A prediction method and device for a weather-climate integrated intelligent large model

Abstract

The present invention relates to the technical field of mathematical models, and discloses a prediction method and device for a weather-climate integrated intelligent large model. In this method, weather information and climate information for a prediction period are integrated to obtain weather-climate characteristics; original weather-climate change trend characteristics are calculated based on weather information and climate information for a historical period; the original weather-climate change trend characteristics and power consumption change trend characteristics are input into an attention mechanism model to obtain target weather-climate change trend characteristics that conform to the power consumption change trend; the weather-climate characteristics and the target weather-climate change trend characteristics are spliced ​​to generate spliced ​​features; the spliced ​​features are combined with the equipment information and the population information to form a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period. The present invention achieves accurate prediction of the power consumption of the cooling system in a regional centralized cooling scenario.

Description

Technical Field

[0001] The present invention relates to the technical field of mathematical models, and in particular to a prediction method and device for a weather-climate integrated intelligent large model. Background Art

[0002] Centralized cooling offers advantages such as intensive land use, staggered capacity reduction, high energy efficiency, intelligent management and control, and the avoidance of redundant investment. It is suitable for deployment in densely populated commercial areas, such as urban central business districts. Users of regional centralized cooling systems can flexibly adjust temperature and humidity levels based on their actual needs, ensuring stable indoor temperatures and humidity.

[0003] District cooling systems rely on water for cooling. They typically consist of energy stations (cold and heat sources), a distribution network, cooling (heat) exchange stations, and cooling terminals (such as fan coil units). Chilled water is often transported through a district cooling network between energy stations and buildings. The entire cooling system is similar to the human heart and circulatory system. Compared to individual cooling services for individual users, district cooling can significantly reduce electricity consumption.

[0004] In the prior art, the power consumption of a centralized cooling system is generally determined based on predicted weather temperatures. This prediction method is significantly different from the actual power consumption, resulting in inaccurate power consumption predictions. In view of this, the present invention is proposed. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a prediction method and device for a weather-climate integrated intelligent large model, which realizes the accurate prediction of the power consumption of the cooling system in the scenario of regional centralized cooling.

[0006] In a first aspect, an embodiment of the present invention provides a prediction method for a weather-climate integrated intelligent large model, comprising:

[0007] Collecting equipment information of the cooling system during the forecast period, population information of the region during the forecast period, weather and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods, where the historical periods are adjacent to the forecast period;

[0008] The weather information and climate information of the forecast period are integrated to obtain the weather-climate characteristics;

[0009] Calculate the power consumption trend characteristics based on the power consumption of the cooling system during historical periods;

[0010] Calculating original weather-climate change trend features based on weather and climate information over historical periods; inputting the original weather-climate change trend features and power consumption change trend features into an attention mechanism model to obtain target weather-climate change trend features that match the power consumption change trend;

[0011] Splicing the weather-climate feature with the target weather-climate change trend feature to generate a spliced ​​feature;

[0012] The spliced ​​features, the equipment information, and the population information are combined into a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model.

[0013] Optionally, the population information collected in the region during the forecast period includes:

[0014] Collect historical data on population inflow and outflow and the total permanent population in the area under the same weather and climate conditions;

[0015] Based on the inflow and outflow of the population and the total permanent population, determine the proportion of the population in the region to the total permanent population during the same historical period;

[0016] The population size in the area during the forecast period is determined based on the total permanent population during the forecast period and the ratio.

[0017] Optionally, the calculation of original weather-climate change trend characteristics based on weather information and climate information of a historical period includes:

[0018] Calculate the weather change trend based on the weather information of adjacent periods in the historical period, and calculate the climate change trend based on the climate information of adjacent periods in the historical period;

[0019] The weather change trend and climate change trend are integrated to obtain the original weather-climate change trend characteristics.

[0020] Optionally, the weather information includes temperature, humidity, wind speed and wind direction; and the climate information includes sunlight and precipitation.

[0021] Optionally, inputting the weather-climate characteristics, the target weather-climate change trend characteristics, the equipment information, and the population information into a prediction model to obtain the power consumption of the cooling system during a prediction period includes:

[0022] Splicing the weather-climate feature with the target weather-climate change trend feature to generate a spliced ​​feature;

[0023] The spliced ​​features, the equipment information, and the population information are combined into a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period.

[0024] Optionally, weather information and climate information for the forecast period are fused to obtain weather-climate characteristics, including:

[0025] Normalize the weather information of the forecast period to generate the first column vector;

[0026] Normalize the climate information at the prediction time to generate the second column vector;

[0027] Taking a weighted sum of the first column vector and the second column vector to obtain weather-climate characteristics;

[0028] Among them, the weights of the first column vector and the second column vector are trained together with the attention mechanism model and the prediction model.

[0029] In a second aspect, an embodiment of the present invention provides a prediction device for a weather-climate integrated intelligent large-scale model, comprising:

[0030] a collection module for collecting equipment information of the cooling system during the forecast period, population information in the region during the forecast period, weather and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods, wherein the historical periods are adjacent to the forecast period;

[0031] A fusion module is used to fuse the weather information and climate information of the forecast period to obtain weather-climate characteristics;

[0032] A calculation module, used to calculate the power consumption change trend characteristics based on the power consumption of the cooling system during historical periods;

[0033] A mapping module is configured to calculate raw weather-climate change trend features based on weather and climate information over a historical period; input the raw weather-climate change trend features and the power consumption change trend features into an attention mechanism model to obtain target weather-climate change trend features that conform to the power consumption change trend;

[0034] A prediction module, configured to combine the weather-climate characteristics with target weather-climate change trend characteristics to generate combined characteristics;

[0035] The spliced ​​features, the equipment information, and the population information are combined into a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model.

[0036] The embodiments of the present invention have the following technical effects:

[0037] 1. The present invention generally uses a combination of weather and climate information to predict the power consumption of the cooling system, which is more accurate than prediction based on weather temperature alone.

[0038] 2. The present invention takes into account the equipment information of the cooling system and the difference in power consumption caused by different equipment models / types.

[0039] 3. The present invention inputs the original weather-climate change trend characteristics and the power consumption change trend characteristics into the attention mechanism model to obtain the target weather-climate change trend characteristics that conform to the power consumption change trend. Through the calculation of the change trend and the introduction of the attention mechanism model, the target weather-climate change trend characteristics describe how the change trend affects power consumption, that is, the impact characteristics of weather-climate on power consumption are extracted. This impact characteristic, as the input feature of the prediction model, can significantly improve the accuracy of the power consumption prediction value.

[0040] 4. This invention employs large-scale modeling technology through feature collection, fusion, attention mapping, and prediction to ultimately determine the cooling system's power consumption during the predicted period. This combined large-scale model possesses enhanced learning and generalization capabilities, and its overall prediction accuracy increases with the number of training samples. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 This is a flow chart of a prediction method for a weather-climate integrated intelligent large model provided by an embodiment of the present invention;

[0043] Figure 2 This is a structural diagram of a weather-climate integrated intelligent large model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0045] As mentioned in the background, existing technologies typically determine the power consumption of centralized cooling systems based on predicted weather temperatures, but actual power consumption can vary significantly. This is because power consumption is influenced not only by weather but also by other factors. With so many factors at play, figuring out how to handle them and building a comprehensive model requires significant creative effort.

[0046] The method provided by this invention is applicable to predicting the power consumption of district cooling systems. District cooling involves setting up a centralized refrigeration station within a building complex to produce chilled water for air conditioning, which is then supplied to individual buildings, such as residential communities, through a circulating water piping system. The cooling water pumps, control devices, and other components of the cooling system consume electricity, and this power consumption is positively correlated with the amount of cooling provided. The method provided by this invention is executed by a prediction device based on a large, integrated weather-climate intelligent model, which is integrated into electronic equipment in the form of hardware and / or software.

[0047] Example 1

[0048] Figure 1 This is a flow chart of a prediction method for a weather-climate integrated intelligent large model provided by an embodiment of the present invention. Figure 2 This is a structural diagram of a weather-climate integrated intelligent large model provided by an embodiment of the present invention.

[0049] See also Figure 1 and Figure 2 , the method provided by the embodiment of the present invention includes the following operations:

[0050] S110 , collecting equipment information of the cooling system during the forecast period, population information in the region during the forecast period, weather information and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods.

[0051] The forecast period refers to tomorrow or the next few hours. The historical period refers to the period adjacent to the forecast period, such as yesterday or the previous few hours.

[0052] The equipment information of the cooling system includes the power and efficiency of the electrical equipment, which directly affects the power consumption, such as the power and efficiency of the chiller and the power and efficiency of the water pump.

[0053] The population information in the region mainly includes the population size. In practical applications, considering that people of different ages and genders have different demands for cooling, it may also include the gender distribution and age distribution of the population. The population information in the region can be obtained through surveys. Preferably, considering that population mobility is affected by weather and climate, for example, when the weather is hot or rainy, a large number of people will gather in the region and will not go out; for example, when the weather is relatively cool, some people will go out to play. In order to accurately reflect the impact of weather and climate on population size, this embodiment collects the population information in the region during the forecast period, including: collecting the inflow and outflow of the population and the total number of permanent residents in the region during the same historical weather and climate; determining the proportion of the population in the region to the total number of permanent residents during the same historical weather and climate based on the inflow and outflow of the population and the total number of permanent residents; determining the population size in the region during the forecast period based on the total number of permanent residents during the forecast period and the proportion.

[0054] For example, compare the weather and climate during the forecast period to a period A during the historical period with the same weather and climate. During this period, count the number of people entering and leaving the area, as well as the total permanent population (B), using cameras installed on porches or entry and exit passes. Subtract the number of people leaving the area from the total permanent population, and add the number of people entering the area to obtain the population C in the area during the same historical weather and climate. C / B is used to determine the ratio of the area's population to the total permanent population, meaning that approximately this proportion of the population would remain in the area under these weather and climate conditions. If the total permanent population during the forecast period is D, multiply D by this ratio to obtain the total population during the forecast period. Since the forecast period is tomorrow or several hours in the future, the total permanent population counted for the current period can be used as the total permanent population for the forecast period.

[0055] The weather information and climate information for the forecast period can be obtained from the meteorological department. The weather information includes temperature, humidity, wind speed and wind direction; and the climate information includes sunlight and precipitation.

[0056] The power consumption of the cooling system during the historical period can be read through the power meter of the cooling system.

[0057] S120: Fusing the weather information and climate information of the forecast period to obtain weather-climate characteristics.

[0058] Weather and climate coexist in time and space, jointly causing people to experience cold and heat. This example explores the combined impact of weather and climate on human perception, which better reflects people's demand for cooling and thus directly affects the power consumption of the cooling system. Specifically, weather information for the forecast period is normalized to generate a first column vector; climate information for the forecast time is also normalized to generate a second column vector; the first and second column vectors are weighted and summed to obtain the weather-climate characteristics.

[0059] Normalization is the process of converting temperature, humidity, wind speed, and direction in weather data, and sunlight and precipitation in climate data, to a uniform scale for feature calculations. Before normalization, weather and climate data must be encoded, using numbers to represent specific information.

[0060] The weights of the first and second column vectors represent the impact on the power consumption of the cooling system. The weights can be pre-set to an initial value and then updated during the training process of the large model.

[0061] S130. Calculate power consumption change trend characteristics based on the power consumption of the cooling system during historical periods.

[0062] For example, if it's May 21st, the historical period includes May 18, May 19, and May 20. The cooling system's daily power consumption can be measured using the meter. The difference between the power consumption on two consecutive days—that is, the difference between May 18 and May 19, and the difference between May 19 and May 20—is calculated. These differences are used as power consumption trend characteristics.

[0063] S140. Calculate original weather-climate change trend characteristics based on weather information and climate information of a historical period; input the original weather-climate change trend characteristics and power consumption change trend characteristics into an attention mechanism model to obtain target weather-climate change trend characteristics that conform to the power consumption change trend.

[0064] The weather trend is calculated based on the weather information of adjacent periods in the historical period, and the climate trend is calculated based on the climate information of adjacent periods in the historical period. Specifically, similar to the power consumption trend feature, the feature codes of the weather information of two adjacent days are subtracted to obtain a number of differences, which are used as the weather trend; the feature codes of the climate information of two adjacent days are subtracted to obtain a number of differences, which are used as the climate trend.

[0065] Then, the weather change trend and the climate change trend are fused to obtain the original weather-climate change trend feature. The fusion method can be to splice the weather change trend and the climate change trend into a string as the original weather-climate change trend feature.

[0066] The original weather-climate change trend features and the power consumption change trend features obtained in S130 are input into the attention mechanism model self-attention to obtain the target weather-climate change trend features output by self-attention that are consistent with the power consumption change trend. Target weather-climate change trend features. The self-attention model selectively filters out a small amount of important information from a large amount of information and focuses on this important information, ignoring most of the unimportant information. In this embodiment, the self-attention model obtains weather-climate change trend features that have a significant impact on the change in power consumption from the original weather-climate change trend features. The present invention uses "change trend" as the input feature of the self-attention model. This is because when the weather and climate change, power consumption will inevitably change accordingly. Only in this dynamic change trend can the "weather-climate features" that have a significant impact on power consumption be found.

[0067] S150. Splicing the weather-climate characteristics with the target weather-climate change trend characteristics to generate spliced ​​characteristics; forming a feature matrix with the spliced ​​characteristics, the equipment information, and the population information, and inputting the matrix into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model.

[0068] The weather-climate characteristics in S150 are obtained in S120 and reflect the characteristics of the forecast period. The target weather-climate change trend characteristics are obtained in S140 and reflect the characteristics of the historical period. The device information and population information are for the forecast period and reflect the impact on power consumption from the physical and social levels, respectively.

[0069] The present invention introduces the features of the above aspects and inputs them into the prediction model for prediction, which is conducive to improving the accuracy of the prediction.

[0070] Optionally, the prediction model is a deep neural network model, such as a BP neural network model or an LSTM model.

[0071] The weather-climate features (dimensions 1×q) are concatenated with the target weather-climate change trend features (dimensions 1×n) to generate concatenated features (dimensions 2×max(n,q)). Any missing features in the concatenated features are padded with 0. The concatenated features are combined with the device information (dimensions 1×m) and the population information (1×p) to form a feature matrix (4×max(n,m,q,p)). Any missing features in the feature matrix are padded with 0. The feature matrix is ​​input into the prediction model to obtain the power consumption of the cooling system during the prediction period.

[0072] It should be noted that, in the embodiment of the present invention, the weights of the first column vector and the second column vector are trained together with the attention mechanism model and the prediction model. During the training process, a sufficient number of samples need to be collected and labeled with the power consumption of the cooling system. The samples include: equipment information of the cooling system in the first period, population information in the area in the first period, weather information and climate information in the first and second periods, and power consumption of the cooling system in the second period; the second period is adjacent to the first period; the second period is earlier than the first period;

[0073] The weather information of the first time period and the climate information of the first time period are integrated to obtain weather-climate characteristics; the power consumption change trend characteristics are calculated according to the power consumption of the cooling system in the second time period; the original weather-climate change trend characteristics are calculated according to the weather information and climate information of the second time period; the original weather-climate change trend characteristics and the power consumption change trend characteristics are input into the attention mechanism model to obtain the target weather-climate change trend characteristics that conform to the power consumption change trend; the weather-climate characteristics, the target weather-climate change trend characteristics, the equipment information and the population information are input into the prediction model to be trained to obtain the power consumption of the cooling system in the prediction period, and the loss function is constructed according to the difference between the power consumption and the label, and the weights of the first column vector and the second column vector, the parameters in the attention mechanism model and the parameters in the prediction model are updated.

[0074] The embodiments of the present invention have the following technical effects:

[0075] 1. The present invention generally uses a combination of weather and climate information to predict the power consumption of the cooling system, which is more accurate than prediction based on weather temperature alone.

[0076] 2. The present invention takes into account the equipment information of the cooling system and the difference in power consumption caused by different equipment models / types.

[0077] 3. The present invention inputs the original weather-climate change trend characteristics and the power consumption change trend characteristics into the attention mechanism model to obtain the target weather-climate change trend characteristics that conform to the power consumption change trend. Through the calculation of the change trend and the introduction of the attention mechanism model, the target weather-climate change trend characteristics describe how the change trend affects power consumption, that is, the impact characteristics of weather-climate on power consumption are extracted. This impact characteristic, as the input feature of the prediction model, can significantly improve the accuracy of the power consumption prediction value.

[0078] 4. This invention employs large-scale modeling technology through feature collection, fusion, attention mapping, and prediction to ultimately determine the cooling system's power consumption during the predicted period. This combined large-scale model possesses enhanced learning and generalization capabilities, and its overall prediction accuracy increases with the number of training samples.

[0079] The embodiment of the present invention further provides a prediction device for a weather-climate integrated intelligent large-scale model, comprising:

[0080] a collection module for collecting equipment information of the cooling system during the forecast period, population information in the region during the forecast period, weather information and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods, wherein the historical periods are adjacent to the forecast period;

[0081] A fusion module is used to fuse the weather information and climate information of the forecast period to obtain weather-climate characteristics;

[0082] A calculation module, used to calculate the power consumption change trend characteristics based on the power consumption of the cooling system during historical periods;

[0083] A mapping module is configured to calculate raw weather-climate change trend features based on weather and climate information over a historical period; input the raw weather-climate change trend features and the power consumption change trend features into an attention mechanism model to obtain target weather-climate change trend features that conform to the power consumption change trend;

[0084] A prediction module is used to splice the weather-climate characteristics with the target weather-climate change trend characteristics to generate spliced ​​characteristics; the spliced ​​characteristics are combined with the equipment information and the population information to form a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model.

[0085] The device provided by the present invention is used to execute the prediction method of the weather-climate integrated intelligent large model provided by the above embodiment, and has specific corresponding technical effects.

[0086] It should be noted that the terms used in the present invention are only for describing specific embodiments and are not intended to limit the scope of this application. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "an", "a kind of" and / or "the" do not specifically refer to the singular and may also include the plural. The terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method or device comprising a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also include elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "comprise a..." do not exclude the presence of other identical elements in the process, method or device comprising the elements.

[0087] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limitations on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the internal parts of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A prediction method for a weather-climate integrated intelligent large-scale model, characterized in that: include: Collecting equipment information of the cooling system during the forecast period, population information in the region during the forecast period, weather and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods; The historical period is adjacent to the forecast period; wherein collecting population information within the region during the forecast period includes: collecting the population inflow and outflow and the total permanent population within the region during the same historical weather and climate; determining the proportion of the population within the region to the total permanent population during the same historical period based on the population inflow and outflow and the total permanent population; determining the population size within the region during the forecast period based on the total permanent population during the forecast period and the proportion; the equipment information includes the power and efficiency of electrical equipment; Fusing weather information and climate information for the forecast period to obtain weather-climate characteristics, including: normalizing the weather information for the forecast period to generate a first column vector; normalizing the climate information for the forecast time to generate a second column vector; and performing a weighted summation of the first column vector and the second column vector to obtain the weather-climate characteristics; Calculate the power consumption trend characteristics based on the power consumption of the cooling system during historical periods; Calculating original weather-climate change trend features based on weather and climate information over historical periods; inputting the original weather-climate change trend features and power consumption change trend features into an attention mechanism model to obtain target weather-climate change trend features that match the power consumption change trend; Splicing the weather-climate feature with the target weather-climate change trend feature to generate a spliced ​​feature; The spliced ​​features, the device information, and the population information are combined into a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model; During the training process, samples are collected and labeled with labels of the cooling system's power consumption; the samples include: cooling system equipment information for the first period, regional population information for the first period, weather and climate information for the first and second periods, and cooling system power consumption for the second period; the second period is adjacent to the first period; and the second period is earlier than the first period. The weather information of the first time period and the climate information of the first time period are integrated to obtain weather-climate characteristics; the power consumption change trend characteristics are calculated according to the power consumption of the cooling system in the second time period; the original weather-climate change trend characteristics are calculated according to the weather information and climate information of the second time period; the original weather-climate change trend characteristics and the power consumption change trend characteristics are input into the attention mechanism model to obtain the target weather-climate change trend characteristics that conform to the power consumption change trend; the weather-climate characteristics, the target weather-climate change trend characteristics, the equipment information and the population information are input into the prediction model to be trained to obtain the power consumption of the cooling system in the prediction period, and the loss function is constructed according to the difference between the power consumption and the label, and the weights of the first column vector and the second column vector, the parameters in the attention mechanism model and the parameters in the prediction model are updated.

2. The method according to claim 1, characterized in that The calculation of original weather-climate change trend characteristics based on weather information and climate information of historical periods includes: Calculate the weather change trend based on the weather information of adjacent periods in the historical period, and calculate the climate change trend based on the climate information of adjacent periods in the historical period; The weather change trend and climate change trend are integrated to obtain the original weather-climate change trend characteristics.

3. The method according to claim 2, characterized in that The weather information includes temperature, humidity, wind speed and wind direction; the climate information includes sunlight and precipitation.

4. A weather-climate integrated intelligent large-scale model prediction device, characterized in that: include: A collection module is used to collect equipment information of the cooling system during the forecast period, population information in the region during the forecast period, weather information and climate information during the forecast period and historical periods, and power consumption of the cooling system during historical periods; The historical period is adjacent to the forecast period; wherein collecting population information within the region during the forecast period includes: collecting the population inflow and outflow and the total permanent population within the region during the same historical weather and climate; determining the proportion of the population within the region to the total permanent population during the same historical period based on the population inflow and outflow and the total permanent population; determining the population size within the region during the forecast period based on the total permanent population during the forecast period and the proportion; the equipment information includes the power and efficiency of electrical equipment; a fusion module for fusing weather information and climate information for the forecast period to obtain weather-climate characteristics, including: normalizing the weather information for the forecast period to generate a first column vector; normalizing the climate information for the forecast time to generate a second column vector; and performing a weighted summation of the first column vector and the second column vector to obtain the weather-climate characteristics; A calculation module, used to calculate the power consumption change trend characteristics based on the power consumption of the cooling system during historical periods; A mapping module is configured to calculate raw weather-climate change trend features based on weather and climate information over a historical period; input the raw weather-climate change trend features and the power consumption change trend features into an attention mechanism model to obtain target weather-climate change trend features that conform to the power consumption change trend; a prediction module configured to combine the weather-climate characteristics with target weather-climate change trend characteristics to generate combined characteristics; combine the combined characteristics with the device information and the population information to form a feature matrix, which is input into a prediction model to obtain the power consumption of the cooling system during the prediction period; the prediction model is a deep neural network model; During the training process, samples are collected and labeled with labels of the cooling system's power consumption; the samples include: cooling system equipment information for the first period, regional population information for the first period, weather and climate information for the first and second periods, and cooling system power consumption for the second period; the second period is adjacent to the first period; and the second period is earlier than the first period. The weather information of the first time period and the climate information of the first time period are integrated to obtain weather-climate characteristics; the power consumption change trend characteristics are calculated according to the power consumption of the cooling system in the second time period; the original weather-climate change trend characteristics are calculated according to the weather information and climate information of the second time period; the original weather-climate change trend characteristics and the power consumption change trend characteristics are input into the attention mechanism model to obtain the target weather-climate change trend characteristics that conform to the power consumption change trend; the weather-climate characteristics, the target weather-climate change trend characteristics, the equipment information and the population information are input into the prediction model to be trained to obtain the power consumption of the cooling system in the prediction period, and the loss function is constructed according to the difference between the power consumption and the label, and the weights of the first column vector and the second column vector, the parameters in the attention mechanism model and the parameters in the prediction model are updated.

Images