Cascade air conditioner electricity consumption statistics method

Abstract

The invention provides a cascade air conditioner power consumption statistics method, and relates to the field of intelligent power consumption management, and the method comprises the following steps: obtaining the operation data of a target extension of a cascade air conditioner in a specific time period and the total power consumption of an outdoor unit in a corresponding time period; acquiring a power utilization behavior mode of the target extension set according to the operation data; determining a power utilization influence coefficient of the target extension set according to the power utilization behavior mode and an area proportion of a region where the target extension set is located; and determining the power utilization weight of the target extension according to the proportion of the power utilization influence coefficient to the sum of the power utilization influence coefficients of all the extensions, and further calculating the power consumption of the target extension. According to the method, the use areas and the power consumption behaviors of different individual households are comprehensively considered, the power consumption of the extension sets in different operation times, set temperatures, wind speeds and operation modes is metered, the problem that the shared cost of the cascade air conditioner is difficult to collect is solved, and the power consumption management efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent power consumption management, and in particular to a method for counting power consumption of cascade air conditioners. Background Art

[0002] In the field of intelligent buildings, cascade air conditioning systems have become the mainstream configuration of large office buildings due to their energy efficiency advantages of multi-cycle series design. However, there is a significant contradiction between the existing electricity fee sharing mechanism and the dynamic operation characteristics of cascade air conditioners. Studies have shown that the air conditioning usage intensity in different areas of the same building can vary by 2-3 times, and the temperature setting deviation exceeds 4°C, but the shared costs cannot reflect this dynamic change. How to achieve effective metering and charging of cascade air conditioners is particularly critical.

[0003] At present, the conventional metering and charging methods on the market mainly include different metering methods such as apportionment by area and metering by household according to operating time. Among them: apportionment by area does not take into account the user's usage habits, which is easy to cause users to consume maliciously and waste energy, and does not reflect the principle of fair charging according to the "usage" of cascade air conditioners; metering by time is to apportion the electricity consumption of air-conditioning units according to the user's usage time, but only considers the user's usage time, and does not consider the cooling / heat output of the indoor unit. Different electricity consumption behaviors will also lead to differences in electricity consumption. It can be seen that the existing technology is difficult to comprehensively consider the usage area and electricity consumption behavior of different households, resulting in the inability to accurately estimate the electricity consumption of each extension, thereby causing low efficiency in electricity management. Summary of the invention

[0004] An embodiment of the present invention provides a method for counting electricity consumption of cascaded air conditioners, aiming to solve the problem that the prior art is difficult to comprehensively consider the electricity consumption behaviors under different household usage areas and different operating times, set temperatures, wind speeds and operating modes, resulting in the inability to accurately estimate the electricity consumption of each extension, thereby improving the efficiency of electricity management.

[0005] In order to achieve the above object, the present invention provides a method for counting power consumption of cascade air conditioners, comprising the following steps:

[0006] Obtain the operating data of the target extension of the cascade air conditioner within a specific time period and the total power consumption of the outdoor unit within the corresponding time period;

[0007] Acquire the power consumption behavior pattern of the target extension according to the operation data, specifically input the operation data into a pre-trained deep learning model, and output the power consumption behavior pattern of the target extension, wherein the deep learning model is configured as a PCA model and a SOM model, wherein the PCA model extracts operation data features by dimensionality reduction, and the SOM model defines the power consumption behavior pattern by clustering the operation data features;

[0008] Determine the electricity consumption impact coefficient of the target extension machine according to the electricity consumption behavior pattern and the area proportion of the area where the target extension machine is located; the electricity consumption impact coefficient is determined by the influence degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit;

[0009] Determine the electricity consumption weight of the target extension machine according to the proportion of the electricity consumption impact coefficient of the target extension machine in the sum of the electricity consumption impact coefficients of all extension machines;

[0010] Obtain the electricity consumption of the target extension machine according to the product of the electricity consumption weight and the total electricity consumption of the outdoor unit.

[0011] Further, the operation data includes operation time, set temperature, wind speed, and operation mode.

[0012] Further, the PCA model includes the following steps:

[0013] Collect the operation data samples of the extension machines in the historical period;

[0014] Convert all the operation data samples into a data matrix in numerical form;

[0015] Obtain several principal components by reducing the dimension of the data matrix;

[0016] Obtain the linear expression of the principal components with respect to the operation data for extracting the operation data features.

[0017] Further, the SOM model includes the following steps:

[0018] Initialize the weight vectors of the two-dimensional grid neurons;

[0019] Input the operation data features into the network and find the most matching neuron by calculating the Euclidean distance;

[0020] Then adjust the weights of it and its surrounding neurons according to the Gaussian neighborhood function with the most matching neuron as the center;

[0021] Make the weights converge through multiple iterations to form a stable mapping;

[0022] Cluster adjacent neurons to form a clustering cluster and define the electricity consumption pattern of the clustering cluster.

[0023] Further, the influence degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit is determined by the following steps:

[0024] Collect the electricity consumption behavior patterns and the area proportion data of all extension machines of the cascade air conditioner in the historical period, as well as the total electricity consumption of the outdoor unit in the corresponding time period;

[0025] Statistically calculate the total sum of the area proportions of different electricity consumption behavior patterns;

[0026] Taking the sum of the area proportions of different electricity consumption behavior patterns as the explanatory variable and the total electricity consumption of the outdoor unit as the response variable, a random forest model is constructed;

[0027] Sort the influence degrees of different electricity consumption behavior patterns through the random forest model;

[0028] According to the order of the influence degrees, assign corresponding electricity consumption influence coefficients to different electricity consumption behavior patterns.

[0029] Furthermore, the electricity consumption weight is calculated as follows:

[0030]

[0031] In the formula, D i is the electricity consumption weight of the i-th extension unit; S i is the area proportion of the region where the i-th extension unit is located; X i is the electricity consumption influence coefficient of the i-th extension unit; X sum is the sum of the electricity consumption influence coefficients of all extension units.

[0032] Furthermore, the electricity consumption of the cascaded air conditioner extension unit is calculated as follows:

[0033] F i = D i × Z d ,

[0034] In the formula, F i is the electricity consumption of the i-th extension unit; D i is the electricity consumption weight of the i-th extension unit; Z d is the total electricity consumption of the outdoor unit.

[0035] Furthermore, when the electricity consumption of the target extension unit exceeds the preset electricity consumption threshold, a reminder message is sent to the user of the corresponding extension unit.

[0036] The above technical solution has the following technical effects:

[0037] (1) Through multiple operation data of the cascaded air conditioner extension unit within a specific time period, determine the electricity consumption behavior pattern of the extension unit during this period, and determine the electricity consumption influence coefficient of the extension unit according to the electricity consumption behavior pattern and the area proportion of the region where the extension unit is located; determine the electricity consumption weight of the extension unit and then determine the electricity consumption of the extension unit through the proportion of the electricity consumption influence coefficient. It solves the problem that the prior art is difficult to comprehensively consider the usage area of each household and the electricity consumption behavior under different operation times, set temperatures, wind speeds and operation modes, resulting in inaccurate evaluation of the electricity consumption of each household, and improves the efficiency of comprehensive electricity management.

[0038] (2) Extract the operation data features through the PCA model, and then use the SOM model to cluster the operation data features, so as to efficiently identify the electricity consumption behavior patterns of the extension machines.

[0039] (3) Determine the influence degree of different electricity consumption behavior patterns on the total electricity consumption by constructing a random forest model, solve the problem that it is difficult to estimate the non-linear change of the electricity consumption influence coefficient, and improve the accuracy of electricity consumption evaluation. Description of the Drawings

[0040] Figure 1 It is a schematic flowchart of the method for statistical calculation of the electricity consumption of cascade air conditioners in an embodiment of the present invention;

[0041] Figure 2 It is a schematic flowchart of the method for determining the electricity consumption behavior pattern of the target extension machine in an embodiment of the present invention;

[0042] Figure 3 It is a schematic flowchart of the method for determining the influence degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit in an embodiment of the present invention. Detailed Embodiments

[0043] To further illustrate the embodiments, the present invention provides drawings. These drawings are part of the disclosure of the present invention, which are mainly used to illustrate the embodiments and can be used to explain the operating principles of the embodiments in combination with the relevant descriptions in the specification. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are usually used to represent similar components.

[0044] The present invention will be further described below in conjunction with the drawings and specific embodiments.

[0045] Embodiment 1:

[0046] Figure 1 It is a schematic flowchart of the method for statistical calculation of the electricity consumption of cascade air conditioners in an embodiment of the present invention. The method of this embodiment includes the following steps:

[0047] Obtain the operation data of the target extension machine of the cascade air conditioner within a specific time period and the total electricity consumption of the outdoor unit within the corresponding time period;

[0048] In this embodiment, it is necessary to extract the detailed operation data of the target extension machine within the specified time period from the air conditioning system. The operation data includes operation time, set temperature, wind speed, and operation mode. At the same time, record the total electricity consumption of the outdoor unit within the same time period, and this data can be obtained through an electric meter or the system background.

[0049] Obtain the electricity consumption behavior pattern of the target extension machine according to the operation data;

[0050] In this embodiment, based on the operation data of the first step, the electricity consumption behavior patterns of the extension machines can be identified. For example: high-load patterns such as continuous high-temperature refrigeration, high wind speed, long-time operation, etc., intermittent patterns such as frequent on / off, short-time operation, etc., low-load patterns such as low-temperature heating, low wind speed, night standby, etc. The behavior patterns reflect the actual energy consumption requirements of the extension machines and are the core basis for evaluating their impact on the energy consumption of the outdoor unit.

[0051] Determine the electricity consumption impact coefficient of the target extension machine according to the electricity consumption behavior pattern and the area proportion of the area where the target extension machine is located; the electricity consumption impact coefficient is determined by the impact degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit;

[0052] Traditional energy consumption calculations often only consider a single factor, either simply based on area or only referring to electricity consumption behavior, which makes the energy consumption calculation results deviate greatly from the actual energy consumption situation. In this embodiment, by comprehensively considering the electricity consumption behavior pattern and the area proportion of the area where the target extension machine is located, a more accurate and reasonable electricity consumption impact coefficient can be obtained. Different electricity consumption behavior patterns, such as high-load patterns of continuous refrigeration and heating, intermittent patterns of frequent on / off, low-load patterns of maintaining low temperature, etc., have very different impact degrees on the total electricity consumption of the outdoor unit. At the same time, the area cannot be ignored. The larger the area, the larger the space that needs to be temperature-adjusted, and the more electricity the outdoor unit consumes to meet the demand.

[0053] By comprehensively considering these two key factors, the actual impact of the target extension machine on the total electricity consumption of the outdoor unit can be evaluated more comprehensively and accurately. The electricity consumption impact coefficient determined in this way can not only reflect the energy consumption differences brought about by the operation characteristics of the extension machine itself but also take into account the energy consumption changes caused by the objective conditions of its location area, making the energy consumption calculation and analysis more in line with the actual situation.

[0054] Determine the electricity consumption weight of the target extension machine according to the proportion of the electricity consumption impact coefficient of the target extension machine in the total sum of the electricity consumption impact coefficients of all extension machines; the calculation formula of the electricity consumption weight is as follows:

[0055]

[0056] In the formula, D i is the electricity consumption weight of the i-th extension machine; S i is the area proportion of the area where the i-th extension machine is located; X i is the electricity consumption impact coefficient of the i-th extension machine; X sum is the total sum of the electricity consumption impact coefficients of all extension machines.

[0057] Obtain the electricity consumption of the target extension machine according to the product of the electricity consumption weight and the total electricity consumption of the outdoor unit. The calculation formula of the electricity consumption of the cascaded air conditioner extension machine is as follows:

[0058] F i= D i × Z d ,

[0059] where F i is the power consumption of the i-th extension; D i is the power consumption weight of the i-th extension; Z d is the total power consumption of the outdoor unit.

[0060] Furthermore, when the power consumption of the target extension exceeds a preset power consumption threshold, a reminder message is sent to the user of the corresponding extension.

[0061] In this embodiment, after accurately calculating the power consumption of the target extension, there is an important step, namely, monitoring and warning the power consumption. The system pre-sets a power consumption threshold, which can be comprehensively determined based on multiple factors such as historical power consumption data, the function of the area where the extension is located, and the energy-saving goal. When the calculated power consumption of the target extension exceeds the preset threshold, the system will automatically trigger a reminder mechanism and accurately send the reminder message to the user of the corresponding extension. It enables users to timely understand their own power consumption situation, realize possible high-energy-consuming behaviors, and thus take energy-saving measures such as adjusting the air-conditioning operation mode and setting a more reasonable temperature; on the other hand, it also helps to control the overall energy consumption, prevent the excessive power consumption of individual extensions from affecting the energy consumption balance of the entire system, and promote the rational use and efficient management of energy.

[0062] Embodiment 2:

[0063] Figure 2 FIG. is a schematic flow chart of determining the power consumption behavior pattern of the target extension in an embodiment of the present invention. The method of this embodiment is specifically as follows:

[0064] Input the operation data into a pre-trained deep learning model, and output the power consumption behavior pattern of the target extension. The deep learning model is configured as a PCA model and a SOM model. The PCA model extracts the operation data features through dimensionality reduction, and the SOM model defines the power consumption behavior pattern by clustering the operation data features;

[0065] In the process of calculating the electricity consumption behavior pattern of the target extension machine, an advanced method based on a deep learning model is used. First, the operation data of the target extension machine obtained will be used as input data. The deep learning model used here is composed of a combination of a principal component analysis (PCA) model and a self-organizing map (SOM) model. The PCA model plays a key preprocessing role. Since the original operation data may contain many dimensions and there is information redundancy, the PCA model extracts the most representative and important features from a large amount of operation data through dimensionality reduction technology, removing the information that has little impact on analyzing the electricity consumption behavior pattern or is redundant. This can not only reduce the complexity and computational amount of data processing, but also more clearly highlight the key features, making the subsequent analysis more targeted.

[0066] The operation data features obtained after being processed by the PCA model will be input into the SOM model. The SOM model is an unsupervised learning neural network model, which can perform clustering operations on these data features. Clustering is the process of grouping data points with similar features into one category. In this context, the SOM model will divide the operation situation of the target extension machine into different categories according to the similarity of the operation data features, and each category corresponds to an electricity consumption behavior pattern, such as high-load continuous operation mode, intermittent start-stop mode, low-load stable operation mode, etc.

[0067] Through this deep learning method combining the PCA model and the SOM model, it is able to automatically and relatively accurately identify the electricity consumption behavior pattern of the target extension machine from complex operation data, providing a solid foundation for subsequent work such as calculating the electricity consumption impact coefficient, energy consumption analysis, and formulating energy-saving strategies based on the electricity consumption behavior pattern. Compared with traditional manual analysis or simple statistical methods, it has higher efficiency and accuracy.

[0068] Furthermore, the PCA model includes the following steps:

[0069] Collect the operation data samples of the extension machine in the historical period:

[0070] By collecting the operation data of the extension machine in the past period of time, such as the set temperature, operation mode (cooling, heating, air supply, etc.), wind speed, etc. at different time periods. These historical data cover various possible operation situations of the extension machine, providing rich materials for subsequent analysis. The larger the amount of data collected, the longer the time span, and the richer the operation scenarios covered, the more conducive it is for the subsequent model to accurately extract data features.

[0071] Convert all the operation data samples into a numerical data matrix:

[0072] Since the extension operation data may contain multiple types of information, such as the operation mode is expressed in words such as "cooling" and "heating". In order to facilitate computer processing and analysis, these non-numerical data need to be converted into numerical form. After such a conversion, all the operation data samples form a data matrix, where each row of the matrix represents a sample (i.e. the operation status of the extension at a certain moment or a certain time period), and each column represents a feature (such as temperature, mode, etc.).

[0073] Several principal components are obtained by reducing the dimension of the data matrix:

[0074] After obtaining the data matrix, since the dimension of the data may be very high (i.e., it contains many features), and there may be certain correlations between these features, this will increase the complexity and amount of calculation of data processing. One of the core functions of the PCA model is to reduce the dimension of the data matrix. It maps high-dimensional data to a low-dimensional space by finding the main direction of change in the data, while retaining the main information of the data as much as possible. The new dimension obtained after dimensionality reduction is the principal component. These principal components are linear combinations of the original features, which are orthogonal to each other (i.e., independent of each other) and are sorted according to the size of their contribution to the data variance. The principal component with the greater variance contribution is ranked first, indicating that it contains more original data information.

[0075] Get the linear expression of the principal component about the running data to extract the running data features:

[0076] After obtaining the principal components through dimensionality reduction, it is necessary to determine how each principal component is linearly combined from the features of the original operating data. In other words, it is necessary to find the mathematical relationship between the principal components and the features of the original operating data to obtain a linear expression. Through this linear expression, the weights of each original feature in each principal component can be clearly known. In this way, when analyzing new operating data in the future, the corresponding principal component values can be calculated based on this linear expression, thereby extracting the key features of the operating data. These extracted features can more concisely and accurately reflect the operating status of the extension, providing an important basis for subsequent power consumption behavior pattern analysis and other related energy consumption analysis.

[0077] Furthermore, the SOM model comprises the following steps:

[0078] Initialize the weight vector of neurons in a 2D grid:

[0079] The SOM model constructs a two-dimensional grid structure of neurons. Before starting to process data, it is necessary to initialize the weight vectors for each neuron. The dimensions of these weight vectors are the same as those of the input running data features. The initial values of the weight vectors can be set randomly, and they represent the initial positions of the neurons in the feature space. Each neuron has its own weight vector, and these weight vectors will be continuously adjusted during the subsequent learning process to adapt to the feature distribution of the input data.

[0080] Input the running data features into the network and find the most matching neuron by calculating the Euclidean distance:

[0081] When the running data features are input into the SOM model, the model calculates the Euclidean distance between the weight vector of each neuron and the input data feature vector. The Euclidean distance is a common method for measuring the distance between two vectors, and the smaller the distance, the more similar the two vectors are. By comparing the Euclidean distances of all neurons with the input data features, the neuron with the smallest distance is found, and this neuron is called the "most matching neuron" (also known as the winning neuron). It represents the neuron that can best reflect the input data features in the current state.

[0082] Then, taking the most matching neuron as the center, adjust the weights of it and its surrounding neurons according to the Gaussian neighborhood function:

[0083] After finding the most matching neuron, not only the weight vector of this neuron needs to be adjusted, but also the weights of its surrounding neurons need to be adjusted according to the Gaussian neighborhood function with it as the center. The Gaussian neighborhood function defines a neighborhood range centered on the most matching neuron, and the neurons within this range will be affected. The closer a neuron is to the most matching neuron, the greater the amplitude of its weight adjustment; the farther away, the smaller the amplitude of the weight adjustment. In this way, the neuron grid can better adapt to the local features of the input data, and at the same time, it also helps to form a certain topological structure among the neurons, that is, adjacent neurons also tend to represent similar data in the feature space.

[0084] Through multiple iterations, make the weights converge to form a stable mapping:

[0085] The process of adjusting weights as described above is not performed only once, but rather iterated multiple times for a large number of operating data features. Each time new data features are input, the steps of finding the most matching neurons and adjusting the weights are repeated. As the number of iterations increases, the weight vectors of the neurons gradually converge, that is, they no longer change significantly. Eventually, the neuron grid forms a stable mapping such that similar data features in the feature space are mapped to adjacent neurons in the neuron grid. In this way, the SOM model can map high-dimensional operating data features onto a two-dimensional neuron grid while preserving the topological structure and similarity of the data.

[0086] Cluster adjacent neurons to form clusters, and define the electricity consumption patterns of the clusters:

[0087] After the weights converge to form a stable mapping, clustering operations are performed on the neurons in the two-dimensional grid. Adjacent neurons are grouped into a cluster because adjacent neurons represent similar data features in the feature space, so the operating data features corresponding to the neurons in the same cluster are also similar. According to the characteristics of the data features in each cluster, the corresponding electricity consumption patterns can be defined. For example, if the data features in a certain cluster all indicate that the extension is in a high-load and long-running state, then this cluster can be defined as the "high-load continuous operation electricity consumption pattern".

[0088] Embodiment 3:

[0089] Figure 3 This is a schematic flowchart for determining the influence degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit in an embodiment of the present invention. The influence degree of the electricity consumption behavior pattern on the total electricity consumption of the outdoor unit is determined through the following steps:

[0090] Collect the electricity consumption behavior patterns and the data of the area proportion of the location of all extensions of the cascade air conditioner in the historical period, as well as the total electricity consumption of the outdoor unit in the corresponding time period:

[0091] Comprehensively collect the relevant data of all extensions in the cascade air conditioner system in the past period, including the electricity consumption behavior pattern of each extension and the area proportion of the area where the extension is located. At the same time, the total electricity consumption of the outdoor unit in the corresponding time period should also be accurately recorded. These data cover the operating characteristics and spatial factors of the extensions and the overall energy consumption situation of the outdoor unit.

[0092] Statistically calculate the total sum of the area proportions of different electricity consumption behavior patterns:

[0093] After collecting the data of all extensions, the extensions are classified according to their electricity consumption behavior patterns. For each electricity consumption behavior pattern, the sum of the area ratios of all extensions belonging to that pattern is calculated. For example, calculate the total area ratio of all extensions in the high-load continuous operation mode. This total reflects the spatial scale corresponding to the extensions with this electricity consumption behavior pattern in the entire system. Through such statistics, the spatial distribution of different electricity consumption behavior patterns can be understood macroscopically.

[0094] Using the sum of the area ratios of different electricity consumption behavior patterns as the explanatory variable and the total electricity consumption of the external unit as the response variable, a random forest model is constructed:

[0095] The random forest model is a machine learning algorithm used to analyze the relationship between the sum of the area ratios of different electricity consumption behavior patterns and the total electricity consumption of the external unit. Take the sum of the area ratios of different electricity consumption behavior patterns obtained from the previous statistics as the explanatory variable (independent variable), which represents the factors that may affect the total electricity consumption of the external unit; while the total electricity consumption of the external unit is used as the response variable (dependent variable). By constructing a random forest model, the model will automatically learn the complex non-linear relationship between these explanatory variables and the response variable, and uncover the potential influence law of the area ratios of different electricity consumption behavior patterns on the total electricity consumption of the external unit.

[0096] Rank the influence degrees of different electricity consumption behavior patterns through the said random forest model:

[0097] The random forest model has the characteristic of being able to evaluate the importance of variables. After constructing and training the model, utilize this characteristic of the model to evaluate the importance of the sum of the area ratios of different electricity consumption behavior patterns as the explanatory variable. According to the evaluation results, rank the influence degrees of different electricity consumption behavior patterns on the total electricity consumption of the external unit, and determine which electricity consumption behavior patterns have a greater impact on the total electricity consumption of the external unit when their area ratios change, and which have a smaller impact. This ranking can clearly show the relative status of different electricity consumption behavior patterns in terms of energy consumption impact.

[0098] According to the order of the said influence degrees, assign corresponding electricity consumption influence coefficients to different electricity consumption behavior patterns:

[0099] Based on the previously obtained ranking of the influence degrees of different electricity consumption behavior patterns, assign a corresponding electricity consumption influence coefficient to each electricity consumption behavior pattern. The electricity consumption behavior pattern with a greater influence degree is assigned a higher electricity consumption influence coefficient. The electricity consumption influence coefficient will be used for subsequent analysis and calculation of the electricity consumption of the extensions, so as to more accurately reflect the contribution of the actual electricity consumption of the extensions to the energy consumption of the entire system.

[0100] Although the present invention has been specifically shown and described in connection with preferred embodiments, those skilled in the art should understand that various changes in form and detail may be made to the present invention without departing from the spirit and scope of the present invention as defined by the appended claims, and all such changes are within the scope of protection of the present invention.

Claims

1. A method for statistical calculation of the power consumption of a cascaded air conditioner, characterized in that, It includes the following steps: Obtain the operation data of the target extension unit of the cascaded air conditioner within a specific time period and the total power consumption of the outdoor unit during the corresponding time period; Obtain the electricity consumption behavior pattern of the target extension unit according to the operation data; Determine the electricity consumption influence coefficient of the target extension unit according to the electricity consumption behavior pattern and the area proportion of the region where the target extension unit is located; the electricity consumption influence coefficient is determined by the influence degree of the electricity consumption behavior pattern on the total power consumption of the outdoor unit; Determine the electricity consumption weight of the target extension unit according to the proportion of the electricity consumption influence coefficient of the target extension unit in the sum of the electricity consumption influence coefficients of all extension units; Obtain the electricity consumption of the target extension unit according to the product of the electricity consumption weight and the total power consumption of the outdoor unit.

2. The method for statistically calculating the power consumption of a cascaded air conditioner according to claim 1, characterized in that, The operation data includes operation time, set temperature, wind speed, and operation mode.

3. A method for statistical analysis of the power consumption of a cascaded air conditioner according to claim 1, characterized in that, The obtaining of the electricity consumption behavior pattern of the target extension unit according to the operation data is specifically to input the operation data into a pre-trained deep learning model, and output the electricity consumption behavior pattern of the target extension unit. The deep learning model is configured as a PCA model and a SOM model. The PCA model extracts the operation data features through dimensionality reduction, and the SOM model defines the electricity consumption behavior pattern by clustering the operation data features.

4. A method for statistical calculation of the power consumption of a cascaded air conditioner according to claim 3, characterized in that, The PCA model includes the following steps: Collect the operation data samples of the extension units in the historical period; Convert all the operation data samples into a data matrix in numerical form; Obtain several principal components by dimensionality reduction of the data matrix; Obtain the linear expression of the principal components with respect to the operation data for extracting the operation data features.

5. A method for statistically calculating the power consumption of a cascaded air conditioner according to claim 3, characterized in that, The SOM model includes the following steps: Initialize the weight vectors of the two-dimensional grid neurons; Input the operation data features into the network, and find the most matching neuron by calculating the Euclidean distance; Then adjust the weights of the most matching neuron and its surrounding neurons according to the Gaussian neighborhood function with the most matching neuron as the center; Make the weights converge to form a stable mapping through multiple iterations; Cluster adjacent neurons to form a clustering cluster, and define the electricity consumption pattern of the clustering cluster.

6. The method for statistically calculating the power consumption of a cascaded air conditioner according to claim 1, characterized in that, The influence degree of the electricity consumption behavior pattern on the total power consumption of the outdoor unit is determined through the following steps: Collect the electricity consumption behavior pattern and the area proportion data of all extension units of the cascaded air conditioner in the historical period, and the total power consumption of the outdoor unit during the corresponding time period; Statistical sum of the area proportions of different electricity consumption behavior patterns; Construct a random forest model with the sum of the area proportions of different electricity consumption behavior patterns as the explanatory variable and the total power consumption of the outdoor unit as the response variable; Rank the influence degrees of different electricity consumption behavior patterns through the random forest model; Assign corresponding electricity consumption influence coefficients to different electricity consumption behavior patterns according to the order of the influence degrees.

7. A method for cascaded air conditioner power consumption statistics according to claim 1, characterized in that, The electricity consumption weight is calculated as follows: where D i is the electricity consumption weight of the i-th extension; S i is the area proportion of the region where the i-th extension is located; X i is the electricity consumption influence coefficient of the i-th extension; X sum is the total sum of the electricity consumption influence coefficients of all extensions.

8. A method for statistical calculation of the power consumption of a cascaded air conditioner according to claim 1, characterized in that, The electricity consumption of the extension unit of the cascaded air conditioner is calculated as follows: F i = D i × Z d , Where, F i is the power consumption of the i-th extension; D i is the power consumption weight of the i-th extension; Z d is the total power consumption of the outdoor unit.

9. The method for statistically calculating the power consumption of a cascaded air conditioner according to claim 1, characterized in that, When the electricity consumption of the target extension unit exceeds the preset electricity consumption threshold, send a reminder message to the user of the corresponding extension unit.

Images