Energy efficiency evaluation model construction method and device, terminal and storage medium

Abstract

The present application relates to the technical field of energy, in particular to a kind of energy efficiency evaluation model construction method, device, terminal and storage medium, the present application method is first obtained the classification of evaluation target, then, according to classification, determine the relevant factors of evaluation target, then, according to relevant factors, obtain multiple data samples. Again, input multiple data samples to initial model, train initial model until the output of initial model and energy efficiency deviation is less than threshold value. Finally, the parameters of initial model are fixed as energy efficiency evaluation model. The whole method first determines the classification of target subject, determines relevant factors based on classification, so that it is not necessary to obtain those parameters with low correlation, reduce the complexity of model construction, reduce the influence of irrelevant factors on model construction, construct, train model through existing data, its construction mode is simple, and the energy efficiency accuracy of model output is high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy, in particular to an energy efficiency evaluation model construction method and device, a terminal and a storage medium. BACKGROUND

[0002] Energy saving and efficiency improvement and green development have become an important task of national economic and social development in the world, a key measure to cultivate and develop new energy and enhance green competitiveness, and an important support for healthy and sustainable economic and social development and improvement of ecological environment quality. There is an increasing call to help society save energy, protect the environment, and develop a green and ecological environment, and to reduce the overall carbon emissions and carbon emission intensity. It is an important part of the energy strategy to firmly curb the blind development of high-energy consumption, high-emission and low-level projects, promote the transformation of energy consumption to carbon emission total and intensity, improve pollution reduction and carbon reduction incentives and constraints, and accelerate the formation of a green way of life.

[0003] In actual operation, energy efficiency analysis work has problems such as single type of metering in key energy-using units, limited analysis ability of ordinary managers when using manual analysis, and insufficient professional diagnostic personnel, resulting in quality problems such as missing and stacking of energy efficiency data, inability to accurately and timely dig hidden information from the data, difficulty in accurately and timely discovering high-energy consumption points, inability to provide constructive suggestions for energy saving, failure to play the effect of minute-level collection of the electricity information collection system, and finally inability to meet the effect of accurately providing energy saving and consumption reduction schemes for users.

[0004] At present, the existing energy efficiency analysis method for key energy efficiency users has the following problems:

[0005] 1. For example, the traditional nearest neighbor clustering algorithm is more suitable for long-term existing faults (calculated in weeks or months), and after using cosine similarity and Euclidean distance similarity, this method still focuses more on finding fault causes, and is only effective for single faults, and the automatic diagnosis capability is slightly insufficient.

[0006] 2. Many studies have not been verified by actual data, and there is also a phenomenon that researchers only analyze measured energy efficiency data manually.

[0007] 3. The relatively systematic evaluation index for key energy-consuming users is only available for air conditioners, and the evaluation of other energy-using systems and equipment is not perfect (meaning that there is not enough perfect evaluation standard in technology).

[0008] Therefore, it is necessary to develop and design an energy efficiency evaluation model construction method to solve the problem of inaccurate energy efficiency evaluation model construction in the prior art, which leads to waste of energy resources. SUMMARY

[0009] The embodiment of the present application provides an energy efficiency evaluation model construction method, device, terminal and storage medium, and is used for solving the problem that the energy efficiency evaluation model is not accurate in the prior art.

[0010] In a first aspect, the embodiment of the present application provides an energy efficiency evaluation model construction method, comprising:

[0011] obtaining a classification of an evaluation target, wherein the classification is used to represent a category to which the evaluation target belongs;

[0012] determining a related factor of the evaluation target according to the classification, wherein the related factor is related to the energy efficiency of the evaluation target;

[0013] obtaining a plurality of data samples according to the related factor, wherein the data samples are constructed based on the related factor and the energy efficiency of the evaluation target;

[0014] inputting the plurality of data samples into an initial model, and training the initial model until an output of the initial model deviates from the energy efficiency by less than a threshold value;

[0015] fixing parameters of the initial model as an energy efficiency evaluation model.

[0016] In a possible implementation manner, the related factor is determined through the following steps:

[0017] obtaining a plurality of factor sets, wherein each factor set comprises a plurality of factor pairs composed of a factor of the evaluation target and the energy efficiency of the evaluation target;

[0018] for each factor set, determining a correlation coefficient of the factor and the energy efficiency according to a first formula, wherein the first formula is:

[0019]

[0020] wherein, ρ is the correlation coefficient, cov(X, Y) is the covariance, σ x is the standard deviation of the factor, and σ y is the standard deviation of the energy efficiency;

[0021] the factor with an absolute value of the correlation coefficient higher than a threshold value is taken as the related factor.

[0022] In a possible implementation manner, the related factor is determined through the following steps:

[0023] obtaining a plurality of factor sets, wherein each factor set comprises a plurality of factor pairs composed of a factor of the evaluation target and the energy efficiency of the evaluation target;

[0024] for each factor set, determining a correlation coefficient of the factor and the energy efficiency according to a second formula, wherein the second formula is:

[0025]

[0026] wherein, ρ is a correlation coefficient, n is a number of factor pairs, d i is a ranking difference of factors and energy efficiencies in the factor pair, wherein the ranking difference is a difference between rankings of the factors and the energy efficiencies after the factors and the energy efficiencies are ranked according to the same ranking manner respectively;

[0027] factors with the correlation coefficient absolute value higher than a threshold value are taken as the correlation factors.

[0028] In a possible implementation, the obtaining of the plurality of data samples according to the correlation factors comprises:

[0029] a plurality of evaluation parameter sets are obtained, wherein the evaluation parameter sets are obtained based on an evaluation target, and each evaluation parameter set comprises an energy efficiency and a parameter obtained according to the correlation factors;

[0030] a preset number of parameter sets are randomly selected as initial centers;

[0031] a classification step: for each evaluation parameter set, a distance to each initial center is calculated, and a class in which the initial center with the minimum distance is located is added;

[0032] a center of each class is calculated as a clustering center;

[0033] if a distance between the clustering center and the initial center is greater than a threshold value, the clustering center is taken as an initial center, and the classification step is jumped to;

[0034] otherwise, the clustering center is taken as a data sample.

[0035] In a possible implementation, the calculation of the distance to each initial center for each evaluation parameter set comprises:

[0036] for each evaluation parameter set, a distance to each initial center is calculated according to a third formula, and the third formula is:

[0037]

[0038] wherein, d(x, y) is a distance between the evaluation parameter set and the initial center, m is a number of parameters, x i is a parameter of the evaluation parameter set, y i is a parameter of the initial center, ρ i is a correlation coefficient of the parameter corresponding to the correlation factor, x m+1 is an energy efficiency of the evaluation parameter set, y m+1 is an energy efficiency of the initial center, ρ m+1 is a maximum limit of a value of the correlation coefficient calculation formula.

[0039] In a possible implementation, the initial model is constructed based on a neural network model, including:

[0040] an input layer, a fully connected intermediate layer, and an output layer;

[0041] wherein an activation function of the intermediate layer is:

[0042]

[0043] wherein f(h) is the activation function, a is a constant greater than 0 and less than 1, and h is an input of the activation function.

[0044] In a possible implementation, the inputting of the plurality of data samples into the initial model, the training of the initial model until the output of the initial model and the energy efficiency deviation are less than a threshold value, includes:

[0045] preprocessing the data samples to obtain input variables and labels, wherein the preprocessing is used to adjust weights of each parameter in the data samples, and the labels are used to represent energy efficiencies of the corresponding parameters;

[0046] an input step: inputting the plurality of input variables into the initial model to obtain an output of the initial model;

[0047] determining a residual error according to the output and the labels;

[0048] if the residual error is greater than a threshold value, using a back propagation algorithm to update weights of each node in the initial model and jumping to the input step;

[0049] wherein the preprocessing of the data samples includes:

[0050] processing each data sample as follows:

[0051] taking an energy efficiency of the data sample as a label of the data sample;

[0052] generating the input variables according to each parameter of the data sample and a fourth formula, the fourth formula being:

[0053] x′=x

[0054] wherein x′ is the input variable, ρ is a correlation coefficient of a parameter corresponding to a relevant factor, and x is the parameter.

[0055] In a second aspect, an embodiment of the present application provides an energy efficiency evaluation model construction device, including:

[0056] a category determination module configured to obtain a category of an evaluation target, the category being used to represent a category to which the evaluation target belongs.

[0057] a correlation factor determining module configured to determine a correlation factor of the evaluation target according to the classification, wherein the correlation factor is related to the energy efficiency of the evaluation target;

[0058] a data sample obtaining module configured to obtain a plurality of data samples according to the correlation factor, wherein the data samples are constructed based on the correlation factor and the energy efficiency of the evaluation target;

[0059] a model training module configured to input the plurality of data samples into an initial model, and train the initial model until an output of the initial model deviates from the energy efficiency by less than a threshold value;

[0060] and,

[0061] a parameter fixing module configured to fix parameters of the initial model as an energy efficiency evaluation model.

[0062] In a third aspect, an embodiment of the present application provides a terminal, comprising a memory and a processor, the memory stores a computer program capable of running on the processor, and the processor implements steps of the method according to the first aspect or any possible implementation manner of the first aspect when executing the computer program.

[0063] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements steps of the method according to the first aspect or any possible implementation manner of the first aspect when executed by a processor.

[0064] Compared with the prior art, the embodiment of the present application has the following beneficial effects:

[0065] The energy efficiency evaluation model construction method disclosed by the embodiment of the present application firstly obtains the classification of the evaluation target, then determines the correlation factor of the evaluation target according to the classification, and then obtains a plurality of data samples according to the correlation factor. Then, the plurality of data samples are input into an initial model, and the initial model is trained until the output of the initial model deviates from the energy efficiency by less than a threshold value. Finally, the parameters of the initial model are fixed as an energy efficiency evaluation model. The whole method firstly determines the classification of the target subject, and determines the correlation factor based on the classification. Therefore, it is not necessary to obtain parameters with low correlation, thereby reducing the complexity of model construction, reducing the influence of irrelevant factors on model construction, constructing the model based on the correlation factor, and adjusting the model parameters through the correlation factor. Finally, the model construction method can not concern the relationship between the correlation factor and the energy efficiency, but can construct a complex logical relationship. The model is constructed and trained based on the existing data, the construction method is simple, and the energy efficiency output by the model is accurate. BRIEF DESCRIPTION OF DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings.

[0067] Figure 1 is a flow chart of the energy efficiency evaluation model construction method provided by the embodiments of the present application.

[0068] Figure 2 is a functional block diagram of the energy efficiency evaluation model construction device provided by the embodiments of the present application.

[0069] Figure 3 is a terminal functional block diagram provided by the embodiments of the present application. DETAILED DESCRIPTION

[0070] In the following description, specific details are set forth in order to provide a thorough understanding of the embodiments of the present application. However, persons skilled in the art will understand that the present application can be practiced without these specific details. In other instances, well-known systems, devices and methods have not been described in detail in order to avoid obscuring the present application.

[0071] In order to make the objects, technical solutions and advantages of the present application clearer, the following will be described in detail by specific embodiments in conjunction with the drawings.

[0072] The embodiments of the present application will be described in detail below. The present example is implemented on the premise of the technical solutions of the present application, and detailed embodiments and specific operation processes are given. However, the protection scope of the present application is not limited to the following embodiments.

[0073] Figure 1 is a flow chart of the energy efficiency evaluation model construction method provided by the embodiments of the present application.

[0074] As shown in Figure 1 , it shows the implementation flow chart of the energy efficiency evaluation model construction method provided by the embodiments of the present application, which is described in detail as follows:

[0075] In step 101, the classification of the evaluation target is obtained, which is used to represent the category to which the evaluation target belongs.

[0076] In step 102, according to the classification, the relevant factors of the evaluation target are determined, wherein the relevant factors are related to the energy efficiency of the evaluation target.

[0077] In some embodiments, the related factors are determined by the following steps:

[0078] obtaining a plurality of factor sets, wherein each factor set comprises a plurality of factor pairs consisting of a factor of the evaluation target and an energy efficiency of the evaluation target;

[0079] for each factor set, determining a correlation coefficient of the factor and the energy efficiency according to a first formula, wherein the first formula is:

[0080]

[0081] wherein p is the correlation coefficient, cov(X, Y) is the covariance, σ x is the standard deviation of the factor, and σ y is the standard deviation of the energy efficiency;

[0082] factors with correlation coefficients having absolute values higher than a threshold value are taken as the related factors.

[0083] In some embodiments, the related factors are determined by the following steps:

[0084] obtaining a plurality of factor sets, wherein each factor set comprises a plurality of factor pairs consisting of a factor of the evaluation target and an energy efficiency of the evaluation target;

[0085] for each factor set, determining a correlation coefficient of the factor and the energy efficiency according to a second formula, wherein the second formula is:

[0086]

[0087] wherein p is the correlation coefficient, n is the number of the factor pairs, d i is the ranking difference of the factor and the energy efficiency in the factor pair, wherein the ranking difference is the difference between the ranks of the factor and the energy efficiency after the factor and the energy efficiency are ranked in the same manner;

[0088] factors with correlation coefficients having absolute values higher than a threshold value are taken as the related factors.

[0089] For example, the energy efficiency is affected by a plurality of possible factors, and it is a huge workload to build an expression to express the relationship between the factors and the energy efficiency, and the accuracy is not high.

[0090] For example, for an electrical equipment, the energy efficiency thereof is affected by factors such as the size of each part of the electrical equipment and the material, and is also related to external conditions such as ambient temperature, load rate, and power quality. For different electrical equipment, it is obvious that it is difficult and complicated to build a model containing all known influencing factors, and the workload of data acquisition is also increased.

[0091] Therefore, it is necessary to obtain different influencing factors for different categories of targets, and to construct models according to the influencing factors, which can obviously reduce the complexity of the models and the types and quantities of data obtained.

[0092] For example, for refrigeration equipment, the energy efficiency is negatively related to the ambient temperature, and the correlation is strong. The lower the ambient temperature, the better the refrigeration effect, and the higher the energy efficiency. For motor products, if the motor products work in a full load state, the efficiency is close to the designed efficiency, that is, the energy efficiency is mainly related to the load rate, and the influence of the environment on the energy efficiency is not so great.

[0093] Therefore, it is necessary to select different factors for different types of products. However, once the influencing factors are determined, different types of products can be constructed through a common template model, in which the parameters of the template are adjusted to realize the construction of different models according to the classification of products.

[0094] Most products have many existing experiences that can be used for reference in classification methods, for example, synchronous motors can be classified into DC motors, and electric heating equipment can be classified into electric heating equipment. For some new products, they basically belong to the unsupervised classification method, that is, according to various attributes of the product, similar points are found, and then the similar points are classified into the closest category.

[0095] After the classification is determined, an important step is to select the factors related to the category, that is, the factors that mainly affect the energy efficiency of the products in this category.

[0096] One embodiment is to use the Pearson correlation coefficient determination method, specifically, for a certain category of target, the energy efficiency and factor pairs of multiple samples are obtained, and then the standard deviation of the energy efficiency and the standard deviation of the factor, and the covariance of the energy efficiency and the factor are calculated according to the factor pairs, and then the correlation coefficient of the energy efficiency and the factor can be obtained according to the following formula:

[0097]

[0098] In the formula, ρ is the correlation coefficient, cov(X, Y) is the covariance, σ x is the standard deviation of the factor, and σ y is the standard deviation of the energy efficiency.

[0099] After obtaining the correlation coefficient of each factor, the factors with high correlation coefficients are retained as the related factors, and the factors with weak correlation or no correlation are discarded.

[0100] The Pearson correlation coefficient requires that the factors be quantified first, and also needs to meet certain conditions, for example, the data needs to meet the normal distribution, only after meeting the predetermined conditions, the correlation coefficient can be calculated by the above formula.

[0101] However, in fact, some factors are not easy to be quantified, for example, the degree of unpleasantness of smell, people can divide it into several levels from the most willing to accept to the most difficult to accept, but it cannot be quantified, at this time, another statistical correlation calculation method needs to be used, that is, the Spearman correlation coefficient determination method, in terms of specific operation, the factors and energy efficiency are arranged in a predetermined order, for example, from small to large, because the factors and energy efficiency constitute a data pair, then the sorting difference between the two in the data pair can be easily obtained, and the correlation coefficient can be determined according to the following formula.

[0102]

[0103] In the formula, p is the correlation coefficient, n is the number of factor pairs, d i is the sorting difference of the factors and energy efficiency in the factor pair, wherein the sorting difference is the difference in the ranking of the factors and energy efficiency after the factors and energy efficiency are sorted in the same sorting manner.

[0104] Therefore, after the above steps, the factors that have an impact on the energy efficiency level are determined.

[0105] In step 103, a plurality of data samples are obtained according to the correlation factors, wherein the data samples are constructed based on the correlation factors of the evaluation target and the energy efficiency.

[0106] In some embodiments, step 103 includes:

[0107] A plurality of evaluation parameter sets are obtained, wherein the evaluation parameter sets are obtained based on the evaluation target, and the evaluation parameter sets include the energy efficiency and the parameters obtained according to the plurality of correlation factors.

[0108] A preset number of parameter sets are randomly selected as initial centers.

[0109] Classification step: for each evaluation parameter set, the distance from each initial center is calculated, and the class in which the initial center with the smallest distance is added.

[0110] The center of each class is calculated as a clustering center.

[0111] If the distance between the clustering center and the initial center is greater than a threshold value, the clustering center is taken as the initial center, and the classification step is jumped to.

[0112] Otherwise, the clustering center is taken as a data sample.

[0113] In some embodiments, the calculating the distance between each evaluation parameter set and each initial center comprises:

[0114] The distance between each evaluation parameter set and each initial center is calculated according to a third formula, which is:

[0115]

[0116] In the formula, d(x, y) is the distance between the evaluation parameter set and the initial center, m is the number of parameters, x i is a parameter of the evaluation parameter set, y i is a parameter of the initial center, p i is the correlation coefficient of the parameter corresponding to the relevant factor, x m+1 is the energy efficiency of the evaluation parameter set, y m+1 is the energy efficiency of the initial center, p m+1 is the maximum limit of the value of the correlation coefficient calculation formula.

[0117] Exemplarily, we always expect to obtain enough samples to ensure that the accuracy of the finally determined model is high enough. However, the more data, the more data processing process troubles it brings, so it is necessary to process the data to retain the accuracy of the original data with as little data as possible.

[0118] A feasible way is to use a clustering algorithm to obtain clustering centers of multiple samples, and then construct a model based on the clustering centers. Obviously, the clustering center can represent the data around the clustering center, but the total amount of data is reduced.

[0119] One clustering algorithm is to first determine the number of clusters after clustering, that is, the total amount of data expected to be reduced. For example, the total amount of data before clustering is k, and if the preset number of clustering centers is j (obviously j is less than k), then the amount of data to be reduced is k-j. The fewer the number of clustering centers, the more data is reduced.

[0120] After obtaining the amount of data to be reduced, randomly extract parameter sets as initial centers according to the amount of data to be reduced, for example, extract j parameter sets as initial centers. Then, classify each parameter set into the nearest one of the j classes, respectively. After the classes are divided, calculate the center of the class as the clustering center. If the distance between the initial center and the clustering center is greater than a preset value, the clustering center is taken as the initial center, and the classification and clustering center calculation operations are performed again until the distance between the clustering center and the initial center is less than the preset value. At this time, the clustering center can be used as a sample of the data.

[0121] In the way of calculating the distance between the parameter set and the initial center, one calculation method is as follows:

[0122]

[0123] where d(x, y) is the distance between the evaluation parameter set and the initial center, m is the number of parameters, x i is a parameter of the evaluation parameter set, y i is a parameter of the initial center, p i is the correlation coefficient of the parameter corresponding to the relevant factor, x m+1 is the energy efficiency of the evaluation parameter set, y m+1 is the energy efficiency of the initial center, p m+1 is the maximum limit of the correlation coefficient calculation formula.

[0124] We can see that this calculation formula first has a weight term-correlation coefficient. For parameters with large absolute values of correlation coefficient, the weight of the calculated distance is also larger, which can better reflect the distance factor of important parameters. It is worth noting that the energy efficiency is also part of the parameter set, and the weight is the maximum value of the correlation coefficient calculation formula when calculating the distance, for example, in the case of using the Spearman correlation coefficient calculation formula, the value is 1.

[0125] Through the above steps, we have reduced the amount of data while maintaining sufficient data accuracy.

[0126] In step 104, the plurality of data samples are input into an initial model, and the initial model is trained until the output of the initial model deviates from the energy efficiency by less than a threshold value.

[0127] In some embodiments, the initial model is constructed based on a neural network model, comprising:

[0128] an input layer, a fully connected intermediate layer, and an output layer;

[0129] wherein the activation function of the intermediate layer is:

[0130]

[0131] where f(h) is the activation function, a is a constant greater than 0 and less than 1, and h is the input of the activation function.

[0132] In some embodiments, step 104 comprises:

[0133] preprocessing the data samples to obtain input variables and labels, wherein the preprocessing is used to adjust the weight of each parameter in the data sample, and the labels are used to represent the energy efficiency of the corresponding parameters.

[0134] Input step: inputting a plurality of input variables into an initial model to obtain the output of the initial model.

[0135] According to the output and the label, a residual error is determined.

[0136] If the residual error is greater than a threshold value, a back propagation algorithm is used to update the weight parameters of each node in the initial model and jump to the input step.

[0137] The pre-processing of the data samples includes:

[0138] Each data sample is processed as follows:

[0139] The energy efficiency of the data sample is used as the label of the data sample.

[0140] The input variable is generated according to each parameter of the data sample and a fourth formula, the fourth formula being:

[0141] x' = px

[0142] In the formula, x' is the input variable, p is the correlation coefficient of the parameter corresponding to the relevant factor, and x is the parameter.

[0143] Exemplarily, one embodiment of the initial model is to use a neural network structure. In some application scenarios, a BP neural network model is used, wherein the activation function sufficiently considers the possibility of gradient disappearance in back propagation, and the activation function uses:

[0144]

[0145] In the formula, f(h) is the activation function, a is a constant greater than 0 and less than 1, and h is the input of the activation function.

[0146] In the model training, first, the input variable and the label are made through the pre-processing of the data sample. Then, the input variable is input to the initial model to obtain the output of the initial model. According to the residual error determined by the output and the label, the weight parameters of each neuron node of the initial model are updated through the back propagation algorithm. Then, the input variable is input to the initial model again. If the residual error is less than a preset value, it can be considered that the iteration is completed.

[0147] In the making of the input variable and the label, one embodiment is to use the energy efficiency as the label and the product of the parameter and the correlation coefficient of the parameter corresponding to the relevant factor as the input variable, that is, expressed by the following formula:

[0148] x' = px

[0149] In the formula, x' is the input variable, p is the correlation coefficient of the parameter corresponding to the relevant factor, and x is the parameter.

[0150] In step 105, the parameters of the initial model are fixed as the energy efficiency evaluation model.

[0151] Exemplarily, after the iteration, the output of the initial model deviates from the label by less than a threshold, indicating that the initial model training is completed, and the model construction is completed.

[0152] The energy efficiency evaluation model construction method embodiment of the present application first acquires the classification of the evaluation target, which is used to represent the category to which the evaluation target belongs. Then, according to the classification, the relevant factors of the evaluation target are determined. Next, according to the relevant factors, a plurality of data samples are acquired. Then, the plurality of data samples are input into an initial model, and the initial model is trained until the output of the initial model deviates from the energy efficiency by less than a threshold. Finally, the parameters of the initial model are fixed as the energy efficiency evaluation model. The whole method first determines the classification of the target subject, and determines the relevant factors based on the classification. Therefore, it is not necessary to acquire parameters with low relevance, reducing the complexity of model construction and reducing the influence of irrelevant factors on model construction. Based on the relevant factors, the model is constructed, and the model parameters are trained and adjusted, and finally the model construction is realized. The way of constructing the model can not concern the relationship between the relevant factors and the energy efficiency, but a complex logical relationship can be constructed. The model is constructed and trained based on the existing data, and the construction method is simple, and the energy efficiency output by the model is accurate.

[0153] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0154] The following is the device embodiment of the present application. For details not described in detail, please refer to the corresponding method embodiments described above.

[0155] Figure 2 is the functional block diagram of the energy efficiency evaluation model construction device provided by the embodiment of the present application. Referring to Figure 2 , the energy efficiency evaluation model construction device 2 comprises a category determination module 201, a relevant factor determination module 202, a data sample acquisition module 203, a model training module 204 and a shaping module 205.

[0156] The category determination module 201 is used to acquire the classification of the evaluation target, and the classification is used to represent the category to which the evaluation target belongs.

[0157] The relevant factor determination module 202 is used to determine the relevant factors of the evaluation target according to the classification, wherein the relevant factors are related to the energy efficiency of the evaluation target.

[0158] The data sample acquisition module 203 is used to acquire a plurality of data samples according to the relevant factors, wherein the data samples are constructed based on the relevant factors and the energy efficiency of the evaluation target.

[0159] A model training module 204 is configured to input the plurality of data samples into an initial model and train the initial model until a deviation between an output of the initial model and the energy efficiency is less than a threshold;

[0160] The finalization module 205 is used to fix the parameters of the initial model as an energy efficiency evaluation model.

[0161] Figure 3 This is a functional block diagram of a terminal provided by an embodiment of the present invention. Figure 3 As shown, the terminal 3 of this embodiment includes: a processor 300 and a memory 301, wherein the memory 301 stores a computer program 302 that can be run on the processor 300. When the processor 300 executes the computer program 302, the steps in the above-mentioned energy efficiency evaluation model construction method and embodiment are implemented, for example Figure 1 Steps 101 to 105 are shown.

[0162] Illustratively, the computer program 302 may be divided into one or more modules / units, and the one or more modules / units are stored in the memory 301 and executed by the processor 300 to implement the present invention.

[0163] The terminal 3 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal 3 may include, but is not limited to, a processor 300 and a memory 301. Those skilled in the art will understand that Figure 3 It is only an example of terminal 3 and does not constitute a limitation on terminal 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal may also include input and output devices, network access devices, buses, etc.

[0164] The processor 300 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0165] The memory 301 can be an internal storage unit of the terminal 3, such as a hard disk or a memory of the terminal 3. The memory 301 can also be an external storage device of the terminal 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD), a flash card, and the like equipped on the terminal 3. Further, the memory 301 can also include both the internal storage unit and the external storage device of the terminal 3. The memory 301 is used to store the computer program and other programs and data required by the terminal. The memory 301 can also be used to temporarily store data that has been output or is to be output.

[0166] It can be clearly understood by those skilled in the art that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0167] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can refer to the relevant description of other embodiments.

[0168] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0169] In the embodiments of the present application, it should be understood that the disclosed apparatus / terminal and method can be implemented in other manners. For example, the apparatus / terminal embodiments described above are merely schematic, and the division of the modules or units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0170] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0171] In addition, each functional unit in the various embodiments of the present application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0172] The integrated module / unit, if implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-described embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of the above-described various energy efficiency evaluation model construction methods and energy efficiency evaluation model construction device embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0173] The above-described embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent replacements; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for constructing an energy efficiency evaluation model, characterized in that: include: Obtaining a classification of an evaluation target, wherein the classification is used to characterize the category to which the evaluation target belongs; Determining relevant factors of the evaluation target based on the classification, wherein the relevant factors are related to the energy efficiency of the evaluation target; According to the relevant factors, a plurality of data samples are obtained, wherein the data samples are constructed based on the relevant factors of the evaluation target and energy efficiency; Inputting the plurality of data samples into an initial model, and training the initial model until a deviation between an output of the initial model and the energy efficiency is less than a threshold, wherein the initial model is constructed based on a neural network model, including: an input layer, a fully connected intermediate layer, and an output layer; Fixing the parameters of the initial model as an energy efficiency evaluation model; Wherein, obtaining multiple data samples according to the relevant factors includes: Acquire multiple evaluation parameter sets, wherein the evaluation parameter sets are acquired based on the evaluation target, and the evaluation parameter sets include energy efficiency and parameters acquired based on the multiple related factors; Randomly select a preset number of parameter sets as initial centers; Classification step: For each evaluation parameter set, calculate the distance from each initial center according to the third formula, and add the class where the initial center with the smallest distance is located. The third formula is: Where d(x,y) is the distance between the evaluation parameter set and the initial center, m is the number of parameters, and x i is the parameter of the evaluation parameter set, y i is the parameter of the initial center, ρ i is the correlation coefficient of the parameter corresponding to the relevant factor, x m+1 To evaluate the energy efficiency of the parameter set, y m+1 is the energy efficiency of the initial center, ρ m+1 The maximum limit of the value of the correlation coefficient calculation formula; Calculate the center of each class as the cluster center; If the distance between the cluster center and the initial center is greater than a threshold, the cluster center is used as the initial center and the process jumps to the classification step; Otherwise, the cluster center is used as the data sample.

2. The energy efficiency evaluation model construction method according to claim 1, characterized in that: The relevant factors are determined by the following steps: Obtaining a plurality of factor sets, wherein the factor set includes a plurality of factor pairs consisting of factors of an evaluation target and energy efficiency of the evaluation target; For each factor set, the correlation coefficient between the factor and energy efficiency is determined according to the first formula, wherein the first formula is: Where ρ is the correlation coefficient, cov(X,Y) is the covariance, σ x is the standard deviation of the factor, σ y is the standard deviation of energy efficiency; The factors whose absolute values ​​of the correlation coefficients are higher than the threshold are regarded as correlation factors.

3. The energy efficiency evaluation model construction method according to claim 1, characterized in that: The relevant factors are determined by the following steps: Obtaining a plurality of factor sets, wherein the factor set includes a plurality of factor pairs consisting of factors of an evaluation target and energy efficiency of the evaluation target; For each factor set, the correlation coefficient between the factor and energy efficiency is determined according to the second formula, wherein the second formula is: Where ρ is the correlation coefficient, n is the number of factor pairs, and d i is the ranking difference between the factor and energy efficiency in the factor pair, wherein the ranking difference is the ranking difference between the factor and energy efficiency after the factor and energy efficiency are sorted in the same sorting manner; The factors whose absolute values ​​of the correlation coefficients are higher than the threshold are regarded as correlation factors.

4. The method for constructing an energy efficiency evaluation model according to any one of claims 1 to 3, characterized in that: The activation function of the middle layer of the initial model is: Where f(h) is the activation function, a is a constant greater than 0 and less than 1, and h is the input of the activation function.

5. The energy efficiency evaluation model construction method according to claim 1, characterized in that: Inputting the plurality of data samples into an initial model and training the initial model until a deviation between an output of the initial model and the energy efficiency is less than a threshold value includes: Preprocessing the data samples to obtain input variables and labels, wherein the preprocessing is used to adjust the weights of various parameters in the data samples, and the labels are used to characterize the energy efficiency of the corresponding parameters; Input step: inputting a plurality of the input variables into the initial model to obtain the output of the initial model; Determining a residual based on the output and the label; If the residual is greater than a threshold, a back propagation algorithm is used to update the weight parameters of each node in the initial model and jump to the input step; The preprocessing of the data sample includes: Each of the data samples is processed as follows: The energy efficiency of the data sample is used as the label of the data sample; The input variable is generated according to each parameter of the data sample and the fourth formula, and the fourth formula is: x′=ρx In the formula, x′ is the input variable, ρ is the correlation coefficient of the parameter corresponding to the relevant factor, and x is the parameter.

6. An energy efficiency evaluation model construction device, characterized in that: For implementing the energy efficiency evaluation model construction method according to any one of claims 1 to 5, the energy efficiency evaluation model construction device comprises: A category determination module, used to obtain a category of an evaluation target, where the category is used to characterize the category to which the evaluation target belongs; a relevant factor determination module, configured to determine relevant factors of the evaluation target based on the classification, wherein the relevant factors are related to the energy efficiency of the evaluation target; A data sample acquisition module, configured to acquire a plurality of data samples according to the relevant factors, wherein the data samples are constructed based on the relevant factors of the evaluation target and energy efficiency; a model training module, configured to input the plurality of data samples into an initial model and train the initial model until a deviation between an output of the initial model and the energy efficiency is less than a threshold; as well as, The finalization module is used to fix the parameters of the initial model as an energy efficiency evaluation model.

7. A terminal comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images