CN-122024896-A - Method and system for detecting concentration of harmful gas in blasting mining process of strip mine

CN122024896ACN 122024896 ACN122024896 ACN 122024896ACN-122024896-A

Abstract

The invention discloses a method and a system for detecting concentration of harmful gas in a strip mine blasting exploitation process, and belongs to the technical field of safety monitoring in the mine exploitation process. The method comprises the steps of deploying a multi-source heterogeneous data acquisition terminal in a strip mine blasting area to acquire key multi-dimensional data, introducing a gas fingerprint feature library, carrying out matching calculation on the key multi-dimensional data and the gas fingerprint feature library data by utilizing an identification algorithm to acquire gas components, introducing a cross interference correction model library to carry out inversion calculation on the components of harmful gas to acquire a specific gas concentration value and checking, acquiring all data, carrying out space-time alignment and fusion, constructing a concentration distribution diagram of the harmful gas at the current time, and outputting the concentration distribution diagram in a data form. According to the invention, through the multisource sensing group and the data preprocessing, pattern recognition and time sequence analysis algorithm, the accuracy of detecting the concentration of harmful gas in a complex mixed gas environment is improved.

Inventors

ZHANG QI
ZHU XIAOLONG
WANG HETANG
WANG SHUGUO
WANG YUNPENG
ZHANG YADONG
LI WENCHAO

Assignees

中国矿业大学
新疆鼎飞益智能科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260204

Claims (10)

1. The method for detecting the concentration of harmful gas in the blasting mining process of the strip mine is characterized by comprising the following steps of: In an open pit mine blasting mining area, deploying a multi-source heterogeneous data acquisition terminal, acquiring key multi-dimensional data, preprocessing the data, extracting data features from the key multi-dimensional data, and acquiring multi-dimensional feature vectors; Based on a gas fingerprint feature library, matching and calculating the multidimensional feature vector and gas fingerprint feature library data by using a weighted K nearest neighbor algorithm, searching K nearest neighbors and voting to obtain a gas component identification result; Introducing a cross interference correction model library, calculating a specific value of gas concentration by calling a calculation model and parameters of the gas of the components corresponding to the gas component identification result in the library, and checking; and acquiring space-time position information from the system, fusing and aligning all data information under a uniform time stamp, dividing a monitoring area network, constructing a concentration distribution map, and outputting the concentration distribution map in a data form.
2. The method for detecting the concentration of harmful gases in the process of blasting mining of strip mine according to claim 1, wherein the steps of collecting key multidimensional data and preprocessing the data comprise the following steps: S1, in a blasting operation surface area, arranging N fixed detection nodes in a fan shape along an expected diffusion path, wherein each detection node hardware integration comprises an electrochemical sensor array, a MOS sensor, a miniature weather station, an RTC and a positioning communication mechanism; S2, setting a synchronous sampling mechanism, sending out a time signal by a central processing unit according to the blasting time T, taking the time signal as a time zero point received by all detection nodes, starting a fixed sampling frequency, and collecting key multidimensional data, wherein each collected original data packet is marked as a relative sampling point sequence number k in the node; s3, smoothing the time sequence data flow of each sensor by independently utilizing a moving average filter MAF, wherein the calculation formula of the filtered signals is as follows: Wherein x Filtering [ k ] represents the filtered smoothed output signal value at time index k, 1/M represents the average factor, Σ represents the summation, M represents the sliding window length, k represents the relative sample point number, x Original, original [ k-j ] represents the historical raw data point, j represents the summation cycle variable; And S4, carrying out environment compensation correction on the readings of each sensor, wherein the environment compensation correction comprises electrochemical sensor temperature compensation, electrochemical sensor humidity cross interference compensation and MOS sensor temperature and humidity comprehensive compensation.
3. The method for detecting concentration of harmful gases in a strip mine explosion mining process according to claim 2, wherein extracting data features from key multidimensional data comprises the steps of: s1, packaging all the data subjected to the pretreatment step at the same detection node and the same sampling time k to form a standardized multidimensional feature vector V [ k ]; And S2, packaging the multidimensional feature vector V [ k ] and then carrying out parameter transfer.
4. A method for detecting the concentration of harmful gases in a strip mine explosion mining process according to claim 3, wherein the preparation before matching calculation of the multidimensional feature vector and the gas fingerprint feature library data comprises the following steps: s1, receiving parameters of a multidimensional feature vector V [ k ], extracting feature subsets which are sensitive to gas types and have discriminant at the same time, and forming a simplified feature vector V f for pattern matching; s2, introducing a gas fingerprint feature library F-DB through a data interface, wherein the gas fingerprint feature library F-DB comprises M class samples of known gases.
5. The method for detecting concentration of harmful gas in a strip mine explosion mining process according to claim 4, wherein the matching calculation of the multidimensional feature vector and the gas fingerprint feature library data comprises the following steps: S1, extracting a standard sample V std from a gas fingerprint feature library F-DB, and calculating a weighted Euclidean distance value D m between a simplified feature vector V f and a standard sample V std which are input in real time; S2, selecting K neighbor samples of a weighted Euclidean distance value D m which is close to a real-time input simplified feature vector V f , counting class labels Label of the K neighbor samples, and carrying out weighted voting on the class labels by using the inverse distance to obtain a gas component identification result; And S2, performing confidence calculation on main gas components in the gas component identification result, and finally packaging the gas component identification result and the confidence and outputting the data.
6. The method for detecting concentration of harmful gas in a strip mine explosion mining process according to claim 5, wherein selecting a calculation model and parameters of the gas of the component corresponding to the gas component identification result comprises the steps of: S1, extracting main gas component information and secondary gas component information from a gas component identification result, and simultaneously extracting a gas concentration value acquired by an electrochemical sensor subjected to environment compensation correction; S2, selecting a correction model function and a model parameter set of the corresponding gas component information in a cross interference correction model library according to the gas component information obtained in the step S1, and taking the correction model function and the model parameter set as a preparation model before calculating a concentration specific value; and S3, extracting the observed values of the primary gas component information and the secondary gas component information from the gas concentration values obtained from the electrochemical sensor subjected to the environmental compensation correction in the step S1.
7. The method for detecting the concentration of harmful gases in the process of blasting mining of strip mine according to claim 6, wherein the steps of calculating a specific value of the concentration of the gases and checking include the steps of: S1, calling a preparation model to obtain a correction model function and a model parameter set, substituting an observed value into the model to calculate a specific value of the gas concentration, wherein the situation A, the situation B and the situation C are included when secondary gas components are processed; S2, a strategy adopts a simplified baseline compensation model to be recalled from a cross interference correction model library to carry out baseline compensation only, a strategy adopts an empirical formula of MOS sensor signals and an electrochemical sensor total value to estimate total harmful gas equivalent concentration, wherein the empirical formula is C Total (S) =p*RS MOS +q which is TC+r, C Total (S) represents a total harmful gas equivalent concentration estimated value, RS MOS represents MOS sensor resistance ratio after environmental compensation, TC represents electrochemical sensor array reading sum, p, q and r respectively represent weight coefficients of MOS sensor signals, weight coefficients of electrochemical sensor sum and constant items, and the strategy is free of secondary gas components and only processes main gas components; and S3, carrying out rationality test on the gas concentration specific value C Gas and its preparation method obtained through calculation processing in the step S1-the step S2, wherein the rationality test comprises non-0 constraint and range limiting.
8. The method for detecting concentration of harmful gases in a strip mine explosion mining process according to claim 7, wherein the step of fusing all data information under a uniform time stamp comprises the steps of: The method comprises the following steps that S1, a central processing unit collects data information reported by all detection nodes, wherein the data information comprises a gas concentration specific value C Gas and its preparation method , node coordinates and a time stamp; And S2, carrying out space-time alignment on the data information, screening out effective data in the same time window, and converting the geographic coordinates into plane coordinates suitable for distance calculation to form a discrete observation data set at the current moment.
9. The method for detecting concentration of harmful gas in a strip mine explosion mining process according to claim 8, wherein the dividing the monitoring area network, constructing a concentration profile comprises the steps of: s1, estimating a concentration value for each unknown grid point on a regular grid of the whole monitoring area by using an inverse distance weighting IDW spatial interpolation algorithm based on discrete point data of a discrete observation data set, and generating a continuous two-dimensional concentration distribution matrix covering the whole area; S2, packaging the two-dimensional concentration distribution matrix obtained through calculation and the corresponding metadata, encapsulating the two-dimensional concentration distribution matrix into a standardized format, constructing a visual data packet and outputting the visual data packet; and S3, receiving the visual data packet, and rendering by using drawing software to generate an intuitive concentration space distribution map.
10. Harmful gas concentration detecting system of opencut blasting exploitation process, its characterized in that includes: the first module is used for collecting key multidimensional data, preprocessing the data and extracting data characteristics to construct multidimensional feature vectors; The second module is used for carrying out matching calculation on the multidimensional feature vector and the gas fingerprint feature library data by utilizing a weighted K nearest neighbor algorithm based on the gas fingerprint feature library, searching K nearest neighbors and voting, and obtaining and outputting a gas component identification result; the third module is used for introducing a cross interference correction model library, calling a calculation model and parameters of the gas of the components corresponding to the gas component identification result in the library, calculating a gas concentration specific value, checking, and finally outputting the gas concentration specific value; And the fourth module is used for acquiring space-time position information from the system, fusing and aligning all the information under a uniform time stamp, dividing a monitoring area network, constructing a concentration distribution map, outputting the concentration distribution map in a data form, and presenting the concentration distribution map through drawing software.

Description

Method and system for detecting concentration of harmful gas in blasting mining process of strip mine Technical Field The invention relates to the technical field of safety monitoring in an ore mining process, in particular to a method and a system for detecting concentration of harmful gas in an open-pit mine blasting mining process. Background Mineral resources are used as core basic raw materials for industrial production, development processes of mining technology are deeply bound with progress of human industrial civilization, strip mine mining becomes one of main modes of global mineral resource mining due to the characteristics of high mining efficiency, high resource recovery rate, relatively superior operation conditions and the like, the mining scale of the strip mine is continuously expanded along with rapid development of global industry, large-scale strip mines and ultra-large-scale strip mines are continuously emerging, the blasting technology is developed towards the directions of precision and large dosage, advanced blasting processes such as directional blasting, millisecond delay blasting and the like are gradually applied, the mining efficiency is improved in a crossing mode, meanwhile, the generation amount and the diffusion range of harmful gases after blasting are further enlarged, the gas composition is more complex, and higher requirements are provided for coverage range, detection precision and real-time performance of the detection technology. At present, after blasting operation, the condition that harmful gas and conventional gas are mixed exists in a strip mine monitoring area, the existing detection technology adopts single gas sensor fixed point monitoring, gas sample data of the existing detection technology depends on laboratory tests to obtain concentration values, the concentration values of the harmful gas in the strip mine area after blasting operation cannot be monitored in real time, and the problem of potential safety hazard coefficient rising occurs. Disclosure of Invention The invention aims to provide a method and a system for detecting the concentration of harmful gas in the blasting mining process of an open-pit mine, wherein the system is based on a multi-source heterogeneous data acquisition terminal, performs preprocessing on heterogeneous data by introducing a multi-dimensional data preprocessing technology, calculates and identifies the type of the gas by introducing a gas fingerprint feature library, confirms the concentration value of the harmful gas by introducing a cross interference correction model library to calculate the concentration value of the gas with different components, and finally constructs a distribution map to realize visual monitoring, so that the problems that in the prior art, single gas sensor is adopted for fixed-point monitoring, the gas sample data depends on laboratory test to obtain the concentration value, and the open-pit mine area after blasting operation cannot be monitored in real time, so that the potential safety hazard coefficient is increased are solved. In order to achieve the above purpose, the present invention provides the following technical solutions: The invention relates to a method for detecting the concentration of harmful gas in the blasting mining process of strip mine, which comprises the following steps: In an open pit mine blasting mining area, deploying a multi-source heterogeneous data acquisition terminal, acquiring key multi-dimensional data, preprocessing the data, extracting data features from the key multi-dimensional data, and acquiring multi-dimensional feature vectors; Based on a gas fingerprint feature library, matching and calculating the multidimensional feature vector and gas fingerprint feature library data by using a weighted K nearest neighbor algorithm, searching K nearest neighbors and voting to obtain a gas component identification result; Introducing a cross interference correction model library, calculating a specific value of gas concentration by calling a calculation model and parameters of the gas of the components corresponding to the gas component identification result in the library, and checking; and acquiring space-time position information from the system, fusing and aligning all data information under a uniform time stamp, dividing a monitoring area network, constructing a concentration distribution map, and outputting the concentration distribution map in a data form. Further, the steps of collecting key multidimensional data and preprocessing the data comprise the following steps: S1, in a blasting operation surface area, arranging N fixed detection nodes in a fan shape along an expected diffusion path, wherein each detection node hardware integration comprises an electrochemical sensor array, a MOS sensor, a miniature weather station, an RTC and a positioning communication mechanism; S2, setting a synchronous sampling mechanism, sending out a time signal by a central proc