Method for predicting load of efficient refrigerating machine room of air conditioner

Abstract

The invention provides an air-conditioning efficient refrigerating machine room load prediction method, and belongs to the technical field of refrigerating machine room load prediction. According to the method, the annual complex weather mode and the unit operation mode are classified through the set clustering method, so that the method is simplified and convenient to regulate and control; through respectively considering influence factors of internal and external loads, correlation degree screening is carried out, and the influence factor with the highest correlation degree, especially the influence of solar radiation, is selected, so that the calculation error of the external load in summer is reduced; the load change of the second day is predicted through the weather forecast of the second day, the operation number and operation time of the air conditioning units of the second day are directly regulated, the units are regulated in advance, the operation mode can be continuously optimized and adjusted according to the actual weather condition, manpower and material resources can be reduced, and the method is novel, suitable for different machine rooms and high in practicability. Load distribution of the refrigerating machine room is obtained by integrating and processing various kinds of information such as working information, environment information and weather data of the refrigerating machine room.

Description

Technical Field

[0001] The present invention relates to the technical field of refrigeration machine room load prediction, and particularly relates to a method for predicting the load of an efficient air-conditioning refrigeration machine room. Background Art

[0002] Central air conditioners are major energy consumers in buildings. For the refrigeration machine room that provides the cold source for the central air conditioner, since it includes main power-consuming equipment such as refrigeration hosts, water pumps, and cooling towers, the energy consumption of the refrigeration machine room is generally more than 50% of the total energy consumption of the central air-conditioning system. The advent of the information age has led to the continuous increase in the scale of data machine rooms, bringing a series of problems. Therefore, it is very necessary to conduct load prediction for the operation of the refrigeration machine room and give a regulation plan in advance for energy conservation and consumption reduction.

[0003] A method, system, and network-side server for optimizing the energy consumption control of a refrigeration machine room system disclosed in the Chinese invention patent application publication specification CN118368877B. Although this method establishes an annual load model of the refrigeration machine room system, and fully fits the actual load change trend through various variable parameters, provides suggestions for the selection of automated equipment, and is convenient for equipment selection in the initial design stage of the refrigeration machine room, enabling the equipment to operate under wide operating conditions and high efficiency. However, the existing control system of the refrigeration machine room has a low degree of intelligence, lacks effective adjustment strategies, and only supports some simple start-stop controls. At the same time, when formulating control strategies, it lacks the support of relevant air-conditioning professional knowledge. Currently, using sensors to read temperature data can only achieve real-time load calculation, unable to predict the load and provide an operation strategy in advance. And when calculating the external load, the influence of the solar radiation value is not considered, resulting in a lower calculated external load in summer and a larger calculation error. The existing load prediction methods generally rely on historical load data and are more suitable for daily predictions. When it is necessary to judge the load fluctuation situation in a short time, such as judging the load value one hour later, due to the influence of factors such as weather and indoor environmental conditions, the predicted value often deviates greatly from the actual value, that is, the prediction effect of the existing load prediction methods is not ideal. Therefore, the present application provides a method for predicting the load of an efficient air-conditioning refrigeration machine room to meet the requirements. Summary of the Invention

[0004] (1) Technical Problems to be Solved

[0005] Aiming at the deficiencies of the prior art, the present invention provides a method for predicting the load of an efficient air-conditioning refrigeration machine room, which solves the problems that the existing control system of the refrigeration machine room has a low degree of intelligence, lacks effective adjustment strategies, and due to the influence of factors such as weather and indoor environmental conditions, the predicted value often deviates greatly from the actual value.

[0006] (2) Technical Solutions

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] An air-conditioning high-efficiency refrigeration machine room load prediction method, comprising the following steps:

[0009] Step 1: Obtain the environmental information and working information of the air-conditioning high-efficiency refrigeration machine room at regular intervals. The working information includes the cooling load information of the air-conditioning high-efficiency refrigeration machine room and the real-time power consumption of the refrigeration machine room. Export the cooling load information of the refrigeration machine room and the real-time power consumption of the refrigeration machine room through a port;

[0010] Step 2: Obtain historical hourly weather data and temperature data through a meteorological interface, and export the weather data and temperature data through a port;

[0011] Step 3: Integrate and collect the obtained environmental information, working information, weather data, and temperature data;

[0012] Step 4: Preprocess the exported data;

[0013] Step 5: Analyze the data using Pearson correlation, and screen out the characteristic parameters with strong correlation with the air-conditioning load. That is, the Pearson correlation coefficient ranges from [-1, 1]. The closer it is to 1, the stronger the positive correlation, and the closer it is to -1, the stronger the negative correlation;

[0014] Step 6: Predict the solar radiation value for the next day;

[0015] Step 7: Obtain the hourly dry-bulb temperature for the next day from the weather forecast network, and calculate the outdoor temperature Tout by combining the predicted solar radiation value for the next day in Step 4;

[0016] Step 8: Calculate the outdoor load Qout, indoor load Qin, and total machine room load Qtotal based on the outdoor thermometer Tout in Step 4; where Qtotal = Qout + Qin;

[0017] Step 9: Read the weather forecast meteorological data for the next day, calculate the predicted total machine room load Qtotal according to Steps 4 to 6, combine the rated cooling capacity of the unit, calculate the number of operating units, and send a command to the automatic control system for pre-regulation.

[0018] Preferably, the working information includes the supply water temperature, return water temperature, and supply water flow rate of the chilled water main pipe in the air-conditioning high-efficiency refrigeration machine room, and the control device obtains the cooling load information of the machine room according to the supply water temperature, return water temperature, and supply water flow rate.

[0019] Preferably, the cooling load = specific heat × density × supply water flow rate / 3600, the specific heat = 4.184 kJ / (kg·°C), and the density = 1010 kg / m3.

[0020] Preferably, the environmental information includes environmental temperature information, environmental humidity information, and solar irradiance information.

[0021] Preferably, the weather information in step two includes dry bulb temperature, relative humidity, solar radiation intensity, wind speed, and weather type.

[0022] Preferably, the specific steps of step four are as follows: First, convert non-numerical character data to numerical data after exporting the data; Second, set the threshold of the weather data change range based on the actual weather data, and eliminate the highest or lowest values exceeding the range threshold; Third, search for and fill in the missing values for the corresponding time series to obtain data with a standard format and a complete time series.

[0023] Preferably, when searching for and filling in the missing values for the corresponding time series, the mean value or exponential smoothing method is used to fill in the missing values.

[0024] Preferably, the specific steps of step six are as follows: First, divide the hourly solar radiation values for the whole year into the cooling season, transition season, and heating season; Second, screen the hourly solar radiation values for different seasons according to different weather types to form several single data groups; The weather types include: clear, cloudy, overcast, and rainy; Third, perform k-means mean clustering based on the distance matrix for each of the screened single data groups; Among them, the turning point where the change of the within-cluster error variance method gradually slows down is the optimal number of clusters K; Fourth, evaluate the clustering effect of the single data groups after mean clustering through the silhouette coefficient; Fifth, obtain that there are i types of hourly solar radiation change values when the weather type is clear, j types when it is cloudy, k types when it is overcast, and v types when it is rainy for the corresponding season; Sixth, integrate the 24-hour clustering results for each type and the corresponding hourly dry bulb temperature change values into a data group and store it in the database; Seventh, identify the season in which the second day is located according to the date of the second day; Eighth, locate to the corresponding n clustering clusters in the database according to the weather type of the second day; Ninth, calculate the similarity between the hourly dry bulb temperature derived from the weather forecast website for the second day and the hourly dry bulb temperature change values corresponding to the clustering clusters in the database through the Euclidean distance, and the clustering cluster where the hourly dry bulb temperature change value with the smallest similarity is located is the solar radiation prediction result for the second day.

[0025] Preferably, the outdoor temperature Tout in step seven is calculated by Tout = Tg + ρS / α, where Tg is the dry bulb temperature in the weather forecast, ρ is the solar radiation absorption coefficient, α is the surface heat transfer coefficient of the outer surface of the building envelope, and S is the solar radiation intensity.

[0026] Preferably, in the seventh step, the external load Qout is calculated by the enclosure structure load calculation formula Qout = kF(Tout - Tin), where k is the heat transfer coefficient of the enclosure structure, F is the surface area of the enclosure structure, and Tin is the indoor temperature of the computer room; the internal load Qin is the heat dissipation load of the IT equipment collected by the smart electricity meter. In the eighth step, the number of operating units = the total load of the computer room / the rated cooling capacity of the unit.

[0027] Compared with the prior art, the present invention has at least the following beneficial effects:

[0028] In the above solution, the clustering method is set to classify the complex annual weather patterns and unit operation patterns, making them simpler and easier to control; by separately considering the influencing factors of internal and external loads and performing correlation screening to select the influencing factor with the highest correlation, especially the influence of solar radiation, the calculation error of the external load in summer is reduced; by predicting the load change of the next day based on the weather forecast of the next day and directly adjusting the number of operating units and operating time of the air-conditioning unit of the next day, the advance control of the unit is realized. It can not only continuously optimize and adjust the operation mode according to the actual weather conditions, but also reduce manpower and material resources, and the method is novel and applicable to different computer rooms.

[0029] By integrating and processing various information such as the working information, environmental information, and weather data of the refrigeration machine room to obtain the load distribution of the refrigeration machine room, that is, in the process of load prediction, various factors affecting the load value are comprehensively considered to improve the accuracy and reliability of the detection results.

[0030] In summary, the present invention has the advantages of continuously optimizing and adjusting the operation mode according to the actual weather conditions, reducing manpower and material resources, and comprehensively considering various factors affecting the load value to improve the accuracy and reliability of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a detailed flow schematic diagram of the load prediction method for an efficient air-conditioning refrigeration machine room;

[0032] Figure 2 It is a detailed flow schematic diagram of data preprocessing for the load prediction method of an efficient air-conditioning refrigeration machine room.

[0033] As shown in the figure, in order to clearly show the structure of the embodiments of the present invention, specific structures and devices are marked in the figure, but this is only for schematic needs and is not intended to limit the present invention to this specific structure, device, and environment. Those of ordinary skill in the art can adjust or modify these devices and environments according to specific needs. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] The following will describe in detail a method for predicting the load of an efficient refrigeration machine room of an air conditioner provided by the present invention in conjunction with the accompanying drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the accompanying drawings are only for more specifically describing the embodiments and are not intended to specifically limit the present invention.

[0035] It should be noted that in the specification, when referring to "an embodiment", "embodiment", "exemplary embodiment", "some embodiments", etc., it indicates that the described embodiment may include specific features, structures or characteristics, but not necessarily every embodiment includes such specific features, structures or characteristics. Additionally, when combining an embodiment to describe a specific feature, structure or characteristic, implementing such a feature, structure or characteristic in combination with other embodiments (whether explicitly described or not) should be within the knowledge scope of those skilled in the relevant art.

[0036] Generally, terms can be understood at least in part from their use in the context. For example, at least in part depending on the context, the term "one or more" used herein can be used to describe any feature, structure or characteristic in a singular sense, or can be used to describe a combination of features, structures or characteristics in a plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but rather, at least in part depending on the context, can allow for the existence of other factors that are not necessarily explicitly described.

[0037] It can be understood that the meanings of "on...", "above...", and "over..." in the present invention should be interpreted in the broadest manner, such that "on..." not only means "directly on" something, but also includes the meaning of being "on" something with intervening features or layers therebetween, and "above..." or "over..." not only means "above" or "over" something, but also can include the meaning of being "above" or "over" something with no intervening features or layers therebetween.

[0038] Furthermore, spatial relative terms such as "under...", "below...", "lower", "above...", "upper", etc. are used herein for convenience of description to describe the relationship between one element or feature and another or more elements or features, as shown in the accompanying drawings. The spatial relative terms are intended to cover different orientations in the use or operation of the device in addition to the orientation depicted in the accompanying drawings. The device can be oriented in another manner, and the spatial relative descriptive terms used herein can be correspondingly interpreted similarly.

[0039] As Figure 1As shown, an embodiment of the present invention provides a method for predicting the load of an efficient refrigeration machine room of an air conditioner. By obtaining the environmental information and working information of the refrigeration machine room every fifteen minutes, the working information includes the cooling load information of the refrigeration machine room and the real-time power consumption information of the refrigeration machine room. It should be noted that the staff can adjust the acquisition time interval of the environmental information and working information according to the actual working conditions of the refrigeration machine room. When the changes in the environmental information and working information are large, the acquisition time interval of the environmental information and working information can be shortened to improve the accuracy of load prediction. The environmental information includes the environmental temperature information of the refrigeration machine room, the environmental humidity information of the refrigeration machine room, and the solar irradiance information of the prediction point. The prediction point refers to the moment when the control device obtains information. The cooling load information of the refrigeration machine room is calculated based on three data: the supply water temperature, the return water temperature, and the supply water flow rate of the chilled water main pipe of the water system of the refrigeration machine room. Specifically, the cooling load = specific heat × density × supply water flow rate / 3600, where the specific heat is 4.184 kJ / (kg·°C), the density is 1010 kg / m3, the temperature difference unit between the return water temperature and the supply water temperature is °C, and the unit of the supply water flow rate is m3 / h. Historical hourly weather data is collected through a meteorological interface, and the historical hourly total air-conditioning operation load of the machine room is collected through an intelligent electricity meter. Through wireless signal transmission, the weather data is exported through a port. The weather data includes dry bulb temperature, relative humidity, solar radiation intensity, wind speed, and weather type. The exported data is preprocessed. After exporting the data, the non-numerical character data is converted into numerical data; for example, in the weather type, 1 represents sunny, 2 represents cloudy, 3 represents overcast, and 4 represents rainy. Based on the actual weather data, the change range thresholds of the weather data are set, and the highest or lowest values exceeding the range thresholds are removed. Among them, the change range thresholds of each weather data include dry bulb temperature (°C) [-50, 50], relative humidity (%) [0, 100], solar radiation intensity (W / m2) [0, 1368], wind speed (m / s) [0, 100], weather type [1, 4]. The missing values are searched and filled for the corresponding time series; data with a standard format and a complete time series is obtained, and Pearson correlation is used to analyze the data, and the characteristic parameters with strong correlation with the air-conditioning load are screened out; that is, the value range of the Pearson correlation coefficient is [-1, 1], the closer to 1, the stronger the positive correlation, and the closer to -1, the stronger the negative correlation.

[0040] As Figure 2As shown, in this embodiment, since there are differences in the changes of annual weather parameters during a day, the hourly solar radiation values of the historical whole year are divided into the cooling season, the transitional season, and the heating season. The hourly solar radiation values in different seasons are screened according to different weather types to form several single data groups. The weather types include: sunny, cloudy, overcast, and rainy. K-means mean clustering based on the distance matrix is performed on the screened single data groups respectively, which is beneficial to attributing the complex annual changes to several patterns. Among them, the turning point where the change of the within-cluster error variance method gradually slows down is the optimal number of clusters K. The clustering effect of the single data groups after mean clustering is evaluated through the silhouette coefficient. It is obtained that under the corresponding season, when the weather type is sunny, there are i types of hourly changes in solar radiation, when it is cloudy, there are j types, when it is overcast, there are k types, and when it is rainy, there are v types. The 24-hour clustering results under each type and the corresponding hourly dry-bulb temperature change values are integrated into a data group and stored in the database. The season in which the next day is located is identified according to the date of the next day, and the corresponding n clustering clusters are located in the database according to the weather type of the next day. The similarity is calculated by calculating the Euclidean distance between the hourly dry-bulb temperature derived from the weather forecast network for the next day and the hourly dry-bulb temperature change values corresponding to the clustering clusters in the database. The clustering cluster where the hourly dry-bulb temperature change value with the smallest similarity is located is the solar radiation prediction result for the next day. The hourly dry-bulb temperature for the next day is obtained from the weather forecast network and combined with the predicted solar radiation prediction value in step S4 to calculate the outdoor temperature Tout, Tout = Tg + ρS / α, where Tg is the weather forecast dry-bulb temperature, ρ is the solar radiation absorption coefficient, α is the surface heat transfer coefficient of the outer surface of the building envelope, and S is the solar radiation intensity. The outdoor temperature Tout is used to calculate the external load Qout, the internal load Qin, and the total load Qtotal of the computer room. Among them, the external load Qout is calculated by the building envelope load calculation formula Oout = kF(Tout - Tin), where k is the building envelope heat transfer coefficient, F is the surface area of the building envelope, and Tin is the indoor temperature of the computer room, usually set to 24 - 26°C. The internal load Qin is the heat dissipation load of the IT equipment collected by the smart meter. The total load Qtotal of the computer room = external load Qout + internal load Qin. Since the building information can be changed according to different projects, mainly including the building envelope information, which includes the building envelope heat transfer coefficient k, the surface area of the building envelope F, the solar radiation absorption coefficient ρ, the surface heat transfer coefficient α, etc. The meteorological data of the weather forecast for the next day is read, and the predicted total load Qtotal of the computer room is calculated according to steps four to eight. Combining with the rated cooling capacity of the unit, the number of operating units is calculated. The number of operating units = total load of the computer room / rated cooling capacity of the unit, and an instruction is sent to the automatic control system to implement pre-regulation.

[0041] Taking the Changzhou Tianmu Lake Data Center of China Telecom as an example, at the time of June 17, 2022, the load forecasting method for the operation control of the data room based on weather forecasting was used to predict the total load of the data room on the next day, June 18. The specific steps are as follows: Read the weather forecasting data for June 18 from the port, including the dry-bulb temperature and weather type for 24 hours. Since June 18 belongs to the cooling season and the weather type is cloudy, the hourly dry-bulb temperature is calculated pairwise with the historical hourly dry-bulb temperature in the database, and the clustering cluster with the minimum similarity value is the corresponding predicted value of solar radiation. Calculate the external load (the negative sign represents the heat load) and the total load of the data room based on the predicted dry-bulb temperature and solar radiation value for the next day, as shown in the following table:

[0042]

[0043]

[0044]

[0045] Judge the number of units to be started according to the hourly total load. The rated cooling capacity of a single chiller is 3868 kw, and the predicted total load value for the whole day of June 18 is less than the rated cooling capacity of a single chiller. Therefore, only one chiller should be turned on throughout the day, and an instruction is sent to the automatic control system to implement pre-control.

[0046] All the electrical components appearing in this article are electrically connected to the external main controller and 220V mains electricity, and the main controller can be a conventional known device such as a computer for control.

[0047] When an air-conditioning high-efficiency refrigeration machine room load forecasting method is used, it includes the following steps;

[0048] Step 1: Obtain the environmental information and working information of the air-conditioning high-efficiency refrigeration machine room at regular intervals. The working information includes the cold load information of the air-conditioning high-efficiency refrigeration machine room and the real-time power consumption information of the refrigeration machine room, and export the cold load information of the refrigeration machine room and the real-time power consumption information of the refrigeration machine room through the port;

[0049] Step 2: Obtain the historical hourly weather data and temperature data through the meteorological interface, and export the weather data and temperature data through the port;

[0050] Step 3: Integrate and collect the obtained environmental information, working information, weather data and temperature data;

[0051] Step 4: Preprocess the exported data;

[0052] Step 5: Analyze the data using Pearson correlation, and screen out the characteristic parameters with strong correlation with the air-conditioning load, that is, the Pearson correlation coefficient ranges from [-1, 1]. The closer to 1, the stronger the positive correlation, and the closer to -1, the stronger the negative correlation;

[0053] Step 6. Predict the solar radiation value for the next day;

[0054] Step 7. Obtain the hourly dry-bulb temperature for the next day from the weather forecast network and calculate the outdoor temperature Tout by combining with the predicted solar radiation value for the next day in Step 4;

[0055] Step 8. Calculate the external load Qout, internal load Qin, and total machine room load Qtotal based on the outdoor thermometer Tout in Step 4; where Qtotal = Qout + Qin;

[0056] Step 9. Read the weather forecast meteorological data for the next day, calculate the predicted total machine room load Qtotal according to Steps 4 to 6, combine with the rated cooling capacity of the unit, calculate the number of operating units of the unit, and issue an instruction to the automatic control system for pre-regulation.

[0057] The technical solution provided by the present invention: By obtaining the environmental information and working information of the air-conditioning high-efficiency refrigeration machine room at regular intervals, the working information includes the cold load information of the air-conditioning high-efficiency refrigeration machine room and the real-time power consumption information of the refrigeration machine room, exporting the cold load information of the refrigeration machine room and the real-time power consumption information of the refrigeration machine room through a port, obtaining historical hourly weather data and temperature data through a meteorological interface, exporting the weather data and temperature data through a port, then integrating and collecting the obtained environmental information, working information, weather data and temperature data, preprocessing the exported data, analyzing the data using Pearson correlation, and screening out the characteristic parameters with strong correlation with the air-conditioning load, that is, the Pearson correlation coefficient ranges from [-1, 1], the closer to 1, the stronger the positive correlation, the closer to -1, the stronger the negative correlation, predicting the solar radiation value for the next day, obtaining the hourly dry-bulb temperature for the next day from the weather forecast network and calculating the outdoor temperature Tout by combining with the predicted solar radiation value for the next day in Step 4, calculating the external load Qout, internal load Qin, and total machine room load Qtotal based on the outdoor thermometer Tout in Step 4; where Qtotal = Qout + Qin, reading the weather forecast meteorological data for the next day, calculating the predicted total machine room load Qtotal according to Steps 4 to 6, combining with the rated cooling capacity of the unit, calculating the number of operating units of the unit, and issuing an instruction to the automatic control system for pre-regulation.

[0058] The present invention covers any substitutions, modifications, equivalent methods, and solutions made within the essence and scope of the present invention. For the public to have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, and those skilled in the art can fully understand the present invention without these detailed descriptions. Additionally, to avoid unnecessary confusion to the essence of the present invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0059] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for predicting load of an air-conditioning high-efficiency refrigeration room, characterized in that: The following steps are involved: Step 1: Obtain environmental information and working information of the air-conditioning high-efficiency refrigeration room at regular intervals. The working information includes the cooling load information of the air-conditioning high-efficiency refrigeration room and the real-time power consumption information of the refrigeration room. The cooling load information and the real-time power consumption information of the refrigeration room are exported through the port. Step 2: Obtain historical hourly weather data and temperature data through the meteorological interface, and export the weather data and temperature data through the port; Step 3: Integrate and collect the acquired environmental information, work information, weather data and temperature data; Step 4: Preprocess the exported data; Step 5: Use Pearson correlation to analyze the data and select characteristic parameters with strong correlation with air conditioning load. That is, the value range of Pearson correlation coefficient is [-1,1]. The closer to 1, the stronger the positive correlation, and the closer to -1, the stronger the negative correlation. Step 6: Predict the solar radiation value for the next day; Step 7: Obtain the hourly dry-bulb temperature of the next day from the weather forecast website and calculate the outdoor temperature Tout in combination with the predicted solar radiation value of the next day predicted in step 4; Step 8: The outdoor thermometer Tout in step 4 calculates the external load Qout, the internal load Qin and the total load Qtotal of the computer room; wherein Qtotal = Qout + Qin; Step 9: Read the weather forecast data for the next day, calculate the predicted total load Qtotal of the computer room according to steps 4 to 6, calculate the number of units in operation based on the rated cooling capacity of the units, and send instructions to the automatic control system to implement pre-regulation.

2. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1 is characterized in that: The working information includes the supply water temperature, return water temperature and supply water flow rate of the chilled water main of the air-conditioning high-efficiency refrigeration room, and the control device obtains the cooling load information of the room according to the supply water temperature, return water temperature and supply water flow rate.

3. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1 is characterized in that: The cooling load = specific heat × density × water supply flow / 3600, the specific heat = 4.184 kJ / (kg · ° C), the density = 1010 kg / m3.

4. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The environmental information includes environmental temperature information, environmental humidity information and solar irradiance information.

5. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The weather information in step 2 includes dry-bulb temperature, relative humidity, solar radiation intensity, wind speed and weather type.

6. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The specific steps of step 4 are as follows: first, converting non-numeric character data into numeric data after exporting the data; second, setting a weather data change range threshold based on actual weather data, and eliminating the highest or lowest value that exceeds the range threshold; Third, search and fill missing values ​​in the corresponding time series to obtain data with standard format and complete time series.

7. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The corresponding time series searches for and fills in missing values, and uses mean or exponential smoothing to fill in missing values.

8. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The specific steps of step six are as follows: first, dividing the historical hourly solar radiation values ​​throughout the year into a cooling season, a transition season, and a heating season; Second, the hourly solar radiation values ​​in different seasons are screened according to different weather types to form several single data groups; weather types include: sunny, cloudy, overcast, and rainy; third, the single data groups after screening are clustered by k-means based on the distance matrix; among them, the turning point where the change of the intra-cluster error variance method gradually slows down is the optimal clustering number K; fourth, the clustering effect of the single data group after mean clustering is evaluated by the silhouette coefficient; fifth, the hourly change values ​​of solar radiation in different weather types in the corresponding seasons are i types when it is sunny, j types when it is cloudy, and k types when it is overcast , there are v types of rainy days; Sixth, the 24-hour clustering results and the corresponding hourly dry-bulb temperature change values ​​under each type are integrated into a data group and stored in the database; Seventh, the season of the second day is identified according to the date of the second day; Eighth, the corresponding n clusters are located in the database according to the weather type of the second day; Ninth, the hourly dry-bulb temperature of the second day derived from the weather forecast network and the hourly dry-bulb temperature change values ​​corresponding to the clusters in the database are calculated pairwise by Euclidean distance to obtain the similarity. The cluster with the hourly dry-bulb temperature change value with the smallest similarity is the solar radiation forecast result of the second day.

9. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: The outdoor temperature Tout in step seven is calculated by Tout=Tg+ρS / α, wherein Tg is the weather forecast dry bulb temperature, ρ is the solar radiation absorption coefficient, α is the surface heat transfer coefficient of the outer surface of the enclosure structure, and S is the solar radiation intensity.

10. The method for predicting load of an air-conditioning high-efficiency refrigeration room according to claim 1, characterized in that: In step seven, the external load Qout is calculated by the enclosure structure load calculation formula Qout=kF(Tout-Tin), wherein k is the heat transfer coefficient of the enclosure structure, F is the surface area of ​​the enclosure structure, and Tin is the indoor temperature of the computer room; the internal load Qin is the heat dissipation load of the IT equipment collected by the smart meter, and in step eight, the number of operating units=total load of the computer room / rated cooling capacity of the unit.

Images