Energy-saving control method and device for refrigerator

Abstract

The invention provides a refrigerator energy-saving control method and device. The method comprises the steps that multiple pieces of historical operation data of an HVAC system in a target building are obtained; according to each piece of historical operation data, the refrigeration standard reaching duration corresponding to each piece of historical operation data is determined, and the refrigeration standard reaching duration is the duration needed for achieving the refrigeration effect after the refrigerator is started; according to the multiple pieces of historical operation data and the refrigeration standard reaching duration corresponding to each piece of historical operation data, a first prediction model used for predicting the refrigeration standard reaching duration is obtained through training; and utilizing the first prediction model to predict the refrigeration standard reaching duration of the refrigerator on a target date, and determining the starting time of the refrigerator on the target date according to the refrigeration standard reaching duration of the refrigerator on the target date and the opening starting time of the target building. According to the AI refrigerator energy-saving control method and device, AI refrigerator energy-saving control can be achieved on the premise that a simulation model is not established through limited data, and good balance is achieved between energy saving and human body comfort.

Description

Technical Field

[0001] The present application relates to the field of heating, ventilation and air conditioning (HVAC) control technology, and in particular to a chiller energy-saving control method and device. Background Art

[0002] HVAC system accounts for more than 50% of the energy consumption in building operation. The main energy-consuming equipment in HVAC system includes chillers, water pumps, cooling towers and other terminal equipment. Chillers account for more than 60% of the energy consumption in HVAC system. Therefore, reducing the energy consumption of chillers is of great significance for energy conservation and emission reduction.

[0003] The existing technology usually adjusts the opening and closing time of the chiller manually based on industry experience to reduce the power consumption of the chiller. This method is limited by personal experience, takes fewer factors into consideration, and lacks scientific verification. Another way is to use artificial intelligence (AI) technology to achieve energy saving of chiller equipment.

[0004] At present, the existing AI chiller energy-saving solutions are mainly divided into two categories:

[0005] 1. Energy-saving solutions based on physical models mainly use simulation software such as EnergyPlus or EnergyPlus to establish physical models of buildings and HVAC systems. The physical model combines the influence of external weather conditions, internal personnel and lighting conditions, etc. to calculate indoor temperature, humidity and energy consumption changes. On this basis, combined with the chiller operation or equipment setting points generated by the AI ​​algorithm (such as chilled water outlet temperature, water pump frequency, fan coil gear, etc.), the optimization goal between energy consumption and human comfort is achieved. For example, the Chinese patent application with publication number CN114279042 uses this type of strategy. The cost and threshold for establishing the physical model in this solution are relatively high. In addition, the physical model established by simulation is only a theoretical model, and the accuracy of the final result is difficult to verify.

[0006] Second, data-driven energy-saving solutions. The data-driven approach uses sensor data (indoor temperature, chilled water, and electric meters) obtained in the real environment to use AI algorithms to fit the impact of chiller operation or equipment setting points on energy consumption and indoor environment, thereby achieving the optimization goal. The disadvantage of this solution is that the data-driven approach has high requirements for data diversity, and it is difficult to obtain a large amount of high-quality data to complete the training of the algorithm in actual applications.

[0007] Therefore, a refrigeration machine energy-saving control scheme is proposed, which can realize AI refrigeration machine energy-saving control through limited data without establishing a simulation model. Summary of the invention

[0008] At least one embodiment of the present application provides a refrigeration machine energy-saving control method and device, which can realize AI refrigeration machine energy-saving control through limited data without establishing a simulation model, thereby improving the practicality of the refrigeration machine energy-saving control solution.

[0009] According to one aspect of the present application, at least one embodiment provides a chiller energy-saving control method, including:

[0010] Acquire multiple historical operation data of the HVAC system in the target building, each of the historical operation data includes one or more of the chiller start time, chiller shut-down time, chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point of the HVAC system in a historical working cycle;

[0011] Determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, wherein the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on;

[0012] According to the plurality of historical operation data and the cooling standard reaching time length corresponding to each historical operation data, a first prediction model for predicting the cooling standard reaching time length is obtained by training;

[0013] The first prediction model is used to predict the cooling time required for the refrigeration machine to meet the cooling standard on the target date, and the start time of the refrigeration machine on the target date is determined based on the cooling time required for the refrigeration machine to meet the cooling standard on the target date and the opening start time of the target building.

[0014] Optionally, according to each piece of historical operation data, the cooling standard reaching time corresponding to each piece of historical operation data is determined, including:

[0015] Based on each historical operation data, the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time is calculated to obtain the chiller supply and return water temperature difference;

[0016] When the supply and return water temperature difference of the chiller is maintained within a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the cooling effect, and the cooling standard reaching time corresponding to the historical operation data is determined based on the chiller start-up time and the chiller achieving the cooling effect time.

[0017] Optionally, the preset temperature range is [T-α, T+α], where T is the target temperature and α is the fluctuation amplitude;

[0018] Wherein, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the multiple copies of historical operation data, or, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the season and time period to which the current historical operation data belongs in the multiple copies of historical operation data; the α is determined according to the fluctuation range of the indoor temperature of the target building while meeting the comfort requirements.

[0019] Optionally, training a first prediction model for predicting the time to reach the target according to the plurality of historical operation data and the time to reach the target corresponding to each piece of historical operation data includes:

[0020] Determine at least one characteristic combination, each characteristic combination includes at least one of the following characteristics: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point;

[0021] Based on each piece of the historical operation data and the cooling standard-reaching duration corresponding to the historical operation data, determine the historical feature data and the cooling standard-reaching duration corresponding to each feature combination;

[0022] Generate a training set and a test set corresponding to each feature combination based on the historical feature data and the cooling standard-reaching time corresponding to the feature combination;

[0023] Based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set;

[0024] The prediction model with the best score is used as the first prediction model, and a first feature combination used by the first prediction model is determined.

[0025] Optionally, the at least one prediction model includes one or more of the following models: a multi-layer perceptron model, a K nearest neighbor (KNN) model, a decision tree model, a gradient boosting model, and a random forest model.

[0026] Optionally, using the first prediction model to predict the duration of the refrigeration machine reaching the cooling standard on the target date includes:

[0027] Obtaining a first feature value of the first feature combination on a target date;

[0028] The first characteristic value is input into the first prediction model to obtain the cooling standard-reaching time of the refrigerator on the target date predicted by the first prediction model.

[0029] Optionally, the above method further includes:

[0030] Determine the refrigeration failure duration corresponding to each piece of historical operation data according to each piece of historical operation data, wherein the refrigeration failure duration is the time required for the refrigeration machine to lose the refrigeration effect after being shut down;

[0031] A second prediction model for predicting the refrigeration failure duration is obtained by training according to the plurality of historical operation data and the refrigeration failure duration corresponding to each historical operation data;

[0032] The second prediction model is used to predict the refrigeration failure duration of the HVAC system on the target date, and the shutdown time of the HVAC system on the target date is determined based on the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

[0033] According to another aspect of the present application, at least one embodiment provides a chiller energy-saving control device, including:

[0034] A first acquisition module is used to acquire multiple historical operation data of the HVAC system in the target building, each of which includes one or more of the following: a chiller start time, a chiller shut-down time, a chilled water supply temperature, a chilled water flow rate, a chilled water return temperature, an outdoor temperature, an indoor humidity, a dry-bulb temperature, a wet-bulb temperature, and a dew point of the HVAC system in a historical working cycle;

[0035] A first determination module is used to determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, and the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on;

[0036] A first training module is used to train a first prediction model for predicting the duration of reaching the cooling standard based on the plurality of historical operation data and the duration of reaching the cooling standard corresponding to each piece of historical operation data;

[0037] The first prediction module is used to use the first prediction model to predict the cooling time of the refrigeration machine on the target date, and determine the start time of the refrigeration machine on the target date based on the cooling time of the refrigeration machine on the target date and the opening start time of the target building.

[0038] Optionally, the first determining module is further used to:

[0039] Based on each historical operation data, the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time is calculated to obtain the chiller supply and return water temperature difference;

[0040] When the supply and return water temperature difference of the chiller is maintained within a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the cooling effect, and the cooling standard reaching time corresponding to the historical operation data is determined based on the chiller start-up time and the chiller achieving the cooling effect time.

[0041] Optionally, the first training module is further used to:

[0042] Determine at least one characteristic combination, each characteristic combination includes at least one of the following characteristics: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point;

[0043] Based on each piece of the historical operation data and the cooling standard-reaching duration corresponding to the historical operation data, determine the historical feature data and the cooling standard-reaching duration corresponding to each feature combination;

[0044] Generate a training set and a test set corresponding to each feature combination based on the historical feature data and the cooling standard-reaching time corresponding to the feature combination;

[0045] Based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set;

[0046] The prediction model with the best score is used as the first prediction model, and a first feature combination used by the first prediction model is determined.

[0047] Optionally, the first prediction module is further used to:

[0048] Obtaining a first feature value of the first feature combination on a target date;

[0049] The first characteristic value is input into the first prediction model to obtain the cooling standard-reaching time of the refrigerator on the target date predicted by the first prediction model.

[0050] Optionally, the above device further comprises:

[0051] A second determination module is used to determine, based on each piece of historical operation data, a refrigeration failure duration corresponding to each piece of historical operation data, wherein the refrigeration failure duration is the time required for the refrigeration machine to lose its refrigeration effect after being shut down;

[0052] A second training module is used to train a second prediction model for predicting the refrigeration failure duration according to the multiple historical operation data and the refrigeration failure duration corresponding to each historical operation data;

[0053] The second prediction module is used to predict the refrigeration failure duration of the HVAC system on the target date by using the second prediction model, and determine the closing time of the HVAC system on the target date according to the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

[0054] According to another aspect of the present application, at least one embodiment provides a computer-readable storage medium, on which a program is stored. When the program is executed by a processor, the steps of the method described above are implemented.

[0055] Compared with the prior art, the refrigeration machine energy-saving control method and device provided in the embodiment of the present application have low requirements for data diversity, and can realize AI refrigeration machine energy-saving control through limited data without establishing a simulation model, thereby achieving a better balance between energy saving and human comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0057] Figure 1 A schematic diagram of a flow chart of a chiller energy-saving control method according to an embodiment of the present application;

[0058] Figure 2 A schematic diagram showing the time comparison between reaching the absolute temperature target and reaching the stable low fluctuation target;

[0059] Figure 3 An example diagram for the chiller supply and return water temperature difference to meet the standard;

[0060] Figure 4 A schematic diagram of a structure of a chiller energy-saving control device according to an embodiment of the present application;

[0061] Figure 5 Another structural schematic diagram of the energy-saving control device for a refrigerator according to an embodiment of the present application;

[0062] Figure 6 This is another structural schematic diagram of the refrigeration machine energy-saving control device according to an embodiment of the present application. DETAILED DESCRIPTION

[0063] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0064] It should be understood that the "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in one embodiment" appearing throughout the specification may not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. "And / or" in the specification and claims represents at least one of the connected objects.

[0065] In the various embodiments of the present application, it should be understood that the size of the serial numbers of the following processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0066] The following description provides examples and does not limit the scope, applicability, or configuration set forth in the claims. Changes may be made to the functions and arrangements of the elements discussed without departing from the spirit and scope of the present disclosure. Various examples may appropriately omit, replace, or add various procedures or components. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.

[0067] The energy-saving control method for a chiller provided in the embodiment of the present application can achieve the purpose of energy saving without affecting the comfort of people in the building by optimizing the time points of starting and / or shutting down the chiller. Figure 1 , the method comprises the following steps:

[0068] Step 101, obtaining multiple historical operation data of the HVAC system in the target building, each of the historical operation data includes one or more of the chiller start time, chiller shut-down time, chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point of the HVAC system in a historical working cycle.

[0069] Here, the target building can be various buildings such as shopping malls, office buildings, residential buildings, etc., or it can be one or some rooms in a building. The embodiment of the present application obtains multiple historical operation data of the HVAC system in the target building, and each historical operation data corresponds to a historical working cycle of the HVAC system. Each working cycle is a period of time before the cold machine is turned on to a period of time after the cold machine is turned off. For example, assuming that the operating data of the HVAC system in the target building is periodically collected at a predetermined sampling period, the several collection cycles before the cold machine is turned on can be used as the starting time of the working cycle, and the time between several collection cycles after the next cold machine is turned off can be used as a working cycle. In other words, each working cycle includes one cold machine turning on and one cold machine turning off action.

[0070] The start time or the shut down time of the refrigerator can be obtained based on the working condition data of the refrigerator. In actual situations, this part of the data may be difficult to obtain. The embodiment of the present application can indirectly obtain the start or shut down time of the refrigerator based on the energy consumption data of the refrigerator. For example, the energy consumption data of the refrigerator is collected periodically, and the energy consumption difference of adjacent energy consumption data arranged in time is calculated. When the energy consumption difference between the adjacent first energy consumption and the second energy consumption is greater than the preset energy consumption difference threshold (E T ), when the second energy consumption is greater than the first energy consumption and the second energy consumption is after the first energy consumption, the time corresponding to the second energy consumption can be determined as the cooling machine start time T start Similarly, when the energy consumption difference between the adjacent third energy consumption and the fourth energy consumption is greater than the preset energy consumption difference threshold (E T ), when the third energy consumption is greater than the fourth energy consumption and the fourth energy consumption is after the third energy consumption, it can be determined that the time corresponding to the fourth energy consumption is the cooling machine shutdown time T stop .

[0071] In addition, the energy consumption of the chiller will change significantly after the chiller is turned on. Combined with the chiller's chilled water supply and return water temperature difference, the chiller's on / off time (T start , T stop ) and the corresponding energy consumption difference threshold (E T ), the threshold value can be used to extract the startup and shutdown time of the cold machine in the future (when the sampling frequency remains unchanged).

[0072] The data of chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, dew point, etc. for a historical working cycle can be collected according to the preset sampling period.

[0073] For example, for historical weather data (such as outdoor temperature, indoor humidity, dry-bulb temperature, wet-bulb temperature, dew point, etc.), this part of the data can be obtained through the API provided by public channels, or through internal or external sensors where the target building is located. Its fields usually include timestamp, temperature, humidity, radiation intensity, dew point, rainfall, etc. For the required range of historical data, due to the differences in cooling seasons in different regions (for example, the cooling season in higher latitudes is from July to October, while other areas close to the tropics require cooling all year round), usually at least one cooling season or one year of data is required for training.

[0074] For the historical data of the chiller (chilled water supply temperature, chilled water flow, chilled water return temperature, etc.), this part of the data can be composed of various sensors and operation data installed in the target building, including timestamp, chilled water supply temperature, chilled water return temperature, chilled water flow, chiller electricity consumption, chiller power consumption, etc. The time range is consistent with the historical weather data.

[0075] In addition, the embodiment of the present application can also obtain indoor comfort data, which can be used to assist in obtaining the cooling time of the refrigeration unit. Indoor comfort data can be obtained through temperature and humidity sensors deployed in the building, including timestamp, temperature, humidity, etc.

[0076] Step 102, according to each piece of historical operation data, determine the refrigeration standard time corresponding to each piece of historical operation data, wherein the refrigeration standard time is the time required to achieve the refrigeration effect after the refrigerator is turned on.

[0077] Here, the embodiment of the present application determines a corresponding refrigeration time for each historical operation data. Refer to the following refrigeration heat conduction formula:

[0078] q w =c pw xρxV w xΔt

[0079] Among them, q w represents the energy consumption of the chiller, c pw represents the specific heat capacity of water, ρ represents the density of water, V w It represents the water flow rate, and Δt represents the temperature difference between the supply and return water of the chiller.

[0080] In the energy-saving control of chillers, it is usually necessary to balance energy consumption and human comfort, and reduce energy consumption as much as possible while meeting human comfort. Whether human comfort is met can be determined by whether the chiller achieves the preset cooling effect. Therefore, it is necessary to consider how to measure whether the chiller achieves the cooling effect, or whether it meets the target cooling requirements.

[0081] One way to measure the relevant technology is to meet the absolute temperature standard. For example, the chiller can preset the "chilled water supply temperature", and the setting range is generally 7°~14°. The lower the temperature setting, the higher the cooling efficiency, but the overall energy consumption will increase, generally not less than 6°; the higher the temperature setting, the worse the cooling efficiency, but the energy consumption will decrease. Usually, whether the "chilled water supply temperature" reaches the preset temperature is used as the criterion for judging whether the cooling effect meets the standard. For example, the preset value of the "chilled water supply temperature" of the chiller is 8°. If it reaches the preset value 30 minutes after the chiller is turned on, the time for the chiller to meet the cooling effect standard is 30 minutes.

[0082] However, due to the characteristics of the equipment itself, aging, sensor errors and other reasons, in actual scenarios, many refrigerators will take a long time to meet the absolute temperature standard (sometimes they will never meet the standard), and the refrigerator has actually achieved the cooling effect before the absolute stable standard is met. Therefore, the embodiment of the present application determines the cooling time of the refrigerator based on the algorithm of "stable and low fluctuation of the supply and return water temperature difference (ΔT) of the refrigerator" in the following text. Figure 2 The figure shows the time comparison between the absolute temperature reaching the standard and the stable low fluctuation reaching the standard. It can be seen that before the absolute temperature reaches the standard, the stable low fluctuation reaching the standard has been met. At this time, the refrigerator has achieved the predetermined cooling effect. Therefore, the stable low fluctuation reaching the standard can more accurately determine the actual reaching time of the refrigerator refrigeration. Therefore, in the subsequent refrigerator energy-saving control, the startup time of the refrigerator can be delayed, the actual operation time of the refrigerator can be reduced, and better energy-saving purposes can be achieved.

[0083] Specifically, the embodiment of the present application calculates the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time based on each historical operation data, and obtains the supply and return water temperature difference of the chiller. Then, it is determined whether the supply and return water temperature difference of the chiller is maintained in a preset temperature range and reaches a preset time. When the supply and return water temperature difference of the chiller is maintained in a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the refrigeration effect. At this time, the starting time when the supply and return water temperature difference of the chiller is maintained in a preset temperature range and reaches a preset time can be used as the time when the chiller achieves the refrigeration effect. Then, according to the chiller start-up time in the historical operation data and the time when the chiller achieves the refrigeration effect, the refrigeration standard duration corresponding to the historical operation data is determined. In this way, the embodiment of the present application can determine a refrigeration standard duration corresponding to a threshold for each historical operation data.

[0084] Figure 3 An example of the chiller supply and return water temperature difference reaching the standard is shown. Figure 3 Each dot or five-pointed star in represents the data corresponding to a sampling moment. From the sampling moment starting from the five-pointed star, the supply and return water temperature difference of the chiller is in the preset [T-α, T+α] interval, and the supply and return water temperature difference of the chiller is in the above interval for a predetermined time n( Figure 3 When the cooling machine reaches the cooling effect, the cooling machine can be determined to have reached the cooling effect. At this time, the starting time when the cooling machine supply and return water temperature difference is in the above range (such as the time corresponding to the five-pointed star) can be used as the time when the cooling machine reaches the cooling effect. Then, the time between the time when the cooling machine reaches the cooling effect and the time when the cooling machine is turned on is calculated to obtain the cooling standard time.

[0085] Here, the preset temperature range can be expressed as [T-α, T+α], where T is the target temperature of the target building, α is the fluctuation range, and ΔT=2α. Specifically, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the multiple historical operation data, or the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the season and time period to which the current historical operation data belongs in the multiple historical operation data; α is determined according to the fluctuation range of the indoor temperature of the target building under the premise of meeting the preset comfort requirement.

[0086] For example, ΔT is usually between 4° and 6°. The average temperature difference after the chiller is turned on in the historical operation data can be used as the target value, or the average temperature difference in a typical month can be selected as the target value. The fluctuation range of ΔT is related to the cooling load and the chilled water flow rate. Since the cooling load is seasonal and changes at different times of the day, the average value can also be calculated according to the season and time period to obtain the target value T.

[0087] The fluctuation amplitude α is related to the sampling frequency and the characteristics of the refrigeration machine, and can be evaluated in combination with the temperature changes in the building (the time to reach the comfort requirement after startup). Generally speaking, the larger the α, the shorter the time it takes for the refrigeration machine to meet the standard. The preset time (such as the number of consecutive sampling cycles n) is related to the sampling frequency and the characteristics of the refrigeration machine, and can be evaluated in combination with the temperature changes in the building (the time it takes to reach the comfort requirement after startup). Generally speaking, the larger the n value, the shorter the time it takes for the refrigeration machine to meet the standard. The above parameters are hyperparameters of the AI ​​algorithm and can be obtained during the training and testing phases of historical data, which determine the target range of the algorithm's prediction.

[0088] Step 103: training a first prediction model for predicting the time required to reach the cooling target based on the plurality of historical operation data and the time required to reach the cooling target corresponding to each piece of historical operation data.

[0089] Here, the embodiment of the present application determines at least one feature combination, and each feature combination includes at least one of the following features: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point. Here, the initial supply temperature of chilled water is the chilled water supply temperature at the previous time when the chiller is turned on. Then, based on each piece of the historical operating data and the refrigeration compliance time corresponding to the historical operating data, the historical feature data and the refrigeration compliance time corresponding to each feature combination are determined, so that each piece of historical feature data corresponding to each feature combination and its refrigeration compliance time are a sample data, wherein the refrigeration compliance time is used as a label for the sample data. Then, according to the historical feature data and refrigeration compliance time corresponding to each feature combination, a training set and a test set corresponding to the feature combination are generated.

[0090] For example, assuming that a certain feature combination only includes the chilled water supply temperature, then for this feature combination, the temperature value of the chilled water supply temperature in each historical operation data and the corresponding refrigeration compliance time of each historical operation data can be obtained, and the temperature value and refrigeration compliance time can be used as a sample data. In this way, based on multiple historical operation data and their corresponding refrigeration compliance time, multiple sample data can be obtained, and part of the sample data (for example, 70% of the sample data) can be used as a training set, and the remaining part of the data (for example, 30% of the sample data) can be used as a test set.

[0091] In this way, through the above steps, for each feature combination, a training set and a test set corresponding to the feature combination can be obtained.

[0092] Then, based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set. Then, the prediction model with the best score is used as the first prediction model, and the first feature combination used by the first prediction model is determined. Here, the at least one prediction model includes one or more of the following models: a multi-layer perceptron model, a K nearest neighbor (KNN) model, a decision tree model, a gradient boosting model, a random forest model, etc.

[0093] In practical applications, multiple regression algorithms can be used as candidates for prediction algorithm models. In the end, the algorithm with the best test results or the weighted average of multiple algorithms can be selected as the prediction result. For example, if the sample data volume is small, the algorithm that performed best in training may not perform best in future predictions due to overfitting. Multiple algorithms can be selected and weighted to obtain prediction results.

[0094] Therefore, the embodiment of the present application can try multiple prediction models, and use the training set corresponding to each feature combination to train multiple prediction models, so that a prediction model can be obtained for different combinations of feature combinations and prediction models. Then, according to the model score, the optimal prediction model is selected as the first prediction model, and the first feature combination used to train the first prediction model is determined.

[0095] Specifically, when scoring the prediction model, the embodiment of the present application can use at least one scoring algorithm to score each prediction model, and then perform a weighted sum of the scoring results obtained by each scoring algorithm as the final scoring result of the prediction model. For example, the root mean square error (RMSE), R 2 Scoring algorithm is used for scoring.

[0096] Step 104, using the first prediction model, predicting the cooling time of the refrigeration machine on the target date, and determining the start time of the refrigeration machine on the target date based on the cooling time of the refrigeration machine on the target date and the opening start time of the target building.

[0097] Here, the embodiment of the present application can obtain the first characteristic value of the first characteristic combination on the target date. If the target date is a certain day in the future, the first characteristic value can be a predicted value, for example, a predicted value of the outdoor temperature on a certain day in the future. Then, the first characteristic value is input into the first prediction model to obtain the cooling standard duration of the refrigerator on the target date predicted by the first prediction model. Furthermore, the opening start time of the target building is subtracted from the cooling standard duration of the refrigerator on the target date to obtain the start time of the refrigerator on the target date. For example, assuming that the opening start time of the target building is 8:00 in the morning, and the predicted cooling standard duration is 30 minutes, the refrigerator needs to be turned on at 7:30 in the morning.

[0098] Through the above steps, the embodiment of the present application can determine the time required for the chiller to reach the target refrigeration effect based on the stable and low fluctuation of the chiller supply and return water temperature difference, and then optimize the start time of the chiller on the target date accordingly. The above process has low requirements for data diversity, and can achieve AI chiller energy-saving control through limited data without establishing a simulation model, achieving a good balance between energy saving and human comfort.

[0099] In the above step 103, the embodiment of the present application uses the prediction model with the best score as the first prediction model for predicting the cooling target time. As an alternative implementation, the embodiment of the present application may also use the M prediction models with the best scores as the first prediction model. Then, in step 104, the M prediction models are used to predict the cooling target time of the refrigeration machine on the target date, and the average of the predicted results is taken as the cooling target time of the refrigeration machine on the target date.

[0100] In addition, in the above step 103, the more features selected for predicting the time it takes to reach the cooling standard, the better. The embodiment of the present application can determine the final selected features / feature combinations based on the test results of the algorithm training phase, or can use the correlation between a single feature and the time it takes to reach the cooling standard (such as the Pierce correlation coefficient) as a reference for selecting features (for example, a correlation above 0.1 is generally considered to be high, but in actual applications, the prediction results of features with low correlation are not necessarily bad, and the results in the algorithm training shall prevail). Features generally include but are not limited to outdoor temperature, outdoor humidity, dry-bulb temperature, wet-bulb temperature, initial chilled water supply temperature (the moment before the chiller is turned on), chilled water flow, dew point, etc.

[0101] The above method of the embodiment of the present application can also determine the refrigeration failure time corresponding to each historical operation data according to each historical operation data, and the refrigeration failure time is the time required for the refrigeration effect to be lost after the refrigeration machine is turned off. The refrigeration failure time can be obtained by calculating the time between the time when the indoor temperature of the target building reaches the first temperature and the time when the refrigeration machine is turned off. The first temperature can usually be set according to the temperature that does not meet the human comfort.

[0102] Then, based on the multiple copies of historical operation data and the refrigeration failure duration corresponding to each copy of historical operation data, a second prediction model for predicting the refrigeration failure duration is trained. The specific training method is similar to the training of the first prediction model, for example, at least one feature combination is determined, and each feature combination includes at least one of the following features: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point. Then, based on each copy of the historical operation data and the refrigeration failure duration corresponding to the historical operation data, the historical feature data and refrigeration failure duration corresponding to each feature combination are determined; based on the historical feature data and refrigeration failure duration corresponding to each feature combination, a training set and a test set corresponding to the feature combination are generated; based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set; the prediction model with the best score is used as the second prediction model, and the second feature combination adopted by the second prediction model is determined.

[0103] Then, the second prediction model is used to predict the refrigeration failure duration of the HVAC system on the target date, and the closing time of the HVAC system on the target date is determined based on the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

[0104] It should be noted that when the above method predicts the opening and closing time of the chiller on the target date, if the final feature combination only includes weather features, the opening and closing time of the chiller can be predicted in advance according to the weather forecast data (depending on the time period that can be predicted by the obtained public weather forecast data); if it includes chiller-related features, it is necessary to predict the opening and closing time of the chiller moment by moment on the same day. The feature data selected is the weather at the current moment or chiller-related data.

[0105] In addition, the above embodiments are described only with a single chiller as an example. When the HVAC system includes multiple chillers, chiller scheduling or chiller sequencing can be performed in combination with the cooling load of the target building to determine the optimal start and stop times for each chiller.

[0106] Based on the above method, the present application embodiment also provides a device for implementing the above method. Figure 4 , a structure of a refrigeration machine energy-saving control device provided in an embodiment of the present application, the refrigeration machine energy-saving control device comprises:

[0107] The first acquisition module 401 is used to acquire multiple historical operation data of the HVAC system in the target building, each of which includes one or more of the following: the start time of the chiller, the stop time of the chiller, the chilled water supply temperature, the chilled water flow, the chilled water return temperature, the outdoor temperature, the indoor humidity, the dry bulb temperature, the wet bulb temperature, and the dew point of the HVAC system in a historical working cycle;

[0108] The first determination module 402 is used to determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, and the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on;

[0109] A first training module 403 is used to train a first prediction model for predicting the cooling target-reaching time according to the plurality of historical operation data and the cooling target-reaching time corresponding to each historical operation data;

[0110] The first prediction module 404 is used to use the first prediction model to predict the cooling time of the refrigeration machine on the target date, and determine the start time of the refrigeration machine on the target date based on the cooling time of the refrigeration machine on the target date and the opening start time of the target building.

[0111] Through the above modules, the embodiment of the present application optimizes the start-up time of the refrigerator to achieve energy saving without affecting the human comfort in the building.

[0112] Optionally, the first determining module is further used to:

[0113] Based on each historical operation data, the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time is calculated to obtain the chiller supply and return water temperature difference;

[0114] When the supply and return water temperature difference of the chiller is maintained within a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the cooling effect, and the cooling standard reaching time corresponding to the historical operation data is determined based on the chiller start-up time and the chiller achieving the cooling effect time.

[0115] Optionally, the preset temperature range is [T-α, T+α], where T is the target temperature and α is the fluctuation amplitude;

[0116] Wherein, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the multiple copies of historical operation data, or, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the season and time period to which the current historical operation data belongs in the multiple copies of historical operation data; the α is determined according to the fluctuation range of the indoor temperature of the target building while meeting the comfort requirements.

[0117] Optionally, the first training module is further used to:

[0118] Determine at least one characteristic combination, each characteristic combination includes at least one of the following characteristics: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point;

[0119] Based on each piece of the historical operation data and the cooling standard-reaching duration corresponding to the historical operation data, determine the historical feature data and the cooling standard-reaching duration corresponding to each feature combination;

[0120] Generate a training set and a test set corresponding to each feature combination based on the historical feature data and the cooling standard-reaching time corresponding to the feature combination;

[0121] Based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set;

[0122] The prediction model with the best score is used as the first prediction model, and a first feature combination used by the first prediction model is determined.

[0123] Optionally, the at least one prediction model includes one or more of the following models: a multi-layer perceptron model, a KNN model, a decision tree model, a gradient boosting model, and a random forest model.

[0124] Optionally, the first prediction module is further used to:

[0125] Obtaining a first feature value of the first feature combination on a target date;

[0126] The first characteristic value is input into the first prediction model to obtain the cooling standard-reaching time of the refrigerator on the target date predicted by the first prediction model.

[0127] It should be noted that the various systems provided in the above embodiments are devices corresponding to the above chiller energy-saving control method, and the implementation methods in the above embodiments are applicable to the embodiments of the device, and can also achieve the same technical effects. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effects, and the parts and beneficial effects that are the same as those in the method embodiment will not be specifically described here.

[0128] Please refer to Figure 5 , another structure of the chiller energy-saving control device provided in the embodiment of the present application is compared with Figure 4 The cooling machine energy-saving control device also includes:

[0129] The second determination module 405 is used to determine the refrigeration failure duration corresponding to each piece of historical operation data according to each piece of historical operation data, wherein the refrigeration failure duration is the time required for the refrigeration machine to lose the refrigeration effect after being shut down;

[0130] A second training module 406 is used to train a second prediction model for predicting the refrigeration failure duration according to the plurality of historical operation data and the refrigeration failure duration corresponding to each historical operation data;

[0131] The second prediction module 407 is used to predict the refrigeration failure duration of the HVAC system on the target date by using the second prediction model, and determine the closing time of the HVAC system on the target date according to the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

[0132] Please refer to Figure 6 , a schematic diagram of the structure of another refrigerator energy-saving control device provided in an embodiment of the present application, the device includes: a processor 601, a transceiver 602, a memory 603, a user interface 604 and a bus interface.

[0133] In the embodiment of the present application, the device also includes: a program stored in the memory 603 and executable on the processor 601 .

[0134] The transceiver 602 is used to send and receive data under the control of the processor;

[0135] The processor 601 is configured to read the computer program in the memory and perform the following operations:

[0136] Acquire multiple historical operation data of the HVAC system in the target building, each of the historical operation data includes one or more of the chiller start time, chiller shut-down time, chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point of the HVAC system in a historical working cycle;

[0137] Determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, wherein the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on;

[0138] According to the plurality of historical operation data and the cooling standard reaching time length corresponding to each historical operation data, a first prediction model for predicting the cooling standard reaching time length is obtained by training;

[0139] The first prediction model is used to predict the cooling time required for the refrigeration machine to meet the cooling standard on the target date, and the start time of the refrigeration machine on the target date is determined based on the cooling time required for the refrigeration machine to meet the cooling standard on the target date and the opening start time of the target building.

[0140] It can be understood that in the embodiment of the present application, when the computer program is executed by the processor 601, each process of the above-mentioned refrigeration machine energy-saving control method embodiment can be implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0141] exist Figure 6In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically one or more processors represented by processor 601 and various circuits of memory represented by memory 603 are linked together. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 602 may be a plurality of components, namely, a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium. For different user devices, the user interface 604 may also be an interface capable of externally and internally connecting required devices, and the connected devices include but are not limited to a keypad, a display, a speaker, a microphone, a joystick, and the like.

[0142] The processor 601 is responsible for managing the bus architecture and general processing, and the memory 603 can store data used by the processor 601 when performing operations.

[0143] It should be noted that the device in this embodiment is a device corresponding to the above-mentioned chiller energy-saving control method, and the implementation methods in the above-mentioned embodiments are all applicable to the embodiments of the device, and can also achieve the same technical effects. In the device, the transceiver 602 and the memory 603, as well as the transceiver 602 and the processor 601 can be connected through the bus interface communication, and the function of the processor 601 can also be implemented by the transceiver 602, and the function of the transceiver 602 can also be implemented by the processor 601. It should be noted that the above-mentioned device provided in the embodiment of the present application can implement all the method steps implemented in the above-mentioned method embodiment, and can achieve the same technical effects, and the parts and beneficial effects that are the same as the method embodiment in this embodiment will not be specifically repeated here.

[0144] In some embodiments of the present application, a computer-readable storage medium is further provided, on which a program is stored, and when the program is executed by a processor, the following steps are implemented:

[0145] Acquire multiple historical operation data of the HVAC system in the target building, each of the historical operation data includes one or more of the chiller start time, chiller shut-down time, chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point of the HVAC system in a historical working cycle;

[0146] Determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, wherein the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on;

[0147] According to the plurality of historical operation data and the cooling standard reaching time length corresponding to each historical operation data, a first prediction model for predicting the cooling standard reaching time length is obtained by training;

[0148] The first prediction model is used to predict the cooling time required for the refrigeration machine to meet the cooling standard on the target date, and the start time of the refrigeration machine on the target date is determined based on the cooling time required for the refrigeration machine to meet the cooling standard on the target date and the opening start time of the target building.

[0149] When the program is executed by the processor, it can implement all the implementation methods of the above-mentioned refrigeration machine energy-saving control method and achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0150] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0151] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0152] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0153] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0154] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0155] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0156] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A chiller energy-saving control method, characterized in that: include: Acquire multiple historical operation data of the HVAC system in the target building, each of the historical operation data includes one or more of the chiller start time, chiller shut-down time, chilled water supply temperature, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point of the HVAC system in a historical working cycle; Determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, wherein the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on; According to the plurality of historical operation data and the cooling standard reaching time length corresponding to each historical operation data, a first prediction model for predicting the cooling standard reaching time length is obtained by training; The first prediction model is used to predict the cooling time required for the refrigeration machine to meet the cooling standard on the target date, and the start time of the refrigeration machine on the target date is determined based on the cooling time required for the refrigeration machine to meet the cooling standard on the target date and the opening start time of the target building.

2. The method according to claim 1, characterized in that According to each piece of historical operation data, determine the cooling standard-reaching time corresponding to each piece of historical operation data, including: Based on each historical operation data, the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time is calculated to obtain the chiller supply and return water temperature difference; When the supply and return water temperature difference of the chiller is maintained within a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the cooling effect, and the cooling standard reaching time corresponding to the historical operation data is determined based on the chiller start-up time and the chiller achieving the cooling effect time.

3. The method according to claim 2, characterized in that The preset temperature range is [T-α, T+α], where T is the target temperature and α is the fluctuation amplitude; Wherein, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the multiple copies of historical operation data, or, the target temperature is set according to the average value of the supply and return water temperature difference of the chiller in the season and time period to which the current historical operation data belongs in the multiple copies of historical operation data; the α is determined according to the fluctuation range of the indoor temperature of the target building while meeting the comfort requirements.

4. The method according to claim 1, characterized in that According to the plurality of historical operation data and the cooling standard-reaching time length corresponding to each historical operation data, a first prediction model for predicting the standard-reaching time length is trained, including: Determine at least one characteristic combination, each characteristic combination includes at least one of the following characteristics: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point; Based on each piece of the historical operation data and the cooling standard-reaching duration corresponding to the historical operation data, determine the historical feature data and the cooling standard-reaching duration corresponding to each feature combination; Generate a training set and a test set corresponding to each feature combination based on the historical feature data and the cooling standard-reaching time corresponding to the feature combination; Based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set; The prediction model with the best score is used as the first prediction model, and a first feature combination used by the first prediction model is determined.

5. The method according to claim 4, characterized in that The at least one prediction model includes one or more of the following models: a multi-layer perceptron model, a K nearest neighbor (KNN) model, a decision tree model, a gradient boosting model, and a random forest model.

6. The method according to claim 4, characterized in that Using the first prediction model, predicting the cooling time of the refrigerator to reach the target cooling standard on the target date includes: Obtaining a first feature value of the first feature combination on a target date; The first characteristic value is input into the first prediction model to obtain the cooling standard-reaching time of the refrigerator on the target date predicted by the first prediction model.

7. The method according to any one of claims 1 to 6, characterized in that: Also includes: Determine the refrigeration failure duration corresponding to each piece of historical operation data according to each piece of historical operation data, wherein the refrigeration failure duration is the time required for the refrigeration machine to lose the refrigeration effect after being shut down; A second prediction model for predicting the refrigeration failure duration is obtained by training according to the plurality of historical operation data and the refrigeration failure duration corresponding to each historical operation data; The second prediction model is used to predict the refrigeration failure duration of the HVAC system on the target date, and the shutdown time of the HVAC system on the target date is determined based on the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

8. A cooling machine energy-saving control device, characterized in that: include: A first acquisition module is used to acquire multiple historical operation data of the HVAC system in the target building, each of which includes one or more of the following: a chiller start time, a chiller shut-down time, a chilled water supply temperature, a chilled water flow rate, a chilled water return temperature, an outdoor temperature, an indoor humidity, a dry-bulb temperature, a wet-bulb temperature, and a dew point of the HVAC system in a historical working cycle; A first determination module is used to determine the refrigeration standard time corresponding to each historical operation data according to each historical operation data, and the refrigeration standard time is the time required for the refrigeration machine to achieve the refrigeration effect after it is turned on; A first training module is used to train a first prediction model for predicting the duration of reaching the cooling standard based on the plurality of historical operation data and the duration of reaching the cooling standard corresponding to each piece of historical operation data; The first prediction module is used to use the first prediction model to predict the cooling time of the refrigeration machine on the target date, and determine the start time of the refrigeration machine on the target date based on the cooling time of the refrigeration machine on the target date and the opening start time of the target building.

9. The device according to claim 8, characterized in that The first determining module is further used for: Based on each historical operation data, the temperature difference between the chilled water supply temperature and the chilled water return temperature at the same collection time is calculated to obtain the chiller supply and return water temperature difference; When the supply and return water temperature difference of the chiller is maintained within a preset temperature range and reaches a preset time, it is determined that the chiller has achieved the cooling effect, and the cooling standard reaching time corresponding to the historical operation data is determined based on the chiller start-up time and the chiller achieving the cooling effect time.

10. The device according to claim 8, characterized in that The first training module is further used for: Determine at least one characteristic combination, each characteristic combination includes at least one of the following characteristics: initial supply temperature of chilled water, chilled water flow, chilled water return temperature, outdoor temperature, indoor humidity, dry bulb temperature, wet bulb temperature, and dew point; Based on each piece of the historical operation data and the cooling standard-reaching duration corresponding to the historical operation data, determine the historical feature data and the cooling standard-reaching duration corresponding to each feature combination; Generate a training set and a test set corresponding to each feature combination based on the historical feature data and the cooling standard-reaching time corresponding to the feature combination; Based on the training set corresponding to each feature combination, at least one prediction model is trained, and the trained prediction model is scored based on the test set; The prediction model with the best score is used as the first prediction model, and a first feature combination used by the first prediction model is determined.

11. The device according to claim 10, characterized in that The first prediction module is further used for: Obtaining a first feature value of the first feature combination on a target date; The first characteristic value is input into the first prediction model to obtain the cooling standard-reaching time of the refrigerator on the target date predicted by the first prediction model.

12. The device according to any one of claims 8 to 11, characterized in that Also includes: A second determination module is used to determine, based on each piece of historical operation data, a refrigeration failure duration corresponding to each piece of historical operation data, wherein the refrigeration failure duration is the time required for the refrigeration machine to lose its refrigeration effect after being shut down; A second training module is used to train a second prediction model for predicting the refrigeration failure duration according to the multiple historical operation data and the refrigeration failure duration corresponding to each historical operation data; The second prediction module is used to predict the refrigeration failure duration of the HVAC system on the target date by using the second prediction model, and determine the closing time of the HVAC system on the target date according to the refrigeration failure duration of the HVAC system on the target date and the opening deadline of the target building.

Images