Refrigerating machine room control strategy optimization method and system and network server

Abstract

The invention provides a refrigerating machine room control strategy optimization method and system and a network server. The refrigerating machine room control strategy optimization method comprises the following steps: acquiring operation data and environment data of each device and system of a refrigerating machine room to form a training set; the operation data comprises historical operation data and real-time operation data. Configuring equipment boundary conditions, and establishing a prediction model of the refrigerating machine room. And regularly updating the training set, and updating the prediction model according to the updated training set. And according to the updated prediction model, generating a plurality of optimized operation strategies of the refrigerating machine room, wherein the optimized operation strategies comprise the optimized operation strategy with the optimal energy efficiency, the optimized operation strategy without switching the startup equipment and the optimized operation strategy with the least operation equipment. And the operation strategy of the refrigerating machine room is adjusted according to the optimized operation strategy. According to the refrigerating machine room control strategy optimization method and system and the network server, the operation efficiency and the operation effect of the refrigerating machine room can be improved.

Description

Technical Field

[0001] The present application relates to the field of refrigeration room control technology, and in particular to a refrigeration room control strategy optimization method, system and network service terminal. Background Art

[0002] In the field of refrigeration room control, a series of common pain points exist, such as irrational startup strategies, inappropriate parameter settings, and difficulty updating control strategies in real time. Because traditional room control systems lack any self-learning or optimization capabilities, the main method for improving system energy efficiency currently relies on manually summarizing a series of relatively good startup strategies and set values ​​based on historical data or experience, and then programming these strategies into the controller for execution. This relies heavily on the experience of the commissioning personnel and considers relatively few factors, making it difficult to ensure stable and reliable operation of the refrigeration room at the highest energy efficiency level.

[0003] Therefore, it is necessary to provide an improved refrigeration room control strategy optimization method, system and network server to solve the above problems. Summary of the Invention

[0004] The present application provides a refrigeration room control strategy optimization method, system and network service end for improving the operating efficiency and operating effect of the refrigeration room.

[0005] This application discloses a refrigeration room control strategy optimization method, including:

[0006] Obtaining operating data and environmental data of each device and system in the refrigeration room to form a training set; the operating data includes historical operating data and real-time operating data;

[0007] Configuring equipment boundary conditions and establishing a prediction model for the refrigeration room, the prediction model including a sub-equipment model, a water system model, and an overall system model;

[0008] Regularly updating the training set, and updating the prediction model according to the updated training set;

[0009] Generating several optimized operation strategies for the refrigeration room based on the updated prediction model, wherein the optimized operation strategies include an optimized operation strategy with the best energy efficiency, an optimized operation strategy with no switching of powered-on devices, and an optimized operation strategy with the least running devices;

[0010] The operation strategy of the refrigeration room is adjusted according to the optimized operation strategy.

[0011] Furthermore, the acquisition of operating data and environmental data of each device and system in the refrigeration room to form a training set includes:

[0012] The operation data and the environmental data are obtained, data cleaning and preprocessing are performed, incomplete data and abnormal data are eliminated, and the remaining normal data are classified and packaged to form a training set for the refrigeration room equipment and system.

[0013] Furthermore, the equipment in the refrigeration room includes a host, a cooling water pump and a cooling tower; the sub-equipment model includes a host model, a cooling water pump model and a cooling tower model.

[0014] Furthermore, the operating data includes host operating data, and the host operating data includes evaporator inlet and outlet water temperature, evaporator flow rate, condenser outlet water temperature, condenser flow rate and power;

[0015] The establishing of the prediction model of the host comprises:

[0016] According to the host operation data and the equipment boundary conditions, the cooling capacity, heat rejection, heat balance and COP of the host are obtained;

[0017] Fitting evaporator model, condenser model, compressor model and heat balance model;

[0018] The host model is trained and fitted according to the evaporator model, the condenser model, the compressor model and the heat balance model, and a prediction result of the operating status of the host is output according to the host model.

[0019] Furthermore, the fitting of the evaporator model, the condenser model, the compressor model and the heat balance model includes:

[0020] Predicting the evaporation temperature based on the evaporator inlet and outlet water temperatures and the evaporator flow rate;

[0021] Predicting the condenser outlet water temperature according to the condenser inlet water temperature, the condensing temperature and the condenser flow rate;

[0022] predicting compressor power and pressure ratio according to the cooling capacity, the condensing temperature, and the evaporating temperature;

[0023] predicting evaporation pressure and condensation pressure according to the evaporation temperature and the condensation temperature;

[0024] predicting a heat balance based on the cooling capacity;

[0025] Iteratively predicting the condensing temperature based on the evaporating temperature, the cooling capacity, the condenser outlet water temperature, the heat rejection, and the pressure increase ratio;

[0026] fitting the evaporator model according to the evaporator flow and the cooling capacity;

[0027] fitting the condenser model according to the condenser water inlet temperature, the condensing temperature and the condenser flow rate;

[0028] fitting the compressor model according to the cooling capacity and the pressure increase ratio;

[0029] The heat balance model is fitted according to the condenser flow rate and the refrigeration capacity.

[0030] Furthermore, the operating data also includes the operating frequency of the cooling water pump, the power of the cooling water pump, the cooling water flow rate, the cooling water return temperature, the operating frequency of the cooling tower and the power of the cooling tower; the environmental data includes the outdoor wet-bulb temperature;

[0031] The step of establishing the prediction model for the refrigeration room further includes:

[0032] Fitting the cooling water pump model according to the operating frequency of the cooling water pump, the power of the cooling water pump and the cooling water flow rate, and outputting a total power prediction result of the cooling water pump according to the cooling water pump model;

[0033] The cooling tower model includes a cooling tower fan model and a cooling tower heat dissipation model. The cooling tower fan model is fitted according to the cooling tower operating frequency and the cooling tower power, and a fan power prediction result of the cooling tower is output according to the cooling tower fan model.

[0034] Fitting the cooling tower heat dissipation model according to the outdoor wet-bulb temperature, the cooling water flow rate, the cooling water return temperature, and the cooling tower operating frequency, and outputting prediction results of the cooling water inlet temperature, the heat rejection, and the cooling tower approach temperature according to the cooling tower heat dissipation model;

[0035] The water system model is fitted according to the operating frequency of the cooling water pump, the power of the cooling water pump and the cooling water flow rate, and a prediction result of the critical frequency is output according to the water system model.

[0036] Furthermore, the overall system model includes:

[0037] An input parameter module, used to input the training set data, equipment setting parameters and existing control strategies;

[0038] A training sub-model module is used to train the sub-device model and the water system model, and obtain a set of flow distribution coefficients of the evaporator and the condenser; and

[0039] The prediction module is used to determine whether the sub-device model, the water system model and the overall system model are successfully fitted, and output the prediction result if the fitting is successful, and output the failure reason if the fitting fails.

[0040] Furthermore, the prediction results include:

[0041] Cooling water side, cooling water pump related prediction: input the average frequency of the cooling water pump, and predict the cooling water flow and total power of the cooling water pump under the corresponding control strategy;

[0042] Host and cooling tower related predictions: Iterative calculations are performed on the host and cooling tower sides to predict the cooling water inlet temperature, cooling water outlet temperature and chilled water supply temperature under the corresponding control strategy.

[0043] Furthermore, the predictions related to the host and cooling tower also include:

[0044] Determine whether there is an actual cooling water inlet temperature value;

[0045] If the actual cooling water inlet temperature value does not exist, enter an iterative calculation cycle to calculate the cooling water inlet temperature.

[0046] Furthermore, the iterative calculation loop includes:

[0047] Iteratively update the upper and lower limits of the cooling water inlet temperature;

[0048] Iteratively calculate the cooling water inlet temperature.

[0049] Furthermore, the periodically updating the training set includes:

[0050] The existing training set is classified according to different operating conditions, and the latest and more concentrated data in each operating condition are filtered to form an updated training set.

[0051] Furthermore, the updated prediction model generates several optimized operation strategies for the refrigeration room, including:

[0052] Enter relevant parameters;

[0053] Traverse all power-on combinations, and for each power-on combination, select several frequency setting value combinations to calculate the total power of the refrigeration room corresponding to the power-on combination;

[0054] Obtain the minimum total power corresponding to each of the power-on combinations, and sort all the power-on combinations according to the size of the minimum total power, eliminate the power-on combination corresponding to the largest minimum total power, and forcibly retain the power-on combination in the current operating state and the power-on combination with the least number of startups;

[0055] Continue to traverse all remaining power-on combinations and calculate the total power of the cooling room corresponding to all frequency combinations of each power-on combination;

[0056] Obtaining the startup combination mode and frequency setting combination corresponding to the lowest total power of the refrigeration room, and generating the optimized operation strategy with the best energy efficiency;

[0057] Obtain the frequency setting value combination corresponding to the lowest total power under the current startup combination mode, and generate the optimized operation strategy for the non-switching startup device;

[0058] Obtain a frequency setting value combination corresponding to the lowest total power under the startup combination mode of the least running device, and generate an optimized operation strategy for the least running device.

[0059] Furthermore, the input related parameters include:

[0060] Determine whether the current operating condition is confirmed and whether the host of the refrigeration room is in a shutdown state;

[0061] If the current operating condition is uncertain or the host is in a shutdown state, stop generating the optimized operating strategy;

[0062] If the current operating condition is confirmed and the host is not in a shutdown state, determining whether the operating state of the device is confirmed;

[0063] If the operating state of the equipment is uncertain, input the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply temperature, chilled water return temperature and the equipment boundary conditions;

[0064] If the operating status of the equipment is determined, the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply temperature and chilled water return temperature, the equipment boundary conditions and the actual operating strategy are input.

[0065] This application also discloses a refrigeration room control strategy optimization system, including:

[0066] The acquisition module is used to obtain the operating data and environmental data of each device and system in the refrigeration room to form a training set, and regularly update the training set and update the prediction model based on the updated training set;

[0067] A modeling module, configured to establish a prediction model for the refrigeration room, wherein the prediction model includes a sub-equipment model, a water system model, and an overall system model;

[0068] A prediction module generates several optimized operation strategies for the refrigeration room based on the updated prediction model, wherein the optimized operation strategies include an optimized operation strategy with the best energy efficiency, an optimized operation strategy with no switching of powered-on devices, and an optimized operation strategy with the least running devices;

[0069] An adjustment module is used to adjust the operation strategy of the refrigeration room according to the optimized operation strategy.

[0070] The present application also discloses a network server, comprising: a memory and a processor;

[0071] The memory is used to store computer programs;

[0072] When the processor is used to execute the computer program, it implements the refrigeration room control strategy optimization method as described above.

[0073] The refrigeration room control strategy optimization method, system and network server of the present application obtain the operating data and environmental data of each device and system, establish a prediction model for the refrigeration room, generate an optimized operating strategy for adjustment, and can calculate the energy consumption under any working condition and any operating strategy, breaking through the limitations of traditional manual experience and effectively improving the operating efficiency and operating effect of the refrigeration room control strategy. At the same time, by regularly updating the prediction model, it can adapt to the dynamic changes of equipment performance degradation, load characteristics and environmental parameters, so that the refrigeration room can operate at high energy for a long time. By providing three different optimization operation strategy options with different focuses, namely, optimal energy efficiency, no switching of startup equipment and minimum running equipment, it can meet diverse actual needs.

[0074] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the specification and, together with the description, serve to explain the principles of the specification.

[0076] Figure 1 It is a flow chart of the refrigeration room control strategy optimization method of the present application.

[0077] Figure 2 It is a flowchart of the host prediction model established in this application.

[0078] Figure 3 This is the flow chart of the cooling tower model training in this application.

[0079] Figure 4 This is a flow chart of the present application for obtaining a set of flow distribution coefficients for the evaporator and the condenser.

[0080] Figure 5 This is a flow chart of the relevant predictions for the host and cooling tower of this application.

[0081] Figure 6 This is a flowchart of step S400 of this application.

[0082] Figure 7It is a schematic diagram of the refrigeration room control strategy optimization system of the present application. DETAILED DESCRIPTION

[0083] Here, the technical solutions in the embodiments (or "implementations") of the present application will be clearly and completely described in conjunction with the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements.

[0084] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0085] Next, the embodiments of this specification are described in detail.

[0086] like Figure 1 As shown, the present application provides a method for optimizing a refrigeration room control strategy, comprising the following steps:

[0087] Step S100: Acquire the operating data and environmental data of each device and system in the refrigeration room to form a training set. The operating data includes historical operating data and real-time operating data.

[0088] The equipment in the refrigeration room includes the main unit, cooling water pump and cooling tower. The main unit includes compressor, evaporator and condenser.

[0089] Operational data includes chilled water supply temperature, chilled water return temperature, chilled water flow rate, cooling water supply temperature, cooling water return temperature, cooling water flow rate, the number of cooling water pumps and cooling towers in operation, operating frequency, and power, as well as main unit operating data. Main unit operating data reflects the operating status of each component within the main unit. Specifically, main unit operating data includes evaporator inlet and outlet water temperatures, evaporator flow rate, condenser outlet water temperature, condenser flow rate, and power. Environmental data includes outdoor wet-bulb temperature.

[0090] Because operational and environmental data are mostly collected in refrigeration rooms, data anomalies may occur during data collection and transmission. These include missing data from some devices due to communication anomalies, unsteady data from initial startup or device switching, and significant data anomalies caused by sensor deviations. During data processing, these incomplete and abnormal data must be filtered to prevent them from interfering with subsequent model training or affecting model accuracy.

[0091] Therefore, it is necessary to first obtain operating data and environmental data, perform data cleaning and preprocessing, eliminate incomplete data and abnormal data, and then classify and package the remaining normal data to form a training set for the refrigeration room equipment and system.

[0092] Specifically, in the above process, it is necessary to separate the parameters of each device level, complete the key indicators of each device, record abnormal data, and encapsulate the classified and packaged data into sub-datasets of the corresponding equipment and system. The sub-datasets together form the training set of the refrigeration room equipment and system.

[0093] Step S200: Configure equipment boundary conditions and establish a prediction model for the refrigeration room.

[0094] The equipment boundary conditions are configured based on the host equipment parameters of a specific project, including the compressor type, unit type, upper and lower flow limits of the evaporator and condenser, minimum power percentage, maximum power percentage, pressure drop, etc.

[0095] The prediction model for the refrigeration room includes sub-equipment models, a water system model, and an overall system model. Furthermore, the sub-equipment model includes the main unit model, cooling water pump model, and cooling tower model, while the main unit model includes the compressor model, evaporator model, condenser model, and heat balance model.

[0096] Specifically, if Figure 2 As shown, establishing a host prediction model includes the following steps:

[0097] Step S201: According to the host operation data and the equipment boundary conditions, the cooling capacity, heat rejection, heat balance and COP (Coefficient of Performance) of the host are obtained.

[0098] Step S202: Fitting the evaporator model, the condenser model, the compressor model and the heat balance model.

[0099] As the core equipment in a refrigeration room, the operating status of the main unit is influenced by the synergistic effects of multiple key components. To accurately establish a main unit prediction model, it is necessary to first fit models of the evaporator, condenser, compressor, and heat balance to accurately characterize the operating characteristics of each main unit component and ensure that the main unit model can fully and accurately reflect its actual operating status.

[0100] Specifically, the fitting process predicts the evaporation temperature based on the evaporator inlet and outlet water temperatures and the evaporator flow rate. The condenser outlet water temperature is predicted based on the condenser inlet and condensing temperatures, and the condenser flow rate. The compressor power and pressure ratio are predicted based on the unit's cooling capacity, condensing temperature, and evaporation temperature. The evaporation and condensing pressures are predicted based on the evaporation and condensing temperatures. The heat balance is predicted based on the unit's cooling capacity. The condensing temperature is iteratively predicted based on the evaporation temperature, the unit's cooling capacity, the condenser outlet water temperature, the amount of heat rejected, and the compressor pressure ratio.

[0101] Furthermore, an evaporator model is fitted based on the evaporator flow rate and cooling capacity. A condenser model is fitted based on the condenser inlet water temperature, condensing temperature, and condenser flow rate. A compressor model is fitted based on the cooling capacity of the main unit and the compressor pressure ratio. A heat balance model is fitted based on the condenser flow rate and cooling capacity.

[0102] In this application, the prediction of the condensing temperature is iterative. Since the condensing temperature is affected by a combination of factors such as the evaporation temperature, the cooling capacity of the main unit, the condenser water outlet temperature, the heat rejection, and the compressor pressure ratio, through iterative prediction, the relationship between these factors can be taken into account, and the predicted value of the condensing temperature can be gradually revised to make it closer to the actual situation. At the same time, iterative prediction can continuously update and adjust the predicted results of the condensing temperature based on new data obtained in real time, and promptly reflect the dynamic changes of the system, thereby maintaining a high prediction accuracy under different operating conditions and ensuring good robustness of the model.

[0103] Step S203: training and fitting a host model according to the evaporator model, the condenser model, the compressor model, and the heat balance model, and outputting a prediction result of the host's operating status according to the host model.

[0104] Establishing a prediction model for cooling water pumps includes the following steps:

[0105] According to the cooling water pump operating frequency, cooling water pump power and cooling water flow in the operating data, a cooling water pump model is fitted, and the total power prediction result of the cooling water pump is output according to the cooling water pump model.

[0106] In this application, the prediction model of the cooling tower includes a cooling tower fan model and a cooling tower heat dissipation model.

[0107] According to the cooling tower operating frequency and cooling tower power in the operating data, a cooling tower fan model is fitted, and the cooling tower fan power prediction result is output according to the cooling tower fan model.

[0108] According to the outdoor wet-bulb temperature in the environmental data and the cooling water flow, cooling water return temperature and cooling tower operation frequency in the operation data, the cooling tower heat dissipation model is fitted, and the prediction results of cooling water inlet temperature, heat rejection and cooling tower approach temperature are output according to the cooling tower heat dissipation model.

[0109] Establishing a predictive model for a water system involves the following steps:

[0110] According to the cooling water pump operating frequency, cooling water pump power and cooling water flow in the operating data, a water system model is fitted, and the prediction result of the critical frequency is output according to the water system model.

[0111] The overall system model of the refrigeration room includes an input parameter module, a training sub-model module and a prediction module.

[0112] The input parameter module is used to input training data, device settings, and the current control strategy. The training data input includes at least one of cooling capacity and chilled water return temperature, as well as chilled water flow and outdoor wet-bulb temperature. The input device settings include chilled water supply temperature, average cooling water pump operating frequency, and average cooling tower operating frequency. The current control strategy includes the operating strategy code, startup combination, cooling water pump operating status, cooling tower operating status, and cooling water valve and chilled water valve switching.

[0113] The training sub-model module is used to train the sub-equipment model and the water system model, and obtain the flow distribution coefficient set of the evaporator and condenser.

[0114] When training a host model, first determine whether the host model was built based on the selection report. The selection report contains detailed information such as key parameters, performance curves, and technical specifications required for host modeling. If the host model was built using data from the selection report, then train the established host model. If the host model was built without data from the selection report, then re-build the host model based on the relevant data from the selection report and then train the newly built host model.

[0115] When training the cooling tower model, the influence of temperature drop error needs to be considered and the temperature drop error correction coefficient should be introduced in due time. Figure 3As shown in the figure, if the training set of the cooling tower model is not empty, determine whether the number of cooling towers in operation is 0. If the number of cooling towers in operation is 0, the temperature drop error correction coefficient is the original value, which means that when the cooling tower is not in operation, there may be no actual temperature drop process, or the temperature drop error correction parameter is set according to some default, basic situation in the model. If the number of cooling towers in operation is not 0, determine whether the temperature drop error correction parameter exists. If so, the temperature drop error correction parameter is introduced from the outside to train the cooling tower fan model and the cooling tower heat dissipation model.

[0116] By introducing a temperature drop error correction parameter, we can more accurately simulate the temperature drop process of cooling towers under different operating conditions and with varying numbers of cooling towers, as well as the associated fan operating conditions. This improves the model's ability to predict the overall performance of cooling tower systems, contributing to more efficient and energy-efficient operation and management.

[0117] When training the water system model and cooling water pump model, you must first determine the water system type. Water system types include large parallel systems and one-to-one systems. A large parallel system refers to multiple cooling water pumps connected in parallel to serve multiple hosts, allowing for flexible operation. A one-to-one system refers to each host being assigned its own cooling water pump. If the system type is a one-to-one system, you must first filter out out-of-frequency cooling pump data from the operating data, update the relevant training set, and then train the water system model and cooling water pump model. If the system type is a large parallel model, you can directly train the water system model and cooling water pump model.

[0118] After the cooling water pump model is successfully fitted, the predicted critical frequency and pump combination are output. After the water system model is successfully fitted, if the model's fitting coefficient is greater than 0.9, the critical frequency is calculated repeatedly. Once the critical frequency is found, the critical frequency and pipe network combination are output. The critical frequency refers to the lowest effective frequency during variable frequency operation of the cooling water pump. A frequency below this value can lead to a sharp drop in system efficiency or equipment failure. The pump combination is the optimal combination consisting of the minimum number of operating pumps and their frequencies while meeting current demand. The pipe network combination is the optimal hydraulic distribution solution for different loop combinations achieved by adjusting valve openings or switching pipelines.

[0119] like Figure 4As shown, when obtaining the flow distribution coefficient set of the evaporator and condenser, the existing startup strategy is first traversed to determine whether only a single host is turned on. If so, the flow distribution coefficient is directly set to 1, and all flows are distributed to the host. If it is not a single host running, each running host is continued to be traversed to determine whether the host has its own evaporator and condenser flow meter. If the host has its own flow meter, the flow distribution coefficient is calculated according to the flow ratio. If the host does not have its own flow meter, it is further determined whether there is a pipe network resistance coefficient on the evaporation side and the condensation side. If there is a pipe network resistance coefficient, the coefficient is introduced to calculate the flow distribution coefficient. Specifically, in this embodiment, the pipe network resistance coefficient of the evaporation side and the condensation side is introduced from the outside, and the square root is calculated according to the inverse of the corresponding pipe network resistance coefficient. The flow distribution coefficient is obtained by accumulating the square root values. If there is no pipe network resistance coefficient, the flow is evenly distributed according to the number of hosts, that is, the flow distribution coefficient of each host is equal.

[0120] The prediction module is used to determine whether the sub-equipment model, water system model and overall system model are successfully fitted. If the fitting is successful, the prediction result is output; if the fitting fails, the failure reason is output.

[0121] Specifically, in this application, when checking the model's fit, it is necessary to first check whether the chilled water return temperature is null, or whether the cooling capacity, chilled water supply temperature setpoint, and chilled water main flow rate all exist, and then calculate the chilled water return temperature. Secondly, it is necessary to refer to the above calculation results and check the set of operating hosts, cooling water valve switch, chilled water valve switch, wet-bulb temperature, cooling water pump operation, cooling tower operation, chilled water supply temperature setpoint, average cooling water pump operating frequency, average cooling tower operating frequency, chilled water return temperature, and chilled water flow rate. Based on the above checks, it is determined whether each model has been successfully fitted. If unsuccessful, the failure reason is output as: Model not fitted.

[0122] If the fit is successful, the next step is to check whether the key parameters are correct. The key parameters are the parameters related to the above check contents. If the key parameters are incorrect, the failure reason will be output as: Key parameter error.

[0123] If the key parameters are correct, then check whether there is any code related to the host operation strategy in the training set. If not, the failure reason output is: the host model was not successfully fitted. If the host is not turned on, the host operation strategy code has no integer part, and the decimal part is the number of cooling water pumps in operation. If there is a code related to the host operation strategy in the training set, further determine the type of water system and determine the system operation strategy code in the training set. When the water system type is a large parallel system, the integer part of the system operation strategy code is the host operation strategy code, and the decimal part is the number of cooling water pumps in operation. When the water system type is a one-to-one system, the integer part of the system operation strategy code is the host operation strategy code, and the decimal part is omitted.

[0124] Furthermore, the system checks whether the operation strategy codes for each component of the system in the training set are all valid. If not, the model fitting fails, and the failure reason is output as: "This operation strategy does not exist in the training set." If so, the current operation strategy exists in the training set.

[0125] The prediction results include the relevant predictions of the cooling water side, cooling water pump, main engine and cooling tower.

[0126] In the prediction of the host and cooling tower, it is necessary to perform iterative calculations on the host and cooling tower sides to predict the cooling water inlet temperature, cooling water outlet temperature and chilled water supply temperature under the corresponding control strategy. Specifically, Figure 5 As shown, an iterative calculation is performed on the main unit and cooling tower to determine whether an actual cooling water inlet temperature value exists. If so, the actual cooling water inlet temperature value is directly substituted into the predicted result. If not, an iterative calculation loop is entered to try to calculate the cooling water inlet temperature.

[0127] The specific steps of the iterative calculation loop are as follows: set the initial iteration value of the cooling water inlet temperature to the outdoor wet-bulb temperature, add 0.1 to the previous calculation value in each loop, substitute the numerical prediction results related to the host and cooling tower, and traverse the training sets of various devices and systems at the same time to calculate the corresponding cooling water outlet temperature and chilled water supply temperature.

[0128] When calculating the cooling water outlet temperature, if the cooling water valve is closed, the cooling water outlet temperature is equal to the cooling water inlet temperature. If the cooling water valve is open, the host under the cooling water policy is traversed to check whether the host is not in the set of running hosts or is in a shutdown state. If so, the cooling water outlet temperature is equal to the weighted value of the cooling water inlet temperature multiplied by the condenser flow distribution coefficient. If not, the cooling water outlet temperature is equal to the weighted value of each host's condenser outlet temperature multiplied by the condenser flow distribution coefficient.

[0129] When calculating the chilled water supply temperature, if the chilled water valve is closed, the chilled water supply temperature equals the chilled water return temperature. If the chilled water valve is open, the system traverses the hosts under the chilled water policy to check whether the host is not in the set of running hosts or is in a shutdown state. If so, the chilled water supply temperature equals the weighted sum of the chilled water return temperature and the evaporator flow distribution coefficient. If not, the chilled water supply temperature equals the weighted sum of the evaporator outlet temperature and the evaporator flow distribution coefficient for each host.

[0130] After obtaining the calculated cooling water outlet temperature and chilled water supply temperature, check the cooling water flow rate. If the cooling water flow rate is zero, update the cooling tower data to: the cooling water inlet temperature equals the cooling water outlet temperature, the heat rejection is zero, and the cooling tower temperature drop is zero. Output the cooling tower model prediction results based on the updated training set. If the cooling water flow rate is not zero, output the cooling tower model prediction results based on the input model prediction parameters.

[0131] Next, determine whether the actual cooling water inlet temperature value exists. If so, exit the loop and directly use the actual cooling water inlet temperature value to calculate the predicted result. If not, continue the iterative calculation loop.

[0132] The iterative calculation of the cooling water inlet temperature includes the following two steps:

[0133] ① Update the upper and lower limits of the cooling water inlet temperature and calculate the cooling water inlet temperature:

[0134] When the iterative calculation does not converge, if the cooling water inlet temperature in the predicted cooling tower result is lower than the trial calculation parameter value, the upper limit is updated; otherwise, the lower limit is updated.

[0135] When the main engine is shut down and the lower limit of the cooling water inlet temperature is equal to the wet-bulb temperature and there is no low-pressure alarm, the cycle will be directly exited.

[0136] In the low pressure alarm state, update the lower limit value of the cooling water inlet temperature. If the difference between the upper and lower limits of the cooling water inlet temperature is less than or equal to 0.1, exit the loop.

[0137] When the host is stopped, update the upper limit of the cooling water inlet temperature. If the difference between the upper and lower limits of the cooling water inlet temperature is less than or equal to 0.1, exit the loop.

[0138] When the iterative calculation converges and the calculated cooling water inlet temperature value is less than the lower limit of the cooling water inlet temperature, the actual cooling water inlet temperature is equal to the set lower limit of the cooling water inlet temperature.

[0139] In the normal iterative convergence state, normal iterative calculation is performed until convergence and the loop is exited.

[0140] ②Calculate the cooling water inlet temperature:

[0141] When there is an actual cooling water inlet temperature, the cooling water inlet temperature value is the actual cooling water inlet temperature.

[0142] When the upper limit of the cooling water inlet temperature is empty, the cooling water inlet temperature is calculated and added by 1 in a single iteration. If the cooling water inlet temperature is greater than or equal to 50, the loop is exited.

[0143] Under abnormal circumstances, the difference between the upper and lower limits of the cooling water inlet temperature is extremely small. At this time, the calculation results do not converge and the loop is exited.

[0144] Under normal circumstances, the next round of iterative calculation begins, and the calculated cooling water inlet temperature value is the average value of the upper and lower limits of the cooling water inlet temperature.

[0145] After the above steps, the relevant predictions of the host and cooling tower are realized, and the cooling water inlet temperature, cooling water outlet temperature, chilled water supply temperature, host model prediction results and cooling tower model prediction results are output.

[0146] For cooling water and cooling water pump predictions, the average cooling water pump frequency is required to predict the cooling water flow rate and total cooling water pump power under the corresponding control strategy. Specifically, in general, the average cooling water pump frequency is input into the cooling water pump model to predict the cooling water flow rate and total cooling water pump power. If the decimal portion of the cooling water strategy code in the training set is 0 and the water system type is a large parallel system, or if the cooling water strategy code does not exist in the training set, the cooling water flow rate and total cooling water pump power are both 0.

[0147] After the above steps, relevant predictions of the cooling water side and the cooling water pump are realized, and the cooling water flow rate and the total power of the cooling water pump are output.

[0148] After achieving the above predictions, the total power of each device is then predicted. Specifically, the total cooling tower power is predicted based on the number of operating cooling towers. If the number of operating cooling towers is zero, the total cooling tower power is zero. If the number of operating cooling towers is not zero, the total cooling tower power is predicted using the cooling tower model. Each host is traversed, and if a host is shut down, the host results are recorded. Host results include compressor status, alarm status, and individual machine status. If the host is not shut down, the total host power, total cooling water pump power, and total cooling tower power are recorded.

[0149] In this application, by giving environmental parameters (wet-bulb temperature), constraints (upper and lower limits of host flow, upper and lower limits of host power, lower limit of chilled water pump frequency, lower limit of cooling tower fan frequency, lower limit of cooling water return temperature), target startup strategy and set values ​​(chilled water outlet temperature set value, currently turned on host and cooling water pump combination), etc., the total power and operating parameters of the refrigeration room are predicted through the prediction model. Each prediction model is reliable, has a low degree of dependence on training data, is mainly based on the mechanism model, and is supplemented by the machine learning algorithm. The model error can be controlled within 3%, providing reliable technical support for the safe and efficient operation of the refrigeration room.

[0150] Step S300: regularly update the training set and update the prediction model according to the updated training set.

[0151] Specifically, before updating the prediction model, the historical data in the training set is filtered. The existing training sets for each device model are categorized according to different operating conditions. The most recent and concentrated data for each operating condition is selected to form an updated training set. The prediction model is then updated using this updated training set.

[0152] The prediction model is updated regularly, with a customizable update cycle. The updated training set is used to update the prediction model, allowing it to continuously reflect the actual performance of each device in the refrigeration room (such as performance degradation or subsequent maintenance), making the prediction model's output more realistic and reliable.

[0153] At the same time, the prediction model of this application will also be verified. When a new set of operating data is collected on-site in the refrigeration room, the environmental parameters, startup combinations, equipment setting parameters, etc. in this set of operating data will be input into the prediction model, and the prediction model will be used to output a set of total power values. The total power value output by the prediction model is compared with the total power value collected from the actual operating data. By continuously monitoring these two data, the accuracy of the current prediction model can be determined. If a large error occurs for a long time, the output of the optimization result will be suspended until the error of the prediction model returns to the normal range.

[0154] Step S400: Generate several optimized operation strategies for the refrigeration room based on the updated prediction model.

[0155] In this application, the optimized operation strategy includes the optimized operation strategy with the best energy efficiency, the optimized operation strategy without switching the powered-on devices, and the optimized operation strategy with the least running devices.

[0156] like Figure 6 As shown, step S400 includes the following steps:

[0157] S410. Input relevant parameters.

[0158] The determination of relevant parameters needs to take into account the current operating conditions, host shutdown status and equipment operating status.

[0159] Specifically, determine whether the current operating conditions are certain and whether the host is in a shutdown state. If the current operating conditions are uncertain or the host is in a shutdown state, stop generating the optimized operation strategy. If the current operating conditions are certain and the host is not in a shutdown state, determine whether the device's operating status is certain. If the device's operating status is uncertain, input the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply and return temperatures, and device boundary conditions. If the device's operating status is certain, input the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply and return temperatures, device boundary conditions, and the actual operation strategy.

[0160] S420: traverse all power-on combinations, select several frequency setting value combinations for each power-on combination, and calculate the total power of the refrigeration room corresponding to the power-on combination.

[0161] In this embodiment, each power-on combination mode takes 25 frequency setting values.

[0162] S430. Obtain the lowest total power corresponding to each power-on combination, sort all power-on combinations according to the lowest total power, eliminate the power-on combination corresponding to the largest lowest total power, and forcibly retain the power-on combination in the current operating state and the power-on combination with the least number of power-ons.

[0163] S440: Continue traversing all remaining power-on combinations, and calculate the total power of the refrigeration room corresponding to all frequency combinations of each power-on combination.

[0164] For the remaining power-on combinations after screening, the total power of the refrigeration room corresponding to all possible frequency combinations under each power-on combination is calculated in detail with 1 Hz as the interval, and the energy consumption of different combinations is further accurately analyzed.

[0165] S450: Obtain the startup combination and frequency setting combination corresponding to the lowest total power of the refrigeration room, and generate an optimized operation strategy with the best energy efficiency. This strategy can minimize the energy consumption of the refrigeration room.

[0166] S460: Obtain the frequency setting value combination corresponding to the lowest total power under the current startup combination mode, and generate an optimized operation strategy that does not switch the startup device. This strategy is suitable for scenarios where frequent device switching is undesirable, and can reduce equipment wear and maintenance costs.

[0167] S470: Obtain the frequency setting value combination corresponding to the lowest total power when the least running device is powered on, and generate an optimized operation strategy for the least running device. This strategy can reduce the number of running devices, further reducing energy consumption and device loss.

[0168] By generating a variety of different types of optimized operation strategies, the needs of different users in different scenarios can be met. The refrigeration room provides an intelligent operation solution that is both economical, reliable and low-carbon, improving the flexibility and adaptability of the refrigeration room operation.

[0169] Step S500: Adjust the operation strategy of the refrigeration room according to the optimized operation strategy.

[0170] In some cases, the number of powered-on devices, the combination of powered-on devices, the operating frequency of the devices, etc. can be dynamically adjusted according to the optimized operation strategy to make the refrigeration room operate in an ideal state.

[0171] like Figure 7 As shown, the present application also provides a refrigeration room control strategy optimization system, including:

[0172] The modeling module is used to establish a prediction model for the refrigeration room. The prediction model includes a sub-equipment model, a water system model and an overall system model.

[0173] The acquisition module is used to obtain the operating data and environmental data of each device and system in the refrigeration room to form a training set, and regularly update the training set and update the prediction model based on the updated training set.

[0174] The prediction module generates several optimized operation strategies for the refrigeration room based on the updated prediction model. The optimized operation strategies include the optimal operation strategy with the best energy efficiency, the optimal operation strategy without switching on the equipment, and the optimal operation strategy with the least running equipment.

[0175] The adjustment module is used to adjust the operation strategy of the refrigeration room according to the optimized operation strategy.

[0176] The present application also provides a network server, comprising a memory and a processor;

[0177] Memory is used to store computer programs;

[0178] When the processor is used to execute the computer program, it implements the refrigeration room control strategy optimization method as described above.

[0179] The refrigeration room control strategy optimization method, system and network server of the present application obtain the operating data and environmental data of each device and system, establish a prediction model for the refrigeration room, generate an optimized operating strategy for adjustment, and can calculate the energy consumption under any working condition and any operating strategy, breaking through the limitations of traditional manual experience and effectively improving the operating efficiency and operating effect of the refrigeration room control strategy. At the same time, by regularly updating the prediction model, it can adapt to the dynamic changes of equipment performance degradation, load characteristics and environmental parameters, so that the refrigeration room can operate at high energy for a long time. By providing three different optimization operation strategy options with different focuses, namely, optimal energy efficiency, no switching of startup equipment and minimum running equipment, it can meet diverse actual needs.

[0180] It should be noted that the technical solutions or technical features described in the above embodiments can be combined or supplemented with each other without conflict. The scope of protection of this application is not limited to the precise structures described in the above embodiments and shown in the accompanying drawings; all modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of this application shall be included in the scope of protection of this application.

Claims

1. A method for optimizing a refrigeration room control strategy, characterized in that: include: Obtaining operating data and environmental data of each device and system in the refrigeration room to form a training set; the operating data includes historical operating data and real-time operating data; Configuring equipment boundary conditions and establishing a prediction model for the refrigeration room, the prediction model including a sub-equipment model, a water system model, and an overall system model; Regularly updating the training set, and updating the prediction model according to the updated training set; Generating several optimized operation strategies for the refrigeration room based on the updated prediction model, wherein the optimized operation strategies include an optimized operation strategy with the best energy efficiency, an optimized operation strategy with no switching of powered-on devices, and an optimized operation strategy with the least running devices; The operation strategy of the refrigeration room is adjusted according to the optimized operation strategy.

2. The refrigeration room control strategy optimization method according to claim 1, characterized in that: The acquisition of operating data and environmental data of each device and system in the refrigeration room to form a training set includes: The operation data and the environmental data are obtained, data cleaning and preprocessing are performed, incomplete data and abnormal data are eliminated, and the remaining normal data are classified and packaged to form a training set for the refrigeration room equipment and system.

3. The refrigeration room control strategy optimization method according to claim 1, characterized in that: The equipment in the refrigeration room includes a host computer, a cooling water pump and a cooling tower; the sub-equipment model includes a host computer model, a cooling water pump model and a cooling tower model.

4. The refrigeration room control strategy optimization method according to claim 3, characterized in that: The operating data includes host operating data, which includes evaporator inlet and outlet water temperature, evaporator flow rate, condenser outlet water temperature, condenser flow rate and power; The establishing of the prediction model of the host comprises: According to the host operation data and the equipment boundary conditions, the cooling capacity, heat rejection, heat balance and COP of the host are obtained; Fitting evaporator model, condenser model, compressor model and heat balance model; The host model is trained and fitted according to the evaporator model, the condenser model, the compressor model and the heat balance model, and a prediction result of the operating status of the host is output according to the host model.

5. The refrigeration room control strategy optimization method according to claim 4, characterized in that: The fitting of the evaporator model, the condenser model, the compressor model and the heat balance model includes: Predicting the evaporation temperature based on the evaporator inlet and outlet water temperatures and the evaporator flow rate; Predicting the condenser outlet water temperature according to the condenser inlet water temperature, the condensing temperature and the condenser flow rate; predicting compressor power and pressure ratio according to the cooling capacity, the condensing temperature, and the evaporating temperature; predicting evaporation pressure and condensation pressure according to the evaporation temperature and the condensation temperature; predicting a heat balance based on the cooling capacity; Iteratively predicting the condensing temperature according to the evaporating temperature, the cooling capacity, the condenser outlet water temperature, the heat rejection, and the pressure increase ratio; fitting the evaporator model according to the evaporator flow and the cooling capacity; fitting the condenser model according to the condenser water inlet temperature, the condensing temperature and the condenser flow rate; fitting the compressor model according to the cooling capacity and the pressure increase ratio; The heat balance model is fitted according to the condenser flow rate and the refrigeration capacity.

6. The refrigeration room control strategy optimization method according to claim 3, characterized in that: The operating data also includes the cooling water pump operating frequency, cooling water pump power, cooling water flow, cooling water return temperature, cooling tower operating frequency and cooling tower power; the environmental data includes outdoor wet bulb temperature; The step of establishing the prediction model for the refrigeration room further includes: Fitting the cooling water pump model according to the operating frequency of the cooling water pump, the power of the cooling water pump and the cooling water flow rate, and outputting a total power prediction result of the cooling water pump according to the cooling water pump model; The cooling tower model includes a cooling tower fan model and a cooling tower heat dissipation model. The cooling tower fan model is fitted according to the cooling tower operating frequency and the cooling tower power, and a fan power prediction result of the cooling tower is output according to the cooling tower fan model. Fitting the cooling tower heat dissipation model according to the outdoor wet-bulb temperature, the cooling water flow rate, the cooling water return temperature, and the cooling tower operating frequency, and outputting prediction results of the cooling water inlet temperature, the heat rejection, and the cooling tower approach temperature according to the cooling tower heat dissipation model; The water system model is fitted according to the operating frequency of the cooling water pump, the power of the cooling water pump and the cooling water flow rate, and a prediction result of the critical frequency is output according to the water system model.

7. The refrigeration room control strategy optimization method according to claim 3, characterized in that: The overall system model includes: An input parameter module, used to input the training set data, equipment setting parameters and existing control strategies; A training sub-model module is used to train the sub-device model and the water system model, and obtain a set of flow distribution coefficients of the evaporator and the condenser; and The prediction module is used to determine whether the sub-device model, the water system model and the overall system model are successfully fitted, and output the prediction result if the fitting is successful, and output the failure reason if the fitting fails.

8. The refrigeration room control strategy optimization method according to claim 7, characterized in that: The prediction results include: Cooling water side, cooling water pump related prediction: input the average frequency of the cooling water pump, and predict the cooling water flow and total power of the cooling water pump under the corresponding control strategy; Host and cooling tower related predictions: Iterative calculations are performed on the host and cooling tower sides to predict the cooling water inlet temperature, cooling water outlet temperature and chilled water supply temperature under the corresponding control strategy.

9. The refrigeration room control strategy optimization method according to claim 8, characterized in that: The above predictions on the main engine and cooling tower also include: Determine whether there is an actual cooling water inlet temperature value; If the actual cooling water inlet temperature value does not exist, enter an iterative calculation cycle to calculate the cooling water inlet temperature.

10. The refrigeration room control strategy optimization method according to claim 9, characterized in that: The iterative calculation loop includes: Iteratively update the upper and lower limits of the cooling water inlet temperature; Iteratively calculate the cooling water inlet temperature.

11. The refrigeration room control strategy optimization method according to claim 1, characterized in that: The regularly updating the training set includes: The existing training set is classified according to different operating conditions, and the latest and more concentrated data in each operating condition are filtered to form an updated training set.

12. The refrigeration room control strategy optimization method according to claim 1, characterized in that: The generating of several optimized operation strategies for the refrigeration room according to the updated prediction model includes: Enter relevant parameters; Traverse all power-on combinations, and for each power-on combination, select several frequency setting value combinations to calculate the total power of the refrigeration room corresponding to the power-on combination; Obtain the minimum total power corresponding to each of the power-on combinations, and sort all the power-on combinations according to the size of the minimum total power, eliminate the power-on combination corresponding to the largest minimum total power, and forcibly retain the power-on combination in the current operating state and the power-on combination with the least number of startups; Continue to traverse all remaining power-on combinations and calculate the total power of the cooling room corresponding to all frequency combinations of each power-on combination; Obtaining the startup combination mode and frequency setting combination corresponding to the lowest total power of the refrigeration room, and generating the optimized operation strategy with the best energy efficiency; Obtain the frequency setting value combination corresponding to the lowest total power under the current startup combination mode, and generate the optimized operation strategy for the non-switching startup device; Obtain a frequency setting value combination corresponding to the lowest total power under the startup combination mode of the least running device, and generate an optimized operation strategy for the least running device.

13. The refrigeration room control strategy optimization method according to claim 12, characterized in that: The input related parameters include: Determine whether the current operating condition is confirmed and whether the host of the refrigeration room is in a shutdown state; If the current operating condition is uncertain or the host is in a shutdown state, stop generating the optimized operating strategy; If the current operating condition is confirmed and the host is not in a shutdown state, determining whether the operating state of the device is confirmed; If the operating state of the equipment is uncertain, input the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply temperature, chilled water return temperature and the equipment boundary conditions; If the operating status of the equipment is determined, the overall system model, outdoor wet-bulb temperature, chilled water flow rate, chilled water supply temperature and chilled water return temperature, the equipment boundary conditions and the actual operating strategy are input.

14. A refrigeration room control strategy optimization system, characterized in that: include: The acquisition module is used to obtain the operating data and environmental data of each device and system in the refrigeration room to form a training set, and regularly update the training set and update the prediction model based on the updated training set; A modeling module, configured to establish a prediction model for the refrigeration room, wherein the prediction model includes a sub-equipment model, a water system model, and an overall system model; A prediction module generates several optimized operation strategies for the refrigeration room based on the updated prediction model, wherein the optimized operation strategies include an optimized operation strategy with the best energy efficiency, an optimized operation strategy with no switching of powered-on devices, and an optimized operation strategy with the least running devices; An adjustment module is used to adjust the operation strategy of the refrigeration room according to the optimized operation strategy.

15. A network server, characterized in that: include: Memory and processor; The memory is used to store computer programs; When the processor is used to execute the computer program, it implements the refrigeration room control strategy optimization method according to any one of claims 1 to 13.

Images