Low-carbon construction evaluation method based on data processing

Abstract

The invention discloses a low-carbon construction evaluation method based on data processing, and belongs to the technical field of low-carbon evaluation, and the method comprises the following steps: S1, inputting a historical carbon emission data set of a construction site into an expected detection model, and obtaining an expected error coefficient at each moment; s2, generating an amplitude adjustment value of nonlinear mapping according to the expected error coefficient of each moment, and obtaining a carbon emission mapping value of each moment by using a plurality of power coefficients of nonlinear mapping; s3, constructing an adaptive penalty equation according to the carbon emission mapping values at all moments, and obtaining the optimal penalty weight of the historical carbon emission data set; and S4, determining the carbon emission evaluation of the construction site according to the optimal penalty weight of the historical carbon emission data set. According to the method, the operation result of the historical peak value and the optimal penalty weight is adopted as the evaluation threshold value, the double safety margins are constructed, whether the possibility of occurrence of the extreme emission scene can be mastered or not is ensured, and the evaluation result is accurate and comprehensive.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-carbon assessment, and particularly relates to a low-carbon construction evaluation method based on data processing. Background Art

[0002] In energy conservation and emission reduction, as the main components of greenhouse gases, CO2, CH4, and N2O are the key research objects of energy conservation and emission reduction. Since CO2 accounts for the largest proportion in greenhouse gases, about 60%, the term carbon emission is commonly used as the general term for greenhouse gases. When calculating carbon emissions, other greenhouse gases are converted into equivalent CO2 to calculate the total emissions of greenhouse gases, and then judge the intensity of the greenhouse effect.

[0003] The carbon emissions continuously generated during the construction process have become an important management content in urban development. However, the carbon emission data at the construction site often has characteristics such as time-variation, non-linearity, and uncertainty, which makes it difficult for traditional linear evaluation methods to accurately reflect the carbon emission status at the construction site. Summary of the Invention

[0004] In order to solve the above problems, the present invention proposes a low-carbon construction evaluation method based on data processing.

[0005] The technical solution of the present invention is: a low-carbon construction evaluation method based on data processing includes the following steps: S1. Input the historical carbon emission data set of the construction site into the expected detection model to obtain the expected error coefficient at each moment; S2. Generate the amplitude adjustment value of the non-linear mapping according to the expected error coefficient at each moment, and use several power coefficients of the non-linear mapping to obtain the carbon emission mapping value at each moment; S3. Construct an adaptive penalty equation according to the carbon emission mapping values at all moments to obtain the optimal penalty weight of the historical carbon emission data set; S4. Determine the carbon emission evaluation of the construction site according to the optimal penalty weight of the historical carbon emission data set.

[0006] Further, in S1, the maximum value and the minimum value between the carbon emission value and the expected carbon emission value at each moment are extracted by using the expected detection model, and 1 minus the ratio between the minimum value and the maximum value is used to generate the expected error coefficient at each moment.

[0007] The beneficial effect of the above further solution is: in the present invention, the adopted normalization calculation method maps the error coefficient to the [0,1] interval, eliminates the influence of dimension, and facilitates the direct comparison and comprehensive analysis of cross-moment errors. The selection mechanism of the maximum value and the minimum value naturally focuses on the extreme values of the data, and can effectively identify the sudden abnormal fluctuations of carbon emissions.

[0008] Further, S2 includes the following sub - steps: S21. Take the expected error coefficient at each moment as the amplitude adjustment value of the non - linear mapping; S22. Perform an exponential operation with the base of the natural constant e on the carbon emission value at each moment as the first power coefficient of the non - linear mapping; perform an exponential operation with the base of the natural constant e on the opposite number of the ratio between the carbon emission value at each moment and the average carbon emission value over the total moments as the second power coefficient of the non - linear mapping; S23. According to the amplitude adjustment value, the first power coefficient, and the second power coefficient of the non - linear mapping, perform a non - linear mapping on the carbon emission value at each moment to obtain the carbon emission mapping value at each moment.

[0009] The beneficial effect of the above - mentioned further solution is: In the present invention, the expected error coefficient generated by S1 is directly used as the amplitude adjustment value A to construct an error - driven mapping intensity adjustment mechanism. The carbon emission data at the moment with a larger error will obtain a more significant non - linear transformation amplitude, realizing the automatic focusing on the abnormal emission mode. Through the moment - specific amplitude adjustment value, an associated mapping between the error characteristics in the time dimension and the spatial emission intensity is established.

[0010] The first power function performs a natural exponential operation on the original carbon emission value to strengthen the discrimination of high - emission values. The second power function constructs a standardized deviation index through the ratio operation between the carbon emission value and the global mean, eliminates the influence of dimensions, suppresses the excessive interference of extreme outliers on the mapping process, and at the same time retains the key morphological characteristics.

[0011] Further, in S23, the expression for performing the non - linear mapping is: ; In the formula, represents the carbon emission mapping value, represents the amplitude adjustment value of the non - linear mapping, represents the first power coefficient of the non - linear mapping, represents the second power coefficient of the non - linear mapping, represents the carbon emission value, represents the activation function, represents the average carbon emission value over the total moments, represents the logarithmic function with base 2.

[0012] The beneficial effect of the above - mentioned further solution is: In the present invention, in the first part, the first power function is used to form a self - regulating gating unit to prevent the exponential channel from being oversaturated. In the second part, By introducing periodic modulation, the hyperbolic tangent constraint tanh function compresses the carbon emission mapping values into the interval (-1, 1), ensuring that the oscillation amplitude is controllable and avoiding the problem of gradient disappearance. In the third part, the linear component of the original emission value is retained as a supplement to the residual of the non-linear mapping to prevent information loss.

[0013] Furthermore, S3 includes the following sub-steps: S31: Use the standard deviation of the carbon emission mapping values at all times as the initial weight of the equation; S32: Construct an adaptive penalty equation using the initial weight; S33: Solve the adaptive penalty equation to obtain the optimal penalty weight for the historical carbon emission dataset.

[0014] The beneficial effect of the above further scheme is that in the present invention, using the standard deviation of the carbon emission mapping values as the initial weight integrates the inherent fluctuation characteristics of the data into the initialization process of the adaptive penalty equation. The solution of the adaptive penalty equation can be decomposed into the parallel optimization of the weights at each time, and the obtained optimal penalty weight comprehensively considers the carbon emission changes over the entire duration.

[0015] Furthermore, in S33, the adaptive penalty equation has the following expression: ; In the formula, represents the carbon emission value at the th moment, represents the weight at the th moment, represents the th expected carbon emission value, represents the initial weight of the adaptive penalty equation, represents the regularization coefficient, represents the final result that needs to be obtained by solving the adaptive penalty equation, represents the total number of moments, represents the number of expected carbon emission values, represents the function for finding the input value that minimizes the function.

[0016] The beneficial effect of the above further scheme is that in the present invention, a deviation measure between the predicted value and the expected value is constructed, and the model is ensured to accurately fit the historical emission pattern through the least squares form. Introducing regularization can prevent overfitting. And by adjusting the γ value, the prediction error and the model complexity can be flexibly balanced. The continuous optimization of the final weight captures the optimal penalty weight.

[0017] Furthermore, S4 includes the following sub-steps: S41. Obtain the expected carbon emission value of the construction site at a future moment; S42. Determine whether the product of the maximum carbon emission value in the historical carbon emission dataset and the optimal penalty weight is greater than the expected carbon emission value. If so, the carbon emission evaluation of the construction site fails; otherwise, it passes.

[0018] The beneficial effect of the above further solution is that in the present invention, the historical maximum carbon emission value is used as a benchmark instead of the mean or median to construct a "worst-case" evaluation concept, ensuring that the construction site can grasp the possibility of sudden emissions.

[0019] The beneficial effect of the present invention is that the present invention extracts the expected error coefficient at each moment through the expected detection model, reflecting the deviation analysis of carbon emissions. The expected error coefficient is dynamically generated based on historical data, avoiding the lag of fixed thresholds; the present invention uses the expected error coefficient to perform multi-channel processing on carbon emission values such as logarithmic compression, hyperbolic tangent oscillation, and adaptive linearity, identifying the optimal penalty weight that plays a decisive role in carbon emission prediction; using the operation result of the historical peak value and the optimal penalty weight as the evaluation threshold to construct a double safety margin to ensure whether the possibility of extreme emission scenarios can be grasped, making the evaluation result accurate and comprehensive. Description of the Drawings

[0020] Figure 1 It is a flowchart of a low-carbon construction evaluation method based on data processing. Detailed Embodiment

[0021] The following further describes the embodiments of the present invention with reference to the drawings.

[0022] As Figure 1 shown, the present invention provides a low-carbon construction evaluation method based on data processing, including the following steps: S1. Input the historical carbon emission dataset of the construction site into the expected detection model to obtain the expected error coefficient at each moment; S2. Generate the amplitude adjustment value of the non-linear mapping according to the expected error coefficient at each moment, and use several power coefficients of the non-linear mapping to obtain the carbon emission mapping value at each moment; S3. Construct an adaptive penalty equation according to the carbon emission mapping values at all moments to obtain the optimal penalty weight of the historical carbon emission dataset; S4. Determine the carbon emission evaluation of the construction site according to the optimal penalty weight of the historical carbon emission dataset.

[0023] In the embodiment of the present invention, in S1, the maximum and minimum values between the carbon emission value and the expected carbon emission value at each moment are extracted by using the expected detection model, and 1 minus the ratio between the minimum value and the maximum value is used to generate the expected error coefficient at each moment.

[0024] In the present invention, the adopted normalization calculation method maps the error coefficient to the interval [0, 1], eliminates the influence of dimension, and facilitates the direct comparison and comprehensive analysis of cross-time errors. The selection mechanism of the maximum value and the minimum value naturally focuses on the extreme values of the data, and can effectively identify the sudden abnormal fluctuations of carbon emissions.

[0025] In the embodiment of the present invention, S2 includes the following sub-steps: S21. Take the expected error coefficient at each moment as the amplitude adjustment value of the non-linear mapping; S22. Perform an exponential operation with the natural constant e as the base on the carbon emission value at each moment as the first power coefficient of the non-linear mapping; perform an exponential operation with the natural constant e as the base on the opposite number of the ratio between the carbon emission value at each moment and the average carbon emission value of the total moments as the second power coefficient of the non-linear mapping; S23. According to the amplitude adjustment value, the first power coefficient, and the second power coefficient of the non-linear mapping, perform a non-linear mapping on the carbon emission value at each moment to obtain the carbon emission mapping value at each moment.

[0026] In the present invention, the expected error coefficient generated by S1 is directly used as the amplitude adjustment value A to construct an error-driven mapping intensity adjustment mechanism. The carbon emission data at the moment with a larger error will obtain a more significant non-linear transformation amplitude, realizing the automatic focusing on the abnormal emission mode. Through the moment-specific amplitude adjustment value, an association mapping between the error characteristics in the time dimension and the spatial emission intensity is established.

[0027] The first power function performs a natural exponential operation on the original carbon emission value to strengthen the discrimination of high emission values. The second power function constructs a standardized deviation index through the ratio operation of the carbon emission value and the global mean value, eliminates the influence of dimension, suppresses the excessive interference of extreme outliers on the mapping process, and at the same time retains the key morphological features.

[0028] In the embodiment of the present invention, in S23, the expression for performing the non-linear mapping is: ; In the formula, represents the carbon emission mapping value, represents the amplitude adjustment value of the non-linear mapping, represents the first power coefficient of the non-linear mapping, represents the second power coefficient of the non-linear mapping, represents the carbon emission value, represents the activation function, represents the average carbon emission value of the total moments, represents the logarithmic function with 2 as the base.

[0029] In the present invention, in the first part, a first power function is used to form a self-regulating gating unit to prevent the exponential channel from becoming saturated. In the second part, Periodic modulation is introduced, and the hyperbolic tangent constraint tanh function compresses the carbon emission mapping value into the interval (-1, 1), ensuring that the oscillation amplitude is controllable and avoiding the problem of gradient disappearance. In the third part, the linear component of the original emission value is retained as a residual supplement for the non-linear mapping to prevent information loss.

[0030] In an embodiment of the present invention, S3 includes the following sub-steps: S31. Take the standard deviation of the carbon emission mapping values at all times as the initial weight of the equation; S32. Use the initial weight to construct an adaptive penalty equation; S33. Solve the adaptive penalty equation to obtain the optimal penalty weight of the historical carbon emission data set.

[0031] In the present invention, taking the standard deviation of the carbon emission mapping value as the initial weight integrates the inherent fluctuation characteristics of the data into the initialization process of the adaptive penalty equation. The solution of the adaptive penalty equation can be decomposed into the parallel optimization of the weights at each moment, and the obtained optimal penalty weight comprehensively considers the carbon emission changes over the entire duration.

[0032] In an embodiment of the present invention, in S33, the adaptive penalty equation has the following expression: ; In the formula, represents the carbon emission value at the th moment, represents the weight at the th moment, represents the th expected carbon emission value, represents the initial weight of the adaptive penalty equation, represents the regularization coefficient, represents the final result that needs to be obtained by solving the adaptive penalty equation, represents the total number of moments, represents the number of expected carbon emission values, represents the function for finding the input value that minimizes the function.

[0033] In the present invention, a deviation metric between the predicted value and the expected value is constructed, and the least squares form is used to ensure the accurate fitting of the model to the historical emission pattern. Regularization is introduced to prevent overfitting. And by adjusting the γ value, the prediction error and the model complexity can be flexibly balanced. The continuous optimization of the final weight captures the optimal penalty weight.

[0034] In an embodiment of the present invention, S4 includes the following sub-steps: S41. Obtain the expected carbon emission value of the construction site at a future moment; S42. Determine whether the product of the maximum carbon emission value in the historical carbon emission data set and the optimal penalty weight is greater than the expected carbon emission value. If so, the carbon emission evaluation of the construction site fails; otherwise, it passes.

[0035] In the present invention, the historical maximum carbon emission value is used as a benchmark instead of the mean or median to construct a "worst-case" evaluation concept, ensuring that the construction site can grasp the possibility of sudden emissions.

[0036] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the principles of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations without departing from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A low-carbon construction evaluation method based on data processing, characterized in that It includes the following steps: S1. Input the historical carbon emission dataset of the construction site into the expected detection model to obtain the expected error coefficients at each moment; S2. Generate the amplitude adjustment value of the non-linear mapping according to the expected error coefficients at each moment, and use several power coefficients of the non-linear mapping to obtain the carbon emission mapping values at each moment; S3. Construct an adaptive penalty equation according to the carbon emission mapping values at all moments to obtain the optimal penalty weight of the historical carbon emission dataset; S4. Determine the carbon emission evaluation of the construction site according to the optimal penalty weight of the historical carbon emission dataset.

2. The low-carbon construction evaluation method based on data processing according to claim 1, wherein In S1, the maximum and minimum values between the carbon emission values and the expected carbon emission values at each moment are extracted by using the expected detection model, and the ratio between 1 and the ratio of the minimum value to the maximum value is generated as the expected error coefficient at each moment.

3. The low-carbon construction evaluation method based on data processing according to claim 1, characterized in that S2 includes the following sub-steps: S21. Take the expected error coefficients at each moment as the amplitude adjustment value of the non-linear mapping; S22. Perform the exponential operation with the natural constant e as the base on the carbon emission values at each moment as the first power coefficient of the non-linear mapping; perform the exponential operation with the natural constant e as the base on the negative value of the ratio between the carbon emission values at each moment and the average carbon emission value of the total moments as the second power coefficient of the non-linear mapping; S23. Perform non-linear mapping on the carbon emission values at each moment according to the amplitude adjustment value, the first power coefficient and the second power coefficient of the non-linear mapping to obtain the carbon emission mapping values at each moment.

4. The low-carbon construction evaluation method based on data processing according to claim 3, wherein In S23, the expression for performing non-linear mapping is: ； In the formula, represents the carbon emission mapping value, represents the amplitude adjustment value of the non-linear mapping, represents the first power coefficient of the non-linear mapping, represents the second power coefficient of the non-linear mapping, represents the carbon emission value, represents the activation function, represents the average value of carbon emissions at the total time, represents the logarithmic function with base 2.

5. The low-carbon construction evaluation method based on data processing according to claim 1, characterized in that S3 includes the following sub-steps: S31. Take the standard deviation of the carbon emission mapping values at all moments as the initial weight of the equation; S32. Construct an adaptive penalty equation by using the initial weight; S33. Solve the adaptive penalty equation to obtain the optimal penalty weight of the historical carbon emission dataset.

6. The low-carbon construction evaluation method based on data processing according to claim 5, wherein In S33, the adaptive penalty equation has the following expression: ； In the formula, represents the carbon emission value at the th moment, represents the weight at the th moment, represents the th expected carbon emission value, represents the initial weight of the adaptive penalty equation, represents the regularization coefficient, represents the final result that needs to be obtained by solving the adaptive penalty equation, represents the total number of moments, represents the number of expected carbon emission values, represents the function for finding the input value that minimizes the function.

7. The low-carbon construction evaluation method based on data processing according to claim 1, characterized in that S4 includes the following sub-steps: S41. Obtain the expected carbon emission value at the future moment of the construction site; S42. Judge whether the product of the maximum carbon emission value of the historical carbon emission dataset and the optimal penalty weight is greater than the expected carbon emission value. If so, the carbon emission evaluation of the construction site fails; otherwise, it passes.

Images