Production enterprise-oriented carbon emission prediction method and system

Abstract

The invention belongs to the technical field of carbon emission prediction, and particularly relates to a production enterprise-oriented carbon emission prediction method and system.The method comprises the steps that S1, a data preparation module sends basic data of a current time point to a data decision module, the data decision module continues to obtain basic data of different historical time points before the current time point from the data preparation module; s2, based on the basic data of the current time point and the basic data of different historical time points before the current time point, the data decision module judges whether the basic data of the current time point is used for prediction processing or not; and S3, under the condition that the judgment result of the data decision module is that the basic data of the current time point is used, the prediction module executes prediction processing according to the basic data of the current time point, and otherwise, the prediction module executes prediction processing according to the basic data of the historical time point before the current time point. The prediction accuracy can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of carbon emission prediction, and in particular relates to a carbon emission prediction method and system for production enterprises. Background Art

[0002] At present, manufacturing enterprises are constantly pursuing carbon peak and carbon neutrality goals, and are also placing increasingly high demands on carbon emissions prediction methods for manufacturing enterprises. Therefore, it becomes particularly important to propose a carbon emissions prediction method for manufacturing enterprises with high prediction accuracy.

[0003] The Chinese invention application with publication number CN114066097A discloses a carbon emission prediction method, including: obtaining the annual average growth rate of turnover of different modes of transportation in the target area, the annual average growth rate of transportation route construction mileage and the annual average change rate of carbon emissions, as well as obtaining the annual average growth rate of GDP and the annual average growth rate of permanent population in the target area; obtaining the predicted annual average growth rate of turnover of different modes of transportation in the target area; obtaining the annual average energy utilization optimization rate of different existing energy sources in the target area; obtaining the annual average energy structure optimization rate of different new energy sources in the target area; obtaining the predicted annual average optimization rate of carbon emission intensity of different modes of transportation in the target area; obtaining the predicted carbon emissions of different modes of transportation in the target area in the target year; and obtaining the predicted development trend of carbon emissions in the target area. In addition, the Chinese invention application with publication number CN115829130A discloses a carbon emission prediction method, including the following steps: Step 1: Obtain the carbon emission data of the target area within the statistical time period and establish a corresponding database; Step 2: Calculate the growth rate of carbon emissions within the statistical time period based on the data statistics in the database, and use this growth rate as the initial predicted growth rate of carbon emissions, and use statistical accounting to predict the carbon emissions of the target area within the future statistical time period. When predicting, the factors affecting carbon emissions in related industries within the future statistical time period can be considered within the prediction range. However, the prediction accuracy of the prediction methods in the above invention applications needs to be improved. Summary of the invention

[0004] The present invention is based on the basic data at the current time point and the basic data at different historical time points before the current time point, and determines whether to use the basic data at the current time point for prediction processing. When the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data at the current time point. When the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data at the historical time points before the current time point. The purpose of the present invention is to improve the prediction accuracy.

[0005] In order to achieve the above-mentioned purpose of the invention, the present invention provides a carbon emission prediction method for a production enterprise as described below, which mainly includes the following steps:

[0006] S1, the data preparation module obtains basic data at the current time point, sends the basic data at the current time point to the data decision module, and the data decision module continues to obtain basic data at different historical time points before the current time point from the data preparation module;

[0007] S2, the data decision module determines whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point;

[0008] S3. When the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs the prediction processing according to the basic data at the current time point; when the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs the prediction processing according to the basic data of the historical time point before the current time point.

[0009] As a preferred technical solution of the present invention, in S1, the basic data includes a first basic data part and a second basic data part, wherein the first basic data part includes emission data between one time point, emission data before two time points, and emission data until M time points before, and the second basic data part includes emission data at the current time point and identification data of the next time point of the current time point.

[0010] As a preferred technical solution of the present invention, in S2, the prediction processing includes: the prediction module inputs the basic data into the trained prediction model, the prediction model outputs the prediction result of the emission data at the next time point of the current time point, the prediction module determines whether the prediction result is greater than or equal to a preset threshold, and if so, takes the next time point of the current time point as the final prediction result, and if not, the prediction module generates new basic data based on the prediction result, assuming that the next time point of the current time point has been reached, and repeats this step.

[0011] As a preferred technical solution of the present invention, in S2, the data decision module determines whether to use the basic data at the current time point for prediction processing, including the following steps:

[0012] S211, the data decision module obtains the current final prediction result, which is the final prediction result obtained by the prediction module using the basic data at the current time point to perform the prediction processing, and the data decision module continues to obtain the historical final prediction result, which is the final prediction result of the historical time point before the current time point stored by the data preparation module;

[0013] S212, the data decision module calculates the difference between the current final prediction result and the historical final prediction result, takes the absolute value of the difference as the deviation value of the current time point, and the data decision module continues to obtain the difference and deviation values ​​of different historical time points before the current time point from the data preparation module;

[0014] S213, the data decision module determines whether the deviation value at the current time point is greater than the deviation value threshold, the deviation value threshold is the average value of the deviation values ​​at different historical time points before the current time point, if yes, it is determined not to use the basic data at the current time point, if not, it is determined to use the basic data at the current time point.

[0015] As a preferred technical solution of the present invention, the data decision module determines whether to use the basic data at the current time point for prediction processing, and the step S213 further includes: the data decision module determines the formula δ t >|β t-1 +β t-2 +…+β t-n Is it true, where δ t is the deviation value at the current time point t, β t-1 is the difference between the current time point t and the time point before it, β t-2 is the difference between the two time points before the current time point t, β t-n It is the difference between the n time points before the current time point t. If it is true, it is determined to use the basic data of the current time point. If it is not true, it is determined not to use the basic data of the current time point.

[0016] As a preferred technical solution of the present invention, the data decision module determines whether to use the basic data at the current time point for prediction processing, and the S213 further includes: the data decision module uses the formula σ=(χ m-1 -x s ) 2 +(χ m-2 -x s-1 ) 2 +…+(χ m-k -x s-k+1 ) 2 Calculate the first eigenvalue, where σ is the first eigenvalue, χ m-1 ,xm-2 … m-k are the difference between the time point before the current time point m, the difference between the two time points before the current time point m, and the difference between the k time points before the current time point m, respectively. s ,x s-1 … s-k+1 They are respectively the difference between the time point before the historical time point s, the difference between the two time points before the historical time point s, and so on until the difference between the k time points before the historical time point s. When the first eigenvalue is less than or equal to the preset first eigenvalue threshold, it is determined to use the basic data of the current time point; when the first eigenvalue is greater than the preset first eigenvalue threshold, it is determined not to use the basic data of the current time point.

[0017] As a preferred technical solution of the present invention, in S2, the data decision module determines whether to use the basic data at the current time point for prediction processing, and further includes the following steps:

[0018] S221, the data decision module uses the basic data at the current time point and the basic data at several historical time points before the current time point to form a basic data sequence μ h ,μ h-1 ,…,μ h-r , which are the basic data of the current time point h, the basic data of the historical time point h-1, and the basic data of the historical time point hr;

[0019] S222, the data decision module obtains basic data of other enterprises at several time points from the data preparation module to form a reference basic data sequence v e ,ν e+1 ,…,ν e+r , which are the basic data of other enterprises at time point e, the basic data of other enterprises at time point e+1, and so on until the basic data of other enterprises at time point e+r;

[0020] S223, the data decision module uses the formula Calculate the second eigenvalue, where ρ is the second eigenvalue, D is a function for calculating the distance between two basic data, and determine whether the second eigenvalue is less than a preset second eigenvalue threshold. If not, jump to S222. If yes, determine whether to use the basic data ν e+r The final prediction result obtained by performing the prediction processing is closer to the real result. If yes, it is determined to use the basic data at the current time point. If no, it is determined not to use the basic data at the current time point.

[0021] As a preferred technical solution of the present invention, the prediction model includes a first-layer prediction model and a second-layer prediction model, the first-layer prediction model includes parallel sub-model one and sub-model two, and the second-layer prediction model is composed of sub-model three, wherein sub-model one, sub-model two, and sub-model three are all models composed of an input layer, an intermediate layer, and an output layer.

[0022] As a preferred technical solution of the present invention, the prediction module is trained for the prediction model, comprising the following steps:

[0023] S231, the prediction module obtains a number of training data from the data preparation module, the training data consists of basic data and real emission data, and the prediction module initializes and sets model parameters in the prediction model;

[0024] S232, based on a plurality of training data, the prediction module selects a piece of training data, inputs the first basic data part of the basic data in the selected training data into the sub-model 1, inputs the second basic data part of the basic data in the selected training data into the sub-model 2, then performs a preset operation on the output result of the sub-model 1 and the output result of the sub-model 2, and then inputs the operation result into the sub-model 3, and the sub-model 3 outputs the predicted emission data;

[0025] S233, the prediction module calculates the deviation between the predicted emission data and the actual emission data in the selected training data, and based on the deviation first adjusts the model parameters in the sub-model three, and then adjusts the model parameters in the sub-model one and the sub-model two at the same time, and the prediction module determines whether to end the training. If yes, end all steps, if not, jump to S232.

[0026] The present invention also provides a carbon emission prediction system for production enterprises, which mainly includes the following modules:

[0027] The data preparation module is used to store the basic data at the current time point and the basic data at different historical time points before the current time point;

[0028] A data decision module is used to determine whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point;

[0029] The prediction module is used to perform prediction processing based on the basic data at the current time point when the judgment result of the data decision module is to use the basic data at the current time point, and to perform prediction processing based on the basic data at the historical time point before the current time point when the judgment result of the data decision module is not to use the basic data at the current time point.

[0030] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0031] In the present invention, first, the data preparation module obtains the basic data at the current time point, and sends the basic data at the current time point to the data decision module, and the data decision module continues to obtain the basic data of different historical time points before the current time point from the data preparation module. Secondly, the data decision module determines whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data of different historical time points before the current time point. Finally, when the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data at the current time point. When the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data of the historical time points before the current time point. The present invention determines whether to use the basic data at the current time point for prediction processing based on the impact that the basic data at the current time point will have on the final prediction result, which can significantly improve the prediction accuracy of emission data. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flowchart of the steps of a method for predicting carbon emissions for production enterprises according to the present invention;

[0033] Figure 2 This is a structural diagram of a carbon emission prediction system for manufacturing enterprises according to the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0035] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0036] The present invention provides Figure 1 The carbon emission prediction method for a production enterprise shown is mainly implemented by executing the following steps:

[0037] S1. The data preparation module obtains basic data at the current time point, sends the basic data at the current time point to the data decision module, and the data decision module continues to obtain basic data at different historical time points before the current time point from the data preparation module;

[0038] S2, the data decision module determines whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point;

[0039] S3. When the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs prediction processing based on the basic data at the current time point. When the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs prediction processing based on the basic data of the historical time points before the current time point.

[0040] Specifically, the main steps of the carbon emission prediction method are introduced here. In step S1, the data preparation module obtains the basic data of the current time point and sends the basic data of the current time point to the data decision module. While receiving the basic data of the current time point, the data decision module obtains the basic data of different historical time points before the current time point from the data preparation module. For example, if the current time point is 11 o'clock in the morning, then the different historical time points before the current time point refer to 10 o'clock in the morning, 9 o'clock in the morning, 8 o'clock in the morning, etc. In step S2, in order to ensure the accuracy of the prediction, the data decision module determines whether to use the basic data of the current time point for prediction processing based on the basic data of the current time point and the basic data of different historical time points before the current time point. The specific judgment method will be described below. In step S3, if the judgment result of the data decision module is to use the basic data of the current time point, then the prediction module performs prediction processing according to the basic data of the current time point. If the judgment result of the data decision module is not to use the basic data of the current time point, the prediction module performs prediction processing according to the basic data of the historical time point before the current time point. The historical time point before the current time point here refers to the previous time point of the current time point.

[0041] Furthermore, the basic data includes a first basic data part and a second basic data part, wherein the first basic data part includes emission data between one time point, emission data before two time points, and emission data until M time points before, and the second basic data part includes emission data at the current time point and identification data of the next time point after the current time point.

[0042] Specifically, the content of the basic data is introduced in detail here. The basic data is relative to a certain time point. The basic data includes a first basic data part and a second basic data part. The first basic data part includes emission data between one time point, emission data before two time points, and emission data before M time points. The value of M is set in the actual application scenario. Continuing with the above example, the first basic data part in the basic data at the current time point is the emission data at 10 am, the emission data at 9 am, the emission data at 8 am, etc. The second basic data part includes the emission data at the current time point and the identification data of the next time point of the current time point. The next time point refers to the time point at which emission data needs to be predicted. Continuing with the above example, the second basic data part in the basic data at the current time point is the emission data at 11 am and the identification data at 12 am. The identification data can identify whether the next time point is working time, because whether the next time point is working time will also affect the emission data at the next time point. Of course, the identification data can also identify other content. It should be noted that the emission data at the current time point is the cumulative emission data of the production enterprises up to the current time point. In addition, the above only gives an example of the basic data at the current time point. The basic data at the current time point corresponds to the current time point. Compared with the basic data at historical time points, the corresponding time points are different.

[0043] Furthermore, the prediction processing includes: the prediction module inputs the basic data into the trained prediction model, the prediction model outputs the prediction result of the emission data at the next time point of the current time point, the prediction module determines whether the prediction result is greater than or equal to a preset threshold, and if so, takes the next time point of the current time point as the final prediction result, and if not, the prediction module generates new basic data based on the prediction result, assuming that the next time point of the current time point has been reached, and repeats this step.

[0044] Specifically, here is an introduction to how the prediction process is performed. The prediction module inputs the basic data into the trained prediction model. The training process of the prediction model will be described below. The prediction model outputs the prediction result of the emission data at the next time point of the current time point. Continuing with the above example, the basic data at the current time point is input into the prediction model. The prediction model outputs the prediction result of the emission data at 12 o'clock in the morning. The prediction module determines whether the prediction result is greater than or equal to the preset threshold. The threshold is set according to the actual application situation. If so, the next time point of the current time point is regarded as the final prediction result. In other words, the final prediction result is actually the time point when the cumulative emission data of the production enterprise reaches the above threshold. For example, if the predicted result of the emission data at 12 am is greater than the threshold, then 12 am is taken as the final predicted result. If not, the prediction module generates new basic data based on the predicted result, assuming that the next time point of the current time point has been reached, and repeats this step. For example, if the predicted result of the emission data at 12 am is less than the above threshold, then the basic data for 12 am is generated, except that the second basic data part of the basic data at 12 am contains the predicted result of the emission data at 12 am. Then repeat the above method, that is, start from inputting the basic number at 12 am into the prediction model again and perform subsequent processing until the final prediction result can be obtained.

[0045] Further, in step S2, the data decision module determines whether to use the basic data at the current time point for prediction processing, including the following steps:

[0046] S211, the data decision module obtains the current final prediction result, which is the final prediction result obtained by the prediction module using the basic data at the current time point for prediction processing, and the data decision module continues to obtain the historical final prediction result, which is the final prediction result of the historical time point before the current time point stored by the data preparation module;

[0047] S212, the data decision module calculates the difference between the current final prediction result and the historical final prediction result, and takes the absolute value of the difference as the deviation value of the current time point, and the data decision module continues to obtain the difference and deviation values ​​of different historical time points before the current time point from the data preparation module;

[0048] S213, the data decision module determines whether the deviation value at the current time point is greater than the deviation value threshold. The deviation value threshold is the average value of the deviation values ​​at different historical time points before the current time point. If so, it is determined not to use the basic data at the current time point. If not, it is determined to use the basic data at the current time point.

[0049] Specifically, the first method for the data decision module to determine whether to use the basic data of the current time point is introduced here. In step S211, the data decision module obtains the current final prediction result from the data preparation module. The current final prediction result is actually the final prediction result obtained by the prediction module using the basic data of the current time point for prediction processing. At the same time, the data decision module continues to obtain the historical final prediction result from the data preparation module. The historical final prediction result is actually the final prediction result of the historical time point before the current time point. The historical time point before the current time point refers to the time point before the current time point. Because each time a time point is reached, the final prediction result will be generated based on the basic data of the time point, so the data preparation module not only stores the current final prediction result but also stores the historical final prediction result. In step S212, the data decision module calculates the current final prediction result and the historical final prediction result. The data decision module also obtains the difference and deviation of different historical time points before the current time point from the data preparation module. Because the difference and deviation of the time point are calculated every time a time point is reached, the data preparation module also stores the difference and deviation of different historical time points before the current time point. In step S213, the data decision module determines whether the deviation of the current time point is greater than the deviation threshold, wherein the deviation threshold is the average of the deviations of different historical time points before the current time point. If so, it is believed that the prediction accuracy of not using the basic data of the current time point for prediction processing will be higher, so it is determined not to use the basic data of the current time point. If not, it is believed that the prediction accuracy of using the basic data of the current time point for prediction processing will be higher, so it is determined to use the basic data of the current time point.

[0050] Furthermore, the data decision module determines whether to use the basic data at the current time point for prediction processing. Step S213 also includes: the data decision module determines the formula δ t >|β t-1 +β t-2 +…+β t-n |Whether it is established, where δ t is the deviation value at the current time point t, β t-1 is the difference between the current time point t and the time point before it, β t-2 is the difference between the two time points before the current time point t, β t-n It is the difference between the n time points before the current time point t. If it is true, it is determined to use the basic data of the current time point. If it is not true, it is determined not to use the basic data of the current time point.

[0051] Specifically, the second execution situation of step S213 is introduced here, and the data decision module determines the formula δt >|β t-1 +β t-2 +…+β t-n |Is it true? If so, it is determined to use the basic data at the current time point. If not, it is determined not to use the basic data at the current time point. t is the deviation value at the current time point t, β t-1 is the difference between the current time point t and the time point before it, β t-2 is the difference between the two time points before the current time point t, β t-n is the difference between the current time point t and the n time points before it. The value of n is set in the actual application scenario. For example, if n is 3, the judgment formula is δ 11 >|β 10 +β9+β8|, and if δ 11 is 4, β 10 is -2, β9 is 3, and β8 is 1. Although the deviation value at 11 a.m. is greater than the deviation values ​​at 10 a.m., 9 a.m., and 8 a.m., the judgment formula is valid, and it is considered that the deviation value at 11 a.m. is still within an acceptable range. In this case, the basic data at the current time point is used.

[0052] Further, the data decision module determines whether to use the basic data at the current time point for prediction processing, S213 also includes: the data decision module uses the formula σ=(χ m-1 -x s ) 2 +(χ m-2 -x s-1 ) 2 +…+(χ m-k -x s-k+1 ) 2 Calculate the first eigenvalue, where σ is the first eigenvalue, χ m-1 ,x m-2 … m-k are the difference between the time point before the current time point m, the difference between the two time points before the current time point m, and the difference between the k time points before the current time point m, respectively. s ,x s-1 … s-k+1 They are respectively the difference between the time point before the historical time point s, the difference between the two time points before the historical time point s, and so on until the difference between the k time points before the historical time point s. When the first eigenvalue is less than or equal to the preset first eigenvalue threshold, it is determined to use the basic data of the current time point; when the first eigenvalue is greater than the preset first eigenvalue threshold, it is determined not to use the basic data of the current time point.

[0053] Specifically, the third execution situation of step S213 is introduced here. The data decision module uses the formula σ=(χ m-1 -x s ) 2 +(χ m-2 -x s-1 ) 2 +…+(χ m-k -x s-k+1 ) 2 Calculate the first eigenvalue. If the first eigenvalue is less than or equal to the preset first eigenvalue threshold, determine to use the basic data at the current time point. If the first eigenvalue is greater than the preset first eigenvalue threshold, determine not to use the basic data at the current time point. The first eigenvalue threshold is set according to the actual application situation. In the above formula, σ is the first eigenvalue, χ m-1 ,x m-2 … m-k are the difference between the previous time point of the current time point m, the difference between the two time points before the current time point m, and the difference between the k time points before the current time point m, respectively. s ,x s-1 … s-k+1 They are the difference of the previous time point of historical time point s, the difference of the previous two time points of historical time point s, and the difference of the previous k time points of historical time point s. The above method considers the change of the final prediction result in the time period from the current time point m to the previous k time points of the current time point m, which may be similar to the change of the final prediction result in the time period from the historical time point s to the previous k time points of the historical time point s. If the two are really similar, that is, the first eigenvalue is less than or equal to the first eigenvalue threshold, then it should be considered that the change of the final prediction result in the time period from the current time point m to the previous k time points of the current time point m is acceptable, and it should be determined to use the basic data of the current time point.

[0054] Furthermore, the data decision module determines whether to use the basic data at the current time point for prediction processing, and further includes the following steps:

[0055] S221, the data decision module uses the basic data at the current time point and the basic data at several historical time points before the current time point to form a basic data sequence μ h ,μ h-1 ,…,μ h-r , which are the basic data of the current time point h, the basic data of the historical time point h-1, and the basic data of the historical time point hr;

[0056] S222, the data decision module obtains the basic data of several time points of other enterprises from the data preparation module to form a reference basic data sequence ve ,ν e+1 ,…,ν e+r , which are the basic data of other enterprises at time point e, the basic data of other enterprises at time point e+1, and so on until the basic data of other enterprises at time point e+r;

[0057] S223, data decision module through the formula Calculate the second eigenvalue, where ρ is the second eigenvalue, D is a function for calculating the distance between two basic data, and determine whether the second eigenvalue is less than a preset second eigenvalue threshold. If not, jump to S222. If yes, determine whether to use the basic data ν e+r The final prediction result obtained by the prediction processing is closer to the real result. If yes, it is determined to use the basic data at the current time point. If no, it is determined not to use the basic data at the current time point.

[0058] Specifically, the second method for the data decision module to determine whether to use the basic data at the current time point is introduced here. In step S221, the data decision module uses the basic data at the current time point and the basic data at several historical time points before the current time point to form a basic data sequence μ h ,μ h-1 ,…,μ h-r , where μ h ,μ h-1 ,…,μ h-r The basic data of the current time point h, the basic data of the historical time point h-1, and the basic data of the historical time point hr are respectively, and the historical time point h-1 is the time point before the current time point h, and other historical time points are deduced in this way. In step S222, the data decision module obtains the basic data of several time points of other enterprises from the data preparation module to form a reference basic data sequence v e ,ν e+1 ,…,ν e+r , where ν e ,ν e+1 ,…,ν e+r The basic data of other companies at time point e, the basic data of other companies at time point e+1, and the basic data of other companies at time point e+r. Time point e+1 is the next time point of time e, and other time points are introduced in this way. It should be noted that the production scale, production mode, etc. of other companies are the same as those of the company whose emission data is to be predicted. In step S223, the data decision module uses the formula Calculate the second eigenvalue, where ρ is the second eigenvalue, D is a function for calculating the distance between two basic data, specifically, a function for calculating the distance between two vectors. If the second eigenvalue is less than a preset second eigenvalue threshold, the second eigenvalue threshold is set in an actual application scenario, indicating that the basic data sequence is similar to the reference basic data sequence. Then, it is determined whether the basic data ν is used in other enterprises. e+r The final prediction result obtained by the prediction processing is closer to the actual result, that is, other companies have already generated the actual result after the time point e+r. If so, it is determined to use the basic data at the current time point. If not, it is determined not to use the basic data at the current time point. If the second eigenvalue is greater than or equal to the preset second eigenvalue threshold, jump to step S222 and continue to search for a reference basic data sequence similar to the basic data sequence.

[0059] Furthermore, the prediction model includes a first-layer prediction model and a second-layer prediction model, the first-layer prediction model includes parallel sub-model one and sub-model two, and the second-layer prediction model is composed of sub-model three, wherein sub-model one, sub-model two, and sub-model three are all models composed of an input layer, an intermediate layer, and an output layer.

[0060] Specifically, the structure of the prediction model is introduced here. The prediction model is generally a two-layer structure, consisting of a first-layer prediction model and a second-layer prediction model. The first-layer prediction model includes parallel sub-model one and sub-model two, and the second-layer prediction model consists of sub-model three. In particular, sub-model one, sub-model two, and sub-model three are all neural network models, including an input layer, an intermediate layer, and an output layer.

[0061] Furthermore, the prediction module trains the prediction model, including the following steps:

[0062] S231, the prediction module obtains a number of training data from the data preparation module, the training data consists of basic data and real emission data, and the prediction module initializes and sets model parameters in the prediction model;

[0063] S232, based on the plurality of training data, the prediction module selects a piece of training data, inputs a first basic data portion of the basic data in the selected training data into sub-model one, inputs a second basic data portion of the basic data in the selected training data into sub-model two, then performs a preset operation on an output result of sub-model one and an output result of sub-model two, and then inputs the operation result into sub-model three, and sub-model three outputs predicted emission data;

[0064] S233, the prediction module calculates the deviation between the predicted emission data and the actual emission data in the selected training data, and based on the deviation first adjusts the model parameters in sub-model three, and then adjusts the model parameters in sub-model one and sub-model two at the same time, and the prediction module determines whether to end the training. If yes, end all steps, if not, jump to S232.

[0065] Specifically, the training process of the prediction model is introduced here. In step S231, the prediction module obtains several training data from the data preparation module. The training data consists of basic data and real emission data. For example, if the basic data is the basic data at 7 am, then the real emission data is the real emission data at 8 am. Then the prediction module initializes and sets the model parameters in the prediction model. Specifically, the model parameters can be set randomly. In step S232, the prediction module selects one training data from several training data, inputs the first basic data part of the basic data in the selected training data into sub-model one, and inputs the second basic data part of the basic data in the selected training data into sub-model two, and executes the output result of sub-model one and the output result of sub-model two. The preset operation, which can be specifically an inner product operation, is then input into sub-model three, and sub-model three outputs predicted emission data. In step S233, the prediction module calculates the deviation between the predicted emission data and the actual emission data in the selected training data, and first adjusts the model parameters in sub-model three according to the back propagation of the deviation in sub-model three, and then adjusts the model parameters in sub-model one and sub-model two at the same time according to the reverse deviation propagation from sub-model three. After that, the prediction module determines whether to end the training, specifically whether the prediction accuracy meets the requirements. If so, end all steps. If not, jump to S232, continue to select a training data that has not been used and perform subsequent processing. The above method can quickly train the prediction model and shorten the training time.

[0066] References Figure 2 As shown, the present invention also provides a carbon emission prediction system for production enterprises, including a data preparation module, a data decision module, and a prediction module, which is used to implement a carbon emission prediction method for production enterprises as described above. The functions of each module are as follows:

[0067] The data preparation module is used to store the basic data at the current time point and the basic data at different historical time points before the current time point;

[0068] A data decision module is used to determine whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point;

[0069] The prediction module is used to perform prediction processing based on the basic data at the current time point when the judgment result of the data decision module is to use the basic data at the current time point, and to perform prediction processing based on the basic data at the historical time point before the current time point when the judgment result of the data decision module is not to use the basic data at the current time point.

[0070] In summary, in the present invention, first, the data preparation module obtains the basic data at the current time point, and sends the basic data at the current time point to the data decision module, and the data decision module continues to obtain the basic data of different historical time points before the current time point from the data preparation module. Secondly, the data decision module determines whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data of different historical time points before the current time point. Finally, when the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data at the current time point. When the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs prediction processing according to the basic data of the historical time points before the current time point. The present invention determines whether to use the basic data at the current time point for prediction processing based on the impact that the basic data at the current time point will have on the final prediction result, which can significantly improve the prediction accuracy of emission data.

[0071] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0072] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0073] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0074] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0075] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A carbon emission prediction method for a production enterprise, applied to a carbon emission prediction system for a production enterprise, including a data preparation module, a data decision module, and a prediction module, characterized in that: The method comprises the following steps: S1, the data preparation module obtains basic data at the current time point, sends the basic data at the current time point to the data decision module, and the data decision module continues to obtain basic data at different historical time points before the current time point from the data preparation module; S2, the data decision module determines whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point; S3. When the judgment result of the data decision module is to use the basic data at the current time point, the prediction module performs the prediction processing according to the basic data at the current time point; when the judgment result of the data decision module is not to use the basic data at the current time point, the prediction module performs the prediction processing according to the basic data of the historical time point before the current time point.

2. The method according to claim 1, characterized in that In S1, the basic data includes a first basic data part and a second basic data part, wherein the first basic data part includes emission data between one time point, emission data before two time points, and emission data before M time points, and the second basic data part includes emission data at the current time point and identification data of the next time point after the current time point.

3. The method according to claim 2, characterized in that In S2, the prediction processing includes: the prediction module inputs the basic data into the trained prediction model, the prediction model outputs the prediction result of the emission data at the next time point of the current time point, the prediction module determines whether the prediction result is greater than or equal to a preset threshold, and if so, takes the next time point of the current time point as the final prediction result, and if not, the prediction module generates new basic data based on the prediction result, assuming that the next time point of the current time point has been reached, and repeats this step.

4. The method according to claim 3, characterized in that In S2, the data decision module determines whether to use the basic data at the current time point for prediction processing, including the following steps: S211, the data decision module obtains the current final prediction result, which is the final prediction result obtained by the prediction module using the basic data at the current time point to perform the prediction processing, and the data decision module continues to obtain the historical final prediction result, which is the final prediction result of the historical time point before the current time point stored by the data preparation module; S212, the data decision module calculates the difference between the current final prediction result and the historical final prediction result, and takes the absolute value of the difference as the deviation value of the current time point, and the data decision module continues to obtain the difference and deviation values ​​of different historical time points before the current time point from the data preparation module; S213, the data decision module determines whether the deviation value at the current time point is greater than the deviation value threshold, the deviation value threshold is the average value of the deviation values ​​at different historical time points before the current time point, if yes, it is determined not to use the basic data at the current time point, if not, it is determined to use the basic data at the current time point.

5. The method according to claim 4, characterized in that The data decision module determines whether to use the basic data at the current time point for prediction processing, and the step S213 further includes: the data decision module determines the formula δ t >|β t-1 +β t-2 +…+β t-n |Whether it is established, where δ t is the deviation value at the current time point t, β t-1 is the difference between the current time point t and the time point before it, β t-2 is the difference between the two time points before the current time point t, β t-n It is the difference between the n time points before the current time point t. If it is true, it is determined to use the basic data of the current time point. If it is not true, it is determined not to use the basic data of the current time point.

6. The method according to claim 4, characterized in that The data decision module determines whether to use the basic data at the current time point for prediction processing, and the S213 further includes: the data decision module uses the formula σ=(χ m-1 -x s ) 2 +(χ m-2 -x s-1 ) 2 +…+(χ m-k -x s-k+1 ) 2 Calculate the first eigenvalue, where σ is the first eigenvalue, χ m-1 ,x m-2 … m-k are the difference between the time point before the current time point m, the difference between the two time points before the current time point m, and the difference between the k time points before the current time point m, respectively. s ,x s-1 … s-k+1 They are respectively the difference between the time point before the historical time point s, the difference between the two time points before the historical time point s, and so on until the difference between the k time points before the historical time point s. When the first eigenvalue is less than or equal to the preset first eigenvalue threshold, it is determined to use the basic data of the current time point; when the first eigenvalue is greater than the preset first eigenvalue threshold, it is determined not to use the basic data of the current time point.

7. The method according to claim 3, characterized in that In S2, the data decision module determines whether to use the basic data at the current time point for prediction processing, and further includes the following steps: S221, the data decision module uses the basic data at the current time point and the basic data at several historical time points before the current time point to form a basic data sequence μ h ,μ h-1 ,…,μ h-r , which are the basic data of the current time point h, the basic data of the historical time point h-1, and the basic data of the historical time point hr; S222, the data decision module obtains basic data of other enterprises at several time points from the data preparation module to form a reference basic data sequence v e ,ν e+1 ,…,ν e+r , which are the basic data of other enterprises at time point e, the basic data of other enterprises at time point e+1, and so on until the basic data of other enterprises at time point e+r; S223, the data decision module uses the formula Calculate the second eigenvalue, where ρ is the second eigenvalue, D is a function for calculating the distance between two basic data, and determine whether the second eigenvalue is less than a preset second eigenvalue threshold. If not, jump to S222. If yes, determine whether to use the basic data ν e+r The final prediction result obtained by performing the prediction processing is closer to the real result. If yes, it is determined to use the basic data at the current time point. If no, it is determined not to use the basic data at the current time point.

8. The method according to claim 3, characterized in that The prediction model includes a first-layer prediction model and a second-layer prediction model, wherein the first-layer prediction model further includes parallel sub-model one and sub-model two, and the second-layer prediction model is composed of sub-model three, wherein sub-model one, sub-model two, and sub-model three are all models composed of an input layer, an intermediate layer, and an output layer.

9. The method according to claim 8, characterized in that The prediction module is trained for the prediction model, comprising the following steps: S231, the prediction module obtains a number of training data from the data preparation module, the training data consists of basic data and real emission data, and the prediction module initializes and sets model parameters in the prediction model; S232, based on a plurality of training data, the prediction module selects a piece of training data, inputs the first basic data part of the basic data in the selected training data into the sub-model 1, inputs the second basic data part of the basic data in the selected training data into the sub-model 2, then performs a preset operation on the output result of the sub-model 1 and the output result of the sub-model 2, and then inputs the operation result into the sub-model 3, and the sub-model 3 outputs the predicted emission data; S233, the prediction module calculates the deviation between the predicted emission data and the actual emission data in the selected training data, and based on the deviation first adjusts the model parameters in the sub-model three, and then adjusts the model parameters in the sub-model one and the sub-model two at the same time, and the prediction module determines whether to end the training. If yes, end all steps, if not, jump to S232.

10. A carbon emission prediction system for manufacturing enterprises, used to implement the method according to any one of claims 1 to 9, characterized in that: Includes the following modules: The data preparation module is used to store the basic data at the current time point and the basic data at different historical time points before the current time point; A data decision module is used to determine whether to use the basic data at the current time point for prediction processing based on the basic data at the current time point and the basic data at different historical time points before the current time point; The prediction module is used to perform prediction processing based on the basic data at the current time point when the judgment result of the data decision module is to use the basic data at the current time point, and to perform prediction processing based on the basic data at the historical time point before the current time point when the judgment result of the data decision module is not to use the basic data at the current time point.

Images