

[0060] Furthermore, the collaborative calculation in step S5 specifically includes: coordinating the transfer of target control parameters and calculation parameters between models, and assisting the output of the energy consumption model of the chiller, the energy consumption model of the chilled water pump group, the energy consumption model of the cooling water pump group, and the energy consumption model of the cooling tower group;

[0061] The target control parameters "cooling water pump frequency, number of cooling water pumps, number of cooling tower fans, cooling tower fan frequency" are used as the input of the cooling tower thermal calculation model to obtain the cooling tower approximation degree A ct , further calculate the cooling water outlet temperature T cwo As the input of the chiller energy consumption model, the cooling water outlet temperature is expressed as T cwo =A ct +T wb +T r ;

[0062] The target control parameters "number of chillers" and "chilled water outlet temperature" are combined with T cwo , Q c As the input of the chiller energy consumption model, the chiller energy consumption W is obtained. che ;

[0063] The target control parameters "chilled water pump frequency" and "number of chilled water pumps" are used as the input of the chilled water pump group energy consumption model to obtain the chilled water pump group energy consumption W chwe ;

[0064] The target control parameters "cooling water pump frequency" and "number of cooling water pumps" are used as the input of the cooling water pump group energy consumption model to obtain the cooling water pump group energy consumption W cwe ;

[0065] The target control parameters "cooling tower fan frequency" and "number of cooling tower fans" are used as the input of the cooling tower group energy consumption model to obtain the cooling tower group energy consumption W cte .

[0066] Furthermore, the calculation constraint conditions in step S6 are that the calculation parameters are constrained by objective real physical conditions. When the constraint conditions are not met, the calculation is invalid. The constraint conditions include:

[0067] (1) The target control parameter "number of chillers in operation" should meet the cooling load demand, that is, the sum of the rated cooling capacities of all operating chillers should be greater than or equal to the current cooling load.

[0068] (2) The cooling water flow corresponding to the target control parameters "number of cooling water pumps" and "cooling water pump frequency" should be greater than the minimum flow limit of the chiller to the condenser.

[0069] (3) The intermediate calculation parameter "cooling water inlet temperature of chiller" should be greater than the minimum inlet temperature limit of the chiller condenser, and "cooling water outlet temperature of chiller" should be less than the maximum outlet temperature limit of the condenser.

[0070] (4) The intermediate parameter "Chiller Unit Load Rate PLR" should be greater than the minimum load rate limit of the chiller and less than the maximum load rate limit of the chiller.

[0071] (5) The flow rate ratio of the refrigeration / cooling water pump corresponding to the target control parameter "refrigeration / cooling water pump frequency" should be greater than the minimum flow rate ratio limit of the variable frequency water pump.

[0072] (6) The cooling capacity, cooling water pump flow, cooling water temperature difference and chiller energy consumption parameters should satisfy the thermal balance relationship, and the calculation error should be within 5%. The thermal balance expression is:

[0073]

[0074] Where c is the specific heat capacity, F cw is the cooling water flow rate, and Δt is the cooling water temperature difference.

[0075] Furthermore, the security filtering process in step S6 is based on the calculation result set EER of step 6. s After reverse order, the optimal policy set is determined after policy security filtering rule processing. The policy security filtering rules are as follows:

[0076] 1. When a device fault alarm is detected online, the control parameters corresponding to the faulty device should not be used as the target control parameters; the control parameter combination corresponding to the calculation result set EERs should be excluded if there are parameters related to the faulty device;

[0077] 2. For strategies with similar energy efficiency values, the frequency regulation strategy is preferred, followed by the number of start-stop strategies; similar energy efficiency values ​​refer to energy efficiency values ​​within the range of ±10%;

[0078] 3. For the number of start-stop control strategies, the strategy corresponding to the equipment with low energy consumption is selected first, and the strategy corresponding to the equipment with high energy consumption is selected second; specifically, the cooling tower is started first, followed by the chilled water pump, the cooling water pump, and finally the chiller;

[0079] 4. The target control parameter strategy should comply with the step constraint rule. The step constraint refers to the control parameter adjustment range constraint in each regulation cycle. The adjustment step of each control parameter in each regulation cycle shall not exceed the predetermined value.

[0080] The working principle of the present invention is:

[0081] The present invention is a global optimization control method for an efficient computer room, and is a new method for generating a control strategy for an efficient computer room. Traditional group control uses a single parameter preset value as the control target for PID real-time feedback control, ignoring the overall energy efficiency of the refrigeration room. The present invention proposes a fixed cycle to perform optimization control with the highest overall energy efficiency of the refrigeration room as the target, to achieve overall system regulation.

[0082] Based on the specific cooling load demand and outdoor weather conditions, the present invention adopts forward prediction to find the optimal control strategy. The target control parameter combination is used as input to calculate the corresponding refrigeration room energy efficiency value, and then find the optimal control strategy.

[0083] The present invention establishes energy consumption and thermal models according to the operating data characteristics of the refrigeration room equipment, can predict the energy consumption of the equipment group, and further can obtain the energy efficiency value EER of the refrigeration room.

[0084] The present invention proposes to perform calculation constraint conditions and safety rule filtering processing on the control strategy, which fully considers the boundary conditions of the real world and ensures the safe and stable operation of the refrigeration system.

[0085] The beneficial effects of the present invention are:

[0086] The present invention establishes energy consumption and thermal models for the chillers, refrigeration water pumps, cooling water pumps, cooling towers and other related equipment in each refrigeration room, and optimizes the control room equipment based on the optimal control parameter combination strategy calculated according to the global energy efficiency optimization, so as to ensure the safe and reliable operation of the refrigeration system, thereby achieving the purpose of saving energy and operating costs, and creating an "energy saving and carbon reduction" initiative that meets the national advocacy.

[0087] The present invention continuously optimizes and corrects the energy consumption model of the chiller, the energy consumption model of the refrigerated water pump group, the energy consumption model of the cooling water pump group, the energy consumption model of the cooling tower group, and the cooling tower thermal model, which is beneficial to improving the accuracy of the model and ensuring the energy-saving effect of the control strategy after the system has been running for a long time. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 It is the overall flow chart of the present invention.

[0089] Figure 2 It is a flow chart of optimization calculation within a certain control cycle of the present invention.

[0090] Figure 3 It is a cooperative calculation parameter transfer relationship diagram among the target control parameters, the chiller energy consumption model, the refrigeration water pump group energy consumption model, the cooling water pump group energy consumption model, the cooling tower thermal model, and the cooling tower energy consumption model of the present invention. DETAILED DESCRIPTION

[0091] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0092] The high-efficiency global optimization control method of the computer room of the present invention comprises the following steps:

[0093] Step S1, based on historical operation data training, establish a chiller energy consumption model, a chilled water pump group energy consumption model, a cooling water pump group energy consumption model, a cooling tower group energy consumption model, and a cooling tower thermal model.

[0094] Step S2: As the operation data accumulates, the model parameters of the above step S1 are regularly corrected based on the difference between the measured value and the predicted value.

[0095] Step S3: determine the control time period T, and obtain the cooling load demand Q for the next control period based on the time series prediction model. c .

[0096] Step S4, discretizing each target control parameter at equal intervals to form a Cartesian product combination of the control parameters.

[0097] Step S5: according to the cooling load demand Q c , outdoor dry-bulb temperature, and outdoor wet-bulb temperature are used as inputs, and each Cartesian product combination of control parameters is traversed. The energy consumption model of the chiller, the energy consumption model of the chilled water pump, the energy consumption model of the cooling water pump, the energy consumption model of the cooling tower group, and the thermal model of the cooling tower are combined to collaboratively calculate the system energy efficiency value EER corresponding to each combination.

[0098] Step S6: Under the premise of satisfying the calculation constraints, search for the control parameter set corresponding to the highest energy efficiency value of the system.

[0099] The following is a detailed description of the execution process of each step:

[0100] The energy consumption model of the chiller in step S1 is used to calculate the energy consumption value W of the chiller group under the current traversal operating conditions. che .

[0101]

[0102] As a further embodiment,

[0103]

[0104] Among them, COP i is the energy efficiency value of the i-th unit, PLR is the single unit load rate of the chiller, dT *is the percentage difference between the outlet water temperatures of the two chillers, T dev* is the percentage difference of the outlet water temperature of the evaporator of the chiller, M i is the rated cooling capacity of the i-th unit, M is the sum of the rated cooling capacities of all units, W iche is the energy consumption of the i-th chiller. {a, b, c, …, l} are the target parameters of the chiller energy consumption prediction model.

[0105] Preferably, the energy consumption model parameters {a, b, c, ..., l} of the chiller are obtained based on the measured operation data of the chiller: PLR, dT*, T dev* , and then obtain the model parameters through regression fitting. The measured data structure of the chiller is shown in Table 1 below.

[0106] Table 1

[0107]

[0108] The energy consumption model of the chilled water pump group in step S1 is used to calculate the energy consumption of the chilled water pump group under the current ergodic operating conditions; in the case of adding a frequency conversion device, the energy consumption of the chilled water pump group is expressed as:

[0109] W chwe =N chw *f chw (Q chw )*f chwc (Q chw ,N chw ,ΔP)

[0110] Where: W chwe Indicates the power of the chilled water pump group; N chw Expressed as the number of chilled water pumps in operation; f chw Expressed as the power of a single chilled water pump; Q chw Expressed as the chilled water flow ratio; f chwx It is represented as the power correction model of the chilled water pump group; ΔP is represented as the chilled water control pressure difference.

[0111] As a further embodiment, when a frequency conversion device is installed, the energy consumption of the chilled water pump group is expressed as:

[0112]

[0113] Where: W chwe Indicates the power of the chilled water pump group; Q chw Expressed as chilled water flow ratio; N chw Expressed as the number of running units; f chwc It is represented as the power correction model of the chilled water pump group; ΔP is represented as the chilled water control pressure difference.

[0114] The energy consumption model coefficient of the chilled water pump group {C 1 ,C 2 ,C 3 ,C 4 The model parameters are obtained by regression fitting based on the measured single pump operation data. The measured data structure of the chilled water pump group is shown in Table 2 below.

[0115] Table 2

[0116]

[0117] The power correction function f of the chilled water pump group chwc ,Based on the measured data of the chilled water pump group as the input sample, the power correction model of the chilled water pump group is trained and constructed. The training tool can use the BP artificial neural network, in which the input layer is Q chw、 N chw、 ΔP, the output layer is the power correction factor of the chilled water pump group.

[0118] The power predicted by the energy consumption model of the chilled water pump group is compared with the actual collected power, and the correction coefficient of the measured chilled water pump group is compared with the predicted value obtained by the power correction model of the chilled water pump group. When the relative error exceeds the predetermined value (such as 10%), the power correction model of the chilled water pump group is corrected.

[0119] The energy consumption model of the cooling water pump group is used to calculate the energy consumption of the cooling water pump group under the current ergodic operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling water pump group is expressed as:

[0120] W cwe =N cw *f cw (Q cw )*f cwc (Q cw ,N cw )

[0121] Where: W cwe Indicates the cooling water pump power; Q cw Expressed as cooling water flow ratio, N cw Expressed as the number of cooling water pumps in operation; f cwc It is expressed as the power correction model of the cooling water pump group.

[0122] As a further specific example:

[0123]

[0124] The cooling water pump group energy consumption model coefficient {C 1 ,C 2 ,C 3 ,C 4} Obtained by regression fitting from the measured operating data of a single cooling water pump. The measured data structure of the cooling water pump is shown in the example in Table 3 below.

[0125] Table 3

[0126] Number of cooling water pumps Cooling water flow ratio Cooling water pump power (kW) 1 1 30 1 0.95 25 1 0.9 20 1 0.85 18 … … …

[0127] The cooling water pump group power correction function f cwc , based on the measured data of the cooling water pump group as the input sample, the power correction model of the cooling water pump group is trained and constructed. The training tool can use the BP artificial neural network, in which the input layer is Q cw、 N cw The output layer is the power correction coefficient of the cooling water pump group. The power predicted by the cooling water pump group energy consumption model is compared with the actual collected power, and the measured cooling water pump group correction coefficient is compared with the predicted value obtained by the cooling water pump group power correction model. When the relative error exceeds the predetermined value (such as 10%), the cooling water pump group power correction model is corrected.

[0128] The energy consumption model of the cooling tower group is used to calculate the energy consumption of the cooling tower group under the current traversal operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling tower group is expressed as:

[0129] W cte =N ct *f ct (Q ct )

[0130] Where: W cte Indicates the power of cooling tower group; N ct Expressed as the number of cooling towers in operation; Q ct Expressed as the cooling tower fan flow ratio; f ct (Q cw ) is expressed as the power of a single cooling tower.

[0131] As a more specific example:

[0132]

[0133] {C 1 ,C 2 ,C 3 ,C 4} is the target parameter of the cooling tower group energy consumption prediction model.

[0134] The cooling tower group energy consumption model coefficient {C 1 ,C 2 ,C 3 ,C 4} Obtained by regression analysis of measured operation data. The measured data structure of the cooling tower group is shown in Table 4 below.

[0135] Table 4

[0136] Number of cooling towers Cooling tower fan flow ratio Cooling tower group power (kW) 1 1 15 1 0.95 12 1 0.9 10 1 0.85 9 … … …

[0137] The power predicted by the cooling tower group energy consumption model is compared with the actual collected power, and the measured cooling tower group power is compared with the predicted value obtained by the cooling tower energy consumption model. When the relative error exceeds the predetermined value (such as 10%), the target parameters of the cooling tower group energy consumption model are corrected.

[0138] The thermal model of the cooling tower group is used to calculate the approximation under the current ergodic operating conditions and is expressed as:

[0139] A ct =f ct (T wb ,T r ,LGRatio)

[0140] Among them, A ct Expressed as cooling tower approximation; T wb is the outdoor wet bulb temperature; T r is the cooling water temperature difference; LGRatio is the water-gas ratio expression.

[0141] As a further specific example:

[0142]

[0143] Among them, A ct Expressed as cooling tower approximation; T wb is the outdoor wet bulb temperature; T r is the cooling water temperature difference; LGRatio is the water-gas ratio, which is expressed as {Coeff(1),…,Coeff(27)} are the target parameters of the cooling tower thermal model.

[0144] The cooling tower thermal model coefficients {Coeff(1), ..., Coeff(27)} are obtained by regression analysis of the actual operation data of the cooling tower. The actual data structure of the cooling tower is shown in the example of Table 5 below.

[0145] Table 5

[0146]

[0147] The predicted approximation of the cooling tower thermal model is compared with the actual collected approximation. When the relative error exceeds a predetermined value, the target parameters of the cooling tower thermal model are corrected.

[0148] In step S3, the present invention performs a traversal optimization calculation and control strategy generation every one regulation cycle T. In each regulation cycle, the cooling load and outdoor weather are stable, and the cooling capacity of the refrigeration system meets the cooling load demand.

[0149] As described in step S3, load forecasting is to use a time series forecasting algorithm to obtain the cooling load of the next control cycle as an input value for the system traversal optimization calculation.

[0150] The target control parameters in step S4 include the number of chillers, the number of chilled water pumps, the number of cooling water pumps, the number of cooling towers, the predetermined value of the chilled water outlet temperature, the frequency of the chilled water pumps, the frequency of the cooling water pumps, the frequency of the cooling towers, etc.

[0151] Limited by the objective physical world, the target control parameters are selected according to the actual project. For example, if the chilled water pump of a project is not set to variable frequency, the Cartesian product combination of the control parameters excludes the chilled water pump frequency parameter. The target control parameters have upper and lower limit thresholds. For example, 1≤number of chillers≤total number of chillers.

[0152] The Cartesian product combination of the control parameters in step S4 is to discretize the target control parameters at equal intervals to form a set. Assuming that the i-th target control parameter has n values, this discrete set of control parameters is expressed as {S i1 , S i2 , …, S in Each control parameter set is subjected to Cartesian product operation to form a control parameter Cartesian product combination, expressed as {S 11 , S 21 ,…S x1}, {S 12 , S 22 ,…,S x2},…,{S 1m , S 2c ,…,S xn}.

[0153] The overall calculation process described in step S5 is detailed in the attached Figure 2 According to the predicted cooling load demand, each control parameter Cartesian product combination is traversed, and W is calculated by the energy consumption model of the chiller unit, the energy consumption model of the chilled water pump group, the energy consumption model of the cooling water pump group, and the energy consumption model of the cooling tower group. che , W chwe , W cwe , W cte , and then combined with the cooling load Q c The energy efficiency value of the refrigeration room is obtained as EER = Q c / (W che +W cwe +W cte +W chwe). After traversing each combination and calculating, the energy efficiency value result set EER of the refrigeration room is generated. s ={EER 1 , EER 2 , …, EER m}.

[0154] The step S5 is used for coordinated calculation, which is used to coordinate the transfer of calculation parameters between the models and assist the output of the energy consumption model of each device. Figure 3 The target control parameters "cooling water pump frequency, number of cooling water pumps, number of cooling tower fans, cooling tower fan frequency" are used as the input of the cooling tower thermal calculation model to obtain the cooling tower approximation degree A. ct . Combined with the approximation degree A ct , and further calculate the chiller cooling water outlet temperature T cwo , as the input of the chiller energy consumption model. The cooling water outlet temperature is expressed as T cwo =A ct +T wb +Tr.

[0155] The target control parameters "number of chillers" and "chilled water outlet temperature" are combined with T cwo , Q c As the input of the chiller energy consumption model, the chiller energy consumption W is obtained. che

[0156] The target control parameters "chilled water pump frequency" and "number of chilled water pumps" are used as the input of the chilled water pump group energy consumption model to obtain the chilled water pump group energy consumption W chwe .

[0157] The target control parameters "cooling tower fan frequency" and "number of cooling tower fans" are used as the input of the cooling tower group energy consumption model to obtain the cooling tower group energy consumption W cte .

[0158] The calculation constraint conditions in step S6 are that the calculation parameters are constrained by objective real physical conditions. If the constraint conditions are not met, the calculation is invalid. The constraint conditions are as follows:

[0159] 1. The number of operating chillers in the calculation should meet the cooling load demand, that is, the sum of the rated cooling capacities of all operating chillers should be greater than or equal to the current cooling load.

[0160] 2. The cooling water flow corresponding to the target control parameters "number of cooling water pumps" and "cooling water pump frequency" should be greater than the minimum flow limit of the chiller to the condenser. The minimum flow limit of the chiller to the condenser is determined based on the actual project, such as 70%.

[0161] 3. The calculation intermediate parameter "cooling water inlet temperature of chiller" should be greater than the minimum inlet temperature limit of chiller condenser, and "cooling water outlet temperature of chiller" should be less than the maximum outlet temperature limit of condenser. The minimum inlet temperature limit of chiller condenser is determined according to the actual project, such as 18℃.

[0162] 4. The calculation intermediate parameter "Chiller Unit Load Rate PLR" should be greater than the minimum load rate limit of the chiller and less than the maximum load rate limit of the chiller. The minimum load rate limit of the chiller is determined based on the actual project, such as 40%.

[0163] 5. The flow rate ratio of the refrigeration / cooling water pump corresponding to the target control parameter "refrigeration / cooling water pump frequency" should be greater than the minimum flow rate ratio limit of the variable frequency water pump. The minimum flow rate ratio limit of the variable frequency water pump is determined according to the actual project, such as 30HZ.

[0164] 6. The cooling capacity, cooling water pump flow, cooling water temperature difference and chiller energy consumption parameters should satisfy the thermal balance relationship, and the error should be within 5%. The thermal balance expression is:

[0165]

[0166] Among them, Qc is the cooling capacity, which is also the cooling load; c is the specific heat capacity, F cw is the cooling water flow rate, and Δt is the cooling water temperature difference.

[0167] The safety filtering process described in step S7 means that the control strategy must ensure the safe and stable operation of the refrigeration room equipment. s Arrange in reverse order, and after processing with safety filtering and screening rules, determine the optimal strategy set that satisfies both safe and efficient operation. The safety filtering and screening rules are as follows: 1. When the equipment fault state is detected online, the control parameters corresponding to the faulty equipment should not be used as the target control parameters; the control parameter combination corresponding to the calculation result set EERs should be excluded if there are parameters related to the faulty equipment;

[0168] 2. For strategies with similar energy efficiency values, the frequency regulation strategy is preferred, followed by the number of start-stop strategies. Specifically, similar energy efficiency values ​​mean that the difference in energy efficiency values ​​is within the range of ±10%.

[0169] 3. For the start-stop control of the number of units, start the equipment with low energy consumption first, and then start the equipment with high energy consumption. For example, start the cooling tower fan first, then the chilled water pump, the cooling water pump, and finally the chiller.

[0170] 4. The target control parameter strategy should comply with the step constraint rule, which refers to the control parameter adjustment range constraint for each regulation cycle. The adjustment step of the number of control parameters shall not exceed the preset value for each regulation cycle, such as the preset value of the number of chillers and water pumps is ±1, and the preset value of cooling tower fans is ±3; the adjustment step of the frequency control parameter shall not exceed the preset value for each cycle, such as ±5HZ. The preset value of the chilled water outlet temperature shall not exceed the preset value for each cycle, such as ±0.5℃.

[0171] The technical means disclosed in the scheme of the present invention are not limited to the technical means disclosed in the above-mentioned technical means, but also include technical schemes composed of any combination of the above technical features.

[0172] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An efficient global optimization control method for a computer room, characterized in that: The following steps are involved: Step S1, based on historical operation data, constructing a chiller energy consumption model, a chilled water pump group energy consumption model, a cooling water pump group energy consumption model, a cooling tower group energy consumption model and a cooling tower thermal model; Step S2, as the operation data accumulates, the model parameters of step S1 are regularly corrected based on the difference between the measured value and the model prediction value; Step S3: determine the control time period T, and obtain the cooling load demand Q for the next control period based on time series prediction. c ; Step S4, discretizing each target control parameter at equal intervals to form a Cartesian product combination of control parameters; Step S5: according to the cooling load demand Q c , outdoor dry-bulb temperature, and outdoor wet-bulb temperature as inputs, traverse each Cartesian product combination of control parameters, and combine the chiller energy consumption model, chilled water pump energy consumption model, cooling water pump energy consumption model, cooling tower group energy consumption model, and cooling tower thermal model in step S1 to collaboratively calculate the system energy efficiency value EER corresponding to each combination; Step S6, searching for a control parameter set corresponding to the highest energy efficiency value of the system under the premise of satisfying the calculation constraint conditions; The calculation constraint condition is that the calculation parameters are constrained by objective real physical conditions. When the constraint condition is not met, the calculation is invalid. The constraint condition includes: (1) The target control parameter "number of chillers in operation" meets the cooling load demand, that is, the sum of the rated cooling capacities of all operating chillers should be greater than or equal to the current cooling load. (2) The cooling water flow corresponding to the target control parameters "number of cooling water pumps" and "cooling water pump frequency" should be greater than the minimum flow limit of the chiller to the condenser. (3) The calculated intermediate parameter "cooling water inlet temperature of chiller" is greater than the minimum inlet temperature limit of the chiller condenser, and the "cooling water outlet temperature of chiller" is less than the maximum outlet temperature limit of the condenser. (4) The calculated intermediate parameter "Chiller Unit Load Rate PLR" is greater than the minimum load rate limit of the chiller and less than the maximum load rate limit of the chiller. (5) The flow rate ratio of the refrigeration / cooling water pump corresponding to the target control parameter "refrigeration / cooling water pump frequency" should be greater than the minimum flow rate ratio limit of the variable frequency water pump. (6) The cooling capacity, cooling water pump flow, cooling water temperature difference and chiller energy consumption parameters should satisfy the thermal balance relationship, and the calculation error should be within 5%. The thermal balance expression is: Where c is the specific heat capacity, F cw is the cooling water flow rate, Δt is the cooling water temperature difference; Step S7, after the control parameter set obtained in step S6 is processed by policy security filtering, the optimal operating state and parameters of each electromechanical device are output; Step S8, every fixed control time period T, the calculation of steps S3-S7 is executed once in a loop.

2. The high-efficiency computer room global optimization control method according to claim 1 is characterized in that: The energy consumption model expression of the chiller in step S1 is: Among them, COP i is the energy efficiency value of the i-th unit, PLR is the single unit load rate of the chiller, dT * is the percentage difference between the outlet water temperatures of the two chillers, T * dev is the percentage difference of the outlet water temperature of the evaporator of the chiller, M i is the rated cooling capacity of the i-th unit, M is the sum of the rated cooling capacities of all units, W iche is the energy consumption of the i-th chiller; 3. The efficient global optimization control method for a computer room according to claim 1 or 2, characterized in that: The energy consumption model of the chiller in step S1 is used to calculate the energy consumption value W of the chiller group under the current traversal operating conditions. che , Among them, {a, b, c, …, l} are the target parameters of the chiller energy consumption prediction model.

4. The high-efficiency computer room global optimization control method according to claim 3 is characterized in that: The chiller energy consumption prediction model parameters {a, b, c, ..., l} are calculated based on the measured operation data of the chiller to obtain PLR, dT * , T * dev , and then obtained through regression fitting.

5. The efficient global optimization control method for a computer room according to claim 1 is characterized in that: The energy consumption model of the chilled water pump group in step S1, when a frequency conversion device is installed, the energy consumption of the chilled water pump group is expressed as: W chwe =N chw *f chw (Q chw )*f chwc (Q chw ,N chw ,ΔP) Where: W chwe Indicates the power of the chilled water pump group; N chw Expressed as the number of chilled water pumps in operation; f chw Expressed as the power of a single chilled water pump; Q chw Expressed as the chilled water flow ratio; f chwc It is represented as the power correction model of the chilled water pump group; ΔP is represented as the chilled water control pressure difference.

6. The efficient global optimization control method for a computer room according to claim 1 or 5, characterized in that: The energy consumption model of the chilled water pump group in step S1 is used to calculate the energy consumption of the chilled water pump group under the current ergodic operating conditions; when a frequency conversion device is installed, the energy consumption of the chilled water pump group is expressed as: Where: {C1, C2, C3, C4} are the target parameters of the energy consumption prediction model of the chilled water pump group, The energy consumption model parameters {C1, C2, C3, C4} of the chilled water pump group are obtained by regression fitting based on the measured operation data of a single water pump.

7. The efficient global optimization control method for a computer room according to claim 5 is characterized in that: The power correction model f of the chilled water pump group chwc ,Based on the measured data of the chilled water pump group as the input sample, the power correction model of the chilled water pump group is trained and constructed. The training tool adopts BP artificial neural network, in which the input layer is Q chw、 N chw、 ΔP, the output layer is the power correction factor of the chilled water pump group; The power predicted by the energy consumption model of the chilled water pump group is compared with the actual collected power, and the correction coefficient of the measured chilled water pump group is compared with the predicted value obtained by the power correction model of the chilled water pump group. When the relative error exceeds the predetermined value, the power correction model of the chilled water pump group is corrected.

8. The efficient global optimization control method for a computer room according to claim 1 is characterized in that: The energy consumption model of the cooling water pump group in step S1, when a frequency conversion device is installed, the energy consumption of the cooling water pump group is expressed as: W cwe =N cw *f cw (Q cw )*f cwc (Q cw ,N cw ) Where: W cwe Indicates the cooling water pump power; Q cw Expressed as cooling water flow ratio, N cw Expressed as the number of cooling water pumps in operation; f cwc It is expressed as the power correction model of the cooling water pump group.

9. The efficient global optimization control method for a computer room according to claim 1 or 8, characterized in that: The energy consumption model of the cooling water pump group in step S1 is used to calculate the energy consumption of the cooling water pump group under the current traversal operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling water pump group is expressed as: Where: {C1, C2, C3, C4} are the target parameters of the cooling water pump group energy consumption prediction model, The cooling water pump group energy consumption model parameters {C1, C2, C3, C4} are obtained by regression fitting based on the measured operation data of a single cooling water pump.

10. The efficient global optimization control method for a computer room according to claim 8, characterized in that: The cooling water pump group power correction model f cwc Based on the measured data of the cooling water pump group as the input sample, the power correction model of the cooling water pump group is trained and constructed. The training tool adopts BP artificial neural network, in which the input layer is Q cw、 N cw The output layer is the power correction coefficient of the cooling water pump group. The power predicted by the cooling water pump group energy consumption model is compared with the actual collected power, and the measured cooling water pump group correction coefficient is compared with the predicted value obtained by the cooling water pump group power correction model. When the relative error exceeds the preset value, the cooling water pump group power correction model is corrected.

11. The efficient global optimization control method for a computer room according to claim 1, characterized in that: The energy consumption model of the cooling tower group in step S1, when a frequency conversion device is installed, the energy consumption of the cooling tower group is expressed as: W cte =N ct *f ct (Q ct ) Where: W cte Indicates the power of cooling tower group; N ct Expressed as the number of cooling towers in operation; Q ct Expressed as the cooling tower fan flow ratio; f ct (Q cw ) is expressed as the power of a single cooling tower.

12. The efficient global optimization control method for a computer room according to claim 1 or 11, characterized in that: The energy consumption model of the cooling tower group in step S1 is used to calculate the energy consumption of the cooling tower group under the current traversal operating conditions; when a frequency conversion device is installed, the energy consumption of the cooling tower group is expressed as: Where: {C1, C2, C3, C4} are the target parameters of the cooling tower group energy consumption prediction model; The cooling tower group energy consumption model parameters {C1, C2, C3, C4} are obtained by regression fitting based on the actual operation data of the cooling towers.

13. The efficient global optimization control method for a computer room according to claim 1, characterized in that: In step S1, the power predicted by the cooling tower group energy consumption model is compared with the actually collected power, and the measured cooling tower group power is compared with the predicted value obtained by the cooling tower energy consumption model. When the relative error exceeds a predetermined value, the target parameters of the cooling tower group energy consumption model are corrected.

14. The efficient global optimization control method for a computer room according to claim 1, characterized in that: The thermal model of the cooling tower in step S1, when a frequency conversion device is installed, is expressed as: A ct =f ct (T wb ,T r ,LGRatio) Among them, A ct Expressed as cooling tower approximation; T wb is the outdoor wet bulb temperature; T r is the cooling water temperature difference; LGRatio is the water-gas ratio expressed as:

15. The efficient global optimization control method for a computer room according to claim 1 or 14, characterized in that: The thermal model of the cooling tower group in step S1 is used to calculate the approximation degree A under the current traversal operating condition. ct , expressed as: The cooling tower thermal model parameters {Coeff(1), ...Coeff(27)} are obtained by regression analysis of the actual operating data of the cooling tower.

16. The high-efficiency computer room global optimization control method according to claim 15, characterized in that: The predicted approximation of the cooling tower thermal model is compared with the actual collected approximation. When the relative error exceeds the predetermined value, the parameters of the cooling tower thermal model are corrected.

17. The efficient global optimization control method for a computer room according to claim 1, characterized in that: In step S3, a traversal optimization calculation and control strategy generation are performed every time a control cycle T is passed. The load forecast is to use a time series forecasting algorithm to obtain the cooling load demand Q for the next control cycle T. c , as an input value for system energy efficiency optimization calculation.

18. The efficient global optimization control method for a computer room according to claim 1, characterized in that: The target control parameters in step S4 include the number of chillers, the number of chilled water pumps, the number of cooling water pumps, the number of cooling towers, the chilled water outlet temperature setting value, the chilled water pump frequency, the cooling water pump frequency and the cooling tower frequency; According to the predicted cooling load demand Q c , traverse each control parameter Cartesian product combination, and calculate W by the chiller energy consumption model, the chilled water pump group energy consumption model, the cooling water pump group energy consumption model, and the cooling tower group energy consumption model. che , W chwe , W cwe , W cte , and then combined with the cooling load Q c The energy efficiency value of the refrigeration room is obtained as EER = Q c / (W che +W cwe +W cte +W chwe ), traverse each combination and calculate to generate the energy efficiency value result set EER of the refrigeration room s ={EER1, EER2, ..., EER m }.

19. The efficient global optimization control method for a computer room according to claim 1, characterized in that: The collaborative calculation in step S5 specifically includes: coordinating the transfer of target control parameters and calculation parameters between models, and assisting the output of the energy consumption model of the chiller, the energy consumption model of the chilled water pump group, the energy consumption model of the cooling water pump group, and the energy consumption model of the cooling tower group; The target control parameters "cooling water pump frequency, number of cooling water pumps, number of cooling tower fans, cooling tower fan frequency" are used as the input of the cooling tower thermal calculation model to obtain the cooling tower approximation degree A. ct , further calculate the cooling water outlet temperature T cwo As the input of the chiller energy consumption model, the cooling water outlet temperature is expressed as T cwo =A ct +T wb +T r ; The target control parameters "number of chillers" and "chilled water outlet temperature" are combined with T cwo , Q c As the input of the chiller energy consumption model, the chiller energy consumption W is obtained. che ; The target control parameters "chilled water pump frequency" and "number of chilled water pumps" are used as the input of the chilled water pump group energy consumption model to obtain the chilled water pump group energy consumption W chwe ; The target control parameters "cooling water pump frequency" and "number of cooling water pumps" are used as the input of the cooling water pump group energy consumption model to obtain the cooling water pump group energy consumption W cwe ; The target control parameters "cooling tower fan frequency" and "number of cooling tower fans" are used as the input of the cooling tower group energy consumption model to obtain the cooling tower group energy consumption W cte .

20. The efficient global optimization control method for a computer room according to claim 18, characterized in that: Refrigeration room energy efficiency value result set EER s After reverse order, the optimal policy set is determined after policy security filtering rule processing. The policy security filtering rules are as follows: 1). When a device fault alarm is detected online, the control parameters corresponding to the faulty device should not be used as the target control parameters; Refrigeration room energy efficiency value result set EER s The corresponding control parameter combination should be excluded if it contains parameters related to the faulty equipment; 2) For strategies with similar energy efficiency values, the frequency regulation strategy is preferred, followed by the number of start-stop strategies; similar energy efficiency values ​​refer to energy efficiency values ​​within the range of ±10%; 3). For the number of start-stop control strategies, the strategy corresponding to the device with low energy consumption is selected first, followed by the strategy corresponding to the device with high energy consumption; Specifically, the cooling tower is started first, followed by the chilled water pump, the cooling water pump, and finally the chiller; 4) The target control parameter strategy should comply with the step constraint rule. The step constraint refers to the control parameter adjustment range constraint in each regulation cycle. The adjustment step of each control parameter in each regulation cycle shall not exceed a predetermined value.