CN-121994746-A - Method and device for calculating heat value of coal

CN121994746ACN 121994746 ACN121994746 ACN 121994746ACN-121994746-A

Abstract

The application relates to the technical field of coal detection, in particular to a method and a device for calculating a coal calorific value, wherein the method comprises the steps of obtaining a raw coal sample of target coal and collecting spectrum data of the raw coal sample; the method comprises the steps of preprocessing an original spectrum vector corresponding to spectrum data to obtain a standardized spectrum feature vector, generating a local nonlinear feature and a global mapping relation corresponding to the spectrum feature vector based on the spectrum feature vector and a pre-constructed prediction model, and generating a heat value of target coal by fusing the local nonlinear feature and the global mapping relation. Therefore, the problems that in the related technology, a traditional calorimeter measuring method is long in time consumption, strong in destructiveness and complex in operation, large-scale industrial production and real-time monitoring requirements are difficult to meet, prediction accuracy is unstable due to the characteristics of equipment performance limitation, complex data characteristics and the like of a modeling method based on near infrared spectrum, and a deep learning model for spectrum modeling lacks of interpretation of internal mechanisms are solved.

Inventors

DOU YOUQUAN
SHU XI
LEI MENG
Ren Kelong
ZHAO TIANJU
ZOU LIANG

Assignees

国能南京煤炭质量监督检验有限公司
国电环境保护研究院有限公司
中国矿业大学

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (10)

1. The calculation method of the heat value of the coal is characterized by comprising the following steps of: acquiring a raw coal sample of target coal, and collecting spectrum data of the raw coal sample; preprocessing an original spectrum vector corresponding to the spectrum data to obtain a standardized spectrum characteristic vector; based on the spectral feature vector and a pre-constructed prediction model, generating a local nonlinear feature and a global mapping relation corresponding to the spectral feature vector so as to fuse the local nonlinear feature and the global mapping relation and generate the heat value of the target coal.
2. The method of claim 1, wherein preprocessing the original spectral vector corresponding to the spectral data to obtain a normalized spectral feature vector comprises: Calculating the mean value and standard deviation of the spectrum data based on the original spectrum vector; And based on the mean value and the standard deviation, centering each reflection intensity value in the original spectrum vector to zero mean value and scaling to unit variance to obtain the spectrum characteristic vector.
3. The method of claim 1, further comprising, prior to generating the local nonlinear feature and global mapping corresponding to the spectral feature vector based on the spectral feature vector and the pre-constructed predictive model: constructing radial basis function branches for capturing local nonlinear characteristics of a target spectrum interval in the spectrum characteristic vector; Constructing a linear branch for establishing an overall mapping relation between input features and a predicted target; constructing a single-layer prediction module in the pre-constructed prediction model based on the radial basis function branches and the linear branches; and sequentially superposing a plurality of single-layer prediction modules to obtain the pre-constructed prediction model.
4. The method according to claim 3, wherein generating the local nonlinear feature and global mapping relationship corresponding to the spectral feature vector based on the spectral feature vector and a pre-constructed prediction model comprises: Inputting the spectral feature vector into the pre-constructed prediction model to locate a spectral interval in the spectral feature vector meeting a target condition based on the radial basis function branches in the pre-constructed prediction model; Collecting local responses of each spectrum channel in the spectrum characteristic vector in the spectrum interval, and local responses of each spectrum channel in other spectrum intervals except the spectrum interval; And combining the local response and response weight corresponding to the spectrum interval with the local response and response weight corresponding to the other spectrum interval to obtain the local nonlinear characteristic.
5. The method according to claim 3, wherein generating the local nonlinear feature and global mapping relationship corresponding to the spectral feature vector based on the spectral feature vector and a pre-constructed prediction model comprises: The spectral feature vector is input into the pre-constructed prediction model to establish a global mapping relationship between the spectral feature vector and the heating value based on the linear branches in the pre-constructed prediction model.
6. The method of claim 1, further comprising, after calculating the heating value of the target coal: Identifying a narrowband spectrum interval in the spectrum characteristic vector based on the pre-constructed prediction model; and establishing a mapping relation between the narrow-band spectrum interval and a physical and chemical absorption band in the spectrum data, so as to determine an interpretable result of the prediction model according to the mapping relation.
7. A computing device for the calorific value of coal, comprising: The acquisition module is used for acquiring a raw coal sample of target coal and acquiring spectrum data of the raw coal sample; The processing module is used for preprocessing the original spectrum vector corresponding to the spectrum data to obtain a standardized spectrum characteristic vector; the calculation module is used for generating local nonlinear characteristics and global mapping relations corresponding to the spectral characteristic vectors based on the spectral characteristic vectors and a pre-constructed prediction model so as to fuse the local nonlinear characteristics and the global mapping relations and generate the heat value of the target coal.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the method of calculating the calorific value of coal as claimed in any one of claims 1 to 6.
9. A computer-readable storage medium having stored thereon a computer program, characterized in that the program is executed by a processor for realizing the calculation method of the calorific value of coal as claimed in any one of claims 1 to 6.
10. A computer program product comprising a computer program which, when executed, is adapted to carry out the method of calculating the calorific value of coal as claimed in any one of claims 1 to 6.

Description

Method and device for calculating heat value of coal Technical Field The application relates to the technical field of coal detection, in particular to a method and a device for calculating a heat value of coal. Background In the related technology, the coal calorific value detection technology is mainly divided into two types, namely a traditional calorimeter measurement method, a standard coal calorific value measurement method, an accurate measurement value obtained by completely burning a coal sample, and high accuracy, and a modeling method based on near infrared spectrum, wherein the portable near infrared spectrometer is used for collecting data, and a linear modeling or shallow machine learning model is combined for predicting the coal calorific value, so that the method has the advantages of rapidness, no damage and portability. However, in the related art, the traditional calorimeter measurement method has the problems of long time consumption, strong destructiveness and complex operation, is difficult to meet the requirements of large-scale industrial production and real-time monitoring, the modeling method based on the near infrared spectrum is limited by the characteristics of low resolution of portable equipment, insufficient signal-to-noise ratio and strong environmental sensitivity, near infrared spectrum data has the characteristics of high dimensionality, strong collinearity and nonlinearity, so that the prediction precision is unstable, and a deep learning model for spectrum modeling mostly belongs to a black box model, lacks the interpretability of internal mechanisms, limits the popularization of the deep learning model in practical industrial application, and needs to be solved urgently. Disclosure of Invention The application provides a calculation method and a calculation device for a coal calorific value, which are used for solving the problems that in the related technology, the traditional calorimeter measurement method is long in time consumption, strong in destructiveness and complex in operation, and is difficult to meet the requirements of large-scale industrial production and real-time monitoring, the modeling method based on near infrared spectrum is easy to cause unstable prediction precision due to the characteristics of equipment performance limitation, complex data characteristics and the like, and a deep learning model for spectrum modeling lacks of the interpretation of an internal mechanism, so that the popularization of the deep learning model in practical industrial application is limited. An embodiment of the first aspect of the application provides a calculation method of a coal calorific value, which comprises the following steps of obtaining a raw coal sample of target coal, collecting spectrum data of the raw coal sample, preprocessing an original spectrum vector corresponding to the spectrum data to obtain a standardized spectrum feature vector, and generating local nonlinear features and a global mapping relation corresponding to the spectrum feature vector based on the spectrum feature vector and a pre-constructed prediction model to fuse the local nonlinear features and the global mapping relation to generate the calorific value of the target coal. Optionally, in one embodiment of the present application, the preprocessing the original spectrum vector corresponding to the spectrum data to obtain a normalized spectrum feature vector includes calculating a mean value and a standard deviation of the spectrum data based on the original spectrum vector, and centering each reflection intensity value in the original spectrum vector to zero mean value and scaling to unit variance based on the mean value and the standard deviation to obtain the spectrum feature vector. Optionally, in one embodiment of the present application, before generating the local nonlinear feature and the global mapping relation corresponding to the spectral feature vector based on the spectral feature vector and the pre-built prediction model, the method further includes constructing a radial basis function branch for capturing the local nonlinear feature of the target spectral interval in the spectral feature vector, constructing a linear branch for establishing an overall mapping relation between the input feature and the prediction target, constructing a single-layer prediction module in the pre-built prediction model based on the radial basis function branch and the linear branch, and sequentially stacking a plurality of single-layer prediction modules to obtain the pre-built prediction model. Optionally, in one embodiment of the present application, the generating the local nonlinear feature and the global mapping relationship corresponding to the spectral feature vector based on the spectral feature vector and a pre-constructed prediction model includes inputting the spectral feature vector into the pre-constructed prediction model to locate a spectral interval in the spectral feature