Mine nozzle water distribution liquid level measurement correction method based on piecewise linear regression and local weighted regression

Abstract

The present application relates to a kind of mine nozzle water distribution liquid level measurement correction method based on piecewise linear regression and local weighted regression, belong to coal mine dust reduction technical field.The method includes: obtaining basic data at the mine nozzle to be calibrated, wherein the basic data obtained includes the measured value measured by mine nozzle water distribution liquid level measurement sensor and the standard value obtained by standard equipment;According to the measured value and standard value obtained, establish piecewise linear regression model;Local weighted regression is introduced to dynamically adjust the weight of each section interval;Combining weight function and piecewise linear regression, establish the final error correction model.Real-time acquisition measured value data, using the final error correction model completes measurement correction.The present application can reduce the measured value deviation caused by equipment deviation and environmental factors, can real-time correction new measurement data, sufficiently improve the universality, accuracy of liquid level measurement equipment.

Description

Technical Field

[0001] This invention belongs to the field of coal mine dust suppression technology, and relates to a method for measuring and correcting the water distribution level of mine nozzles based on piecewise linear regression and local weighted regression. Background Technology

[0002] Coal mining generates a large amount of dust, which not only severely pollutes the working environment and affects the health of workers, but also poses safety hazards such as explosions. Therefore, dust control in coal mines is a crucial aspect of ensuring safe production and protecting the health of workers. Various dust control solutions exist. During drilling, methods such as wet rock drilling with pneumatic drills, dry rock drilling with dust collection, and wet drilling with pneumatic drills can be used. During blasting, water-based mud and blasting sprays can significantly reduce dust. Mechanical tunneling faces rely on a well-designed ventilation and dust removal system to control dust. Among spray dust control technologies, high-pressure spraying, dry fog dust suppression, and new wind-water integrated spray devices each have their advantages. In physical covering methods, dust nets, enclosed coal sheds, and surface solidifiers for coal piles can reduce dust. Mechanical ventilation and dust removal equipment, such as negative pressure dust removal systems and local dust removal devices, can effectively handle dust-laden air. Management and operational optimization are also indispensable; controlling stacking and loading / unloading, managing transportation channels, and intelligent monitoring and early warning systems can all contribute to dust control. Among source dust reduction measures, coal seam water injection and coal humidification can reduce dust generation. In addition, there are technologies such as foam dust removal, magnetized water dust suppression, and acoustic atomization dust suppression.

[0003] Currently, a common approach is to use spray systems with nozzles as the primary dust suppression system in coal mines. The nozzles, as a key component of this system, directly affect the effectiveness of dust suppression due to the accuracy and stability of their water distribution. Precise measurement and calibration of the nozzle water distribution level ensures that the dust suppression system sprays only as needed, avoiding both over- and under-spraying, thereby improving dust suppression efficiency and effectively reducing dust concentration in the coal mine working environment. Accurate level measurement and calibration help ensure the stable operation of the dust suppression system, avoiding safety hazards such as explosions caused by dust accumulation, thus guaranteeing safe production in coal mines. Reducing dust pollution improves the working environment, lowers the risk of occupational diseases caused by long-term dust inhalation, and protects the health of workers. Precise level control ensures the rational use of water resources, avoids waste, and meets the requirements of sustainable development.

[0004] In conclusion, accurate measurement and calibration of water distribution and liquid level in dust suppression nozzles used in coal mines are of great significance for improving dust suppression effects, ensuring safe production, protecting the health of workers, and improving resource utilization efficiency.

[0005] Currently, when measuring the water distribution level of nozzles used for dust suppression in coal mines, laser sensors are being used to measure the liquid level height. However, in the liquid level measurement, the measurement data is affected by both systematic errors (such as equipment deviation) and random errors (such as environmental noise, liquid surface fluctuations, etc.), causing the measured value to deviate from the true liquid level. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to provide a method for correcting the liquid level measurement of water distribution in mine nozzles based on piecewise linear regression and local weighted regression, which is used to correct systematic and random errors in liquid level measurement and can correct new measurement data in real time to improve the accuracy of liquid level measurement.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for correcting the water distribution level measurement of mine nozzles based on piecewise linear regression and local weighted regression, the method comprising the following steps:

[0009] S1. Obtain basic data at the mine nozzle to be calibrated. The basic data includes the measured value measured by the mine nozzle water distribution level measurement sensor and the standard value obtained by the standard equipment.

[0010] S2. Based on the obtained measurement values ​​L m and standard value L t Establish a piecewise linear regression model;

[0011] S3. Introduce local weighted regression to dynamically adjust the weights of each segment interval;

[0012] S4. Combining the weighting function and piecewise linear regression, establish the final error correction model.

[0013] Furthermore, in step S1, the multiple measurement data obtained by the liquid level sensor are recorded as the measured value L. m The standard liquid level value obtained through standard equipment is recorded as the standard value L. t During the data acquisition process, the measured value L m and standard value L t The actual liquid level is acquired synchronously and dynamically adjusted according to a preset step size, with the adjustment range covering the historical range of liquid level distribution in the mining nozzle. The data is preprocessed, and the sliding window mean method is used to smooth the noise of liquid level fluctuations while retaining effective trend information.

[0014] Furthermore, the water distribution level measurement sensor for mining nozzles refers to the level sensor actually installed in the dust suppression system for mining; standard equipment refers to equipment used separately for standard level measurement.

[0015] Furthermore, in step S2, the range of measured values ​​is first determined based on the maximum and minimum values ​​of the measured standard values ​​[[L]]. m_min ,L m_max The measured value range is divided into N intervals using a dynamic binning strategy;

[0016] Then, for each of the divided intervals i, i∈N, a piecewise linear regression model is constructed, where the slope α of the linear regression model within each interval is set. i and offset b i Then the piecewise linear regression model for each interval is expressed as: Among them, L m These are real-time measurements, and they are the independent variables of the model; This is the preliminary revised predicted value, which is the dependent variable of the model.

[0017] Furthermore, the slope α of the linear regression model within each interval is determined. i and offset b i At this time, Mahalanobis distance weights are introduced to reduce the impact of outlier measurement points on the fitting. The weight calculation formula is as follows:

[0018]

[0019] Where μ i The average liquid level of interval i is the standard value. This indicates the solution to the covariance matrix. This represents the j-th measurement value within interval i;

[0020] Then, the parameters are solved using matrix operations:

[0021]

[0022] Where X is the measurement matrix, W is the weight diagonal matrix, and Y is the true error vector.

[0023] Furthermore, in step S3, let each interval Ω i There is K i Historical average measurement data points First, extract K from this interval. i Historical average measurement data points feature:

[0024]

[0025] In the above process, Representing the interval Ω i K i Historical average measurement data points The mean, Representing the interval Ωi K i Historical average measurement data points The variance;

[0026] Therefore, based on the real-time measurement data values ​​within the current interval Determine the interval Ω i The real-time weight is expressed as:

[0027]

[0028] In the formula, ε is a positive constant that guarantees the denominator is not zero. That is, the interval Ω i Real-time local weights.

[0029] Furthermore, in step S4, the final error correction model is established as follows:

[0030]

[0031] Therefore, the corrected liquid level value is:

[0032] L corrected =L m +ΔL(L m )

[0033] Real-time acquisition of liquid level sensor data, followed by correction of the liquid level sensor data based on the final error correction model.

[0034] The beneficial effects of this invention are as follows:

[0035] This invention combines piecewise linear regression and locally weighted regression to measure and correct the liquid level distribution in mine nozzles, significantly improving measurement accuracy. The piecewise linear regression model constructs a linear relationship based on different liquid level intervals, effectively fitting the liquid level change trends within those intervals. The introduction of locally weighted regression further enhances the model's adaptability to local data variations by dynamically adjusting the weights of data points within each interval, reducing the impact of outlier measurement points on the overall fitting results. This dual regression mechanism works together to make the corrected liquid level value closer to the true value, improving measurement accuracy.

[0036] The piecewise linear regression model of this invention employs a dynamic bin-division strategy to divide liquid level intervals, automatically determining the optimal number of intervals based on actual conditions, thus avoiding the subjectivity and uncertainty caused by manual division. Simultaneously, local weighted regression, through feature extraction and real-time weight calculation of historical measurement data, dynamically adjusts the data weights within each interval, enhancing the model's local adaptability. This design enables the model to maintain high stability and prediction accuracy even under complex conditions such as varying operating conditions and environmental noise levels, improving the model's robustness and adaptability.

[0037] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0038] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0039] Figure 1 This is a flowchart illustrating the method for measuring and correcting the water distribution level of a mine nozzle based on piecewise linear regression and local weighted regression, as described in an embodiment of the present invention. Detailed Implementation

[0040] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0041] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0042] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0043] Please see Figure 1 This is a method for correcting the water distribution level measurement of mine nozzles based on piecewise linear regression and local weighted regression.

[0044] This embodiment first provides a specific implementation process for a method for correcting the water distribution and liquid level measurement of mine nozzles based on piecewise linear regression and local weighted regression, such as... Figure 1 As shown, it mainly includes the following steps:

[0045] S1. Obtain basic data at the mine nozzle to be calibrated. The basic data includes the measured value measured by the mine nozzle water distribution level measurement sensor and the standard value obtained by the standard equipment.

[0046] Specifically, a mine nozzle water distribution level measurement sensor refers to a level sensor actually installed in a mine dust suppression system, such as a reflective level gauge or a laser sensor. Standard equipment refers to standalone devices used for measuring standard levels, such as pressure-type level standard devices.

[0047] Multiple measurement data obtained through the liquid level sensor are recorded as the measured value L. m The standard liquid level value obtained through standard equipment is recorded as the standard value L. t During the data acquisition process, the measured value L m and standard value L t The actual liquid level is acquired synchronously and dynamically adjusted according to a preset step size, covering the historical range of the water distribution level in the mining nozzle. For example, the actual liquid level value increases in 0.5mm increments, covering the entire range and ensuring segmented calibration accuracy.

[0048] The data is preprocessed by using the sliding window mean method to smooth out the noise of liquid level fluctuations while retaining effective trend information.

[0049] S2. Based on the obtained measurement values ​​L m and standard value L tEstablish a piecewise linear regression model. Specifically, first determine the range of measured values ​​based on the maximum and minimum values ​​of the measured standard values ​​[[L]]. m_min ,L m_max The measured value range is divided into N intervals using a dynamic binning strategy. In this embodiment, the optimal number of intervals N can be automatically determined by combining the K-means clustering algorithm with the liquid level sensor range. The clustering quality is evaluated using the elbow rule, and splitting is stopped when the SSE decrease rate is <5% to avoid overfitting of the model due to excessive segmentation.

[0050] Then, for each of the partitioned intervals i, i∈N, a piecewise linear regression model is constructed, where the slope a of the linear regression model within each interval is... i and offset b i This can be determined using weighted least squares. Specifically, within interval i, an error model is constructed. At this time, Mahalanobis distance weights are introduced to reduce the impact of outlier measurement points on the fitting. The weight calculation formula is as follows:

[0051]

[0052] Where μ i The average liquid level of interval i is the standard value. This indicates the solution to the covariance matrix. This represents the j-th measurement value within interval i.

[0053] Solving for parameters using matrix operations:

[0054]

[0055] Where X is the measurement matrix, W is the weight diagonal matrix, and Y is the true error vector.

[0056] S3. To enhance the local adaptability of the segmented model, local weighted regression is introduced to dynamically adjust the weights of each segment interval. The process is as follows:

[0057] Let Ω be the interval. i There is K i Historical average measurement data points First, extract K from this interval. i Historical average measurement data points feature:

[0058]

[0059] In the above process, Representing the interval Ω i K i Historical average measurement data points The mean, Representing the interval Ω i K i Historical average measurement data points The variance;

[0060] Therefore, based on the real-time measurement data values ​​within the current interval Determine the interval Ω i The real-time weight is expressed as:

[0061]

[0062] In the formula, ε is a positive constant that guarantees the denominator is not zero. That is, the interval Ω i Real-time local weights.

[0063] S4. Combining the weighting function and piecewise linear regression, the final error correction model is established as follows:

[0064]

[0065] Therefore, the corrected liquid level value is:

[0066] L corrected =L m +ΔL(L m )

[0067] Real-time acquisition of liquid level sensor data, followed by correction of the liquid level sensor data based on the final error correction model.

[0068] In this embodiment, the measurement intervals are divided into [100, 120], [120, 150], [150, 180], and [180, 200], and the following measurement data are obtained within each interval:

[0069] Interval [100, 120]: Historical measurement data (measured values): {105, 110, 115}; Corresponding standard data (standard values): {104.5, 109.8, 114.9}.

[0070] Interval [120, 150]: Historical measurement data (measured values): {125, 130, 135, 140, 145}; Corresponding standard data (standard values): {124.6, 129.7, 134.8, 139.9, 145.1}.

[0071] Interval [150, 180]: Historical measurement data (measured values): {155, 160, 165, 170, 175}; Corresponding standard data (standard values): {154.8, 159.6, 164.5, 169.4, 174.3}.

[0072] Interval [180, 200]: Historical measurement data (measured values): {185, 190, 195}; Corresponding standard data (standard values): {184.6, 189.8, 194.9}.

[0073] Suppose a real-time measurement of 155 mm falls within the interval [150, 180]. First, calculate the mean measurement within this interval: (155 + 160 + 165 + 170 + 175) / 5 = 165; then calculate the standard mean: (154.8 + 159.6 + 164.5 + 169.4 + 174.3) / 5 = 164.52.

[0074] Based on the above data, a covariance matrix is ​​constructed, and then the slope and intercept of the linear regression model within this interval are obtained by matrix operations or table lookup: a = 0.998, b = -0.01.

[0075] Then, the real-time weights for each historical average measurement data point {155,160,165,170,175} within this interval are:

[0076] w=[0.8534, 0.9901, 0.6997, 0.3012, 0.0792]

[0077] The correction value is calculated as follows: (0.8534×((0.998×159-0.01)-155)+0.9901×((0.998×159-0.01)-160)+0.6997×((0.998×159-0.01)-165)+0.3012×((0.998×159-0.01)-170)+0.0792×((0.998×159-0.01)-175)) / 5≈--1.46.

[0078] The final liquid level is 159 - 1.46 = 157.64.

[0079] This demonstrates that it possesses the following advantages:

[0080] (1) Improve the accuracy of liquid level measurement

[0081] In this embodiment, by combining piecewise linear regression and locally weighted regression, the measurement accuracy of the water distribution level in mine nozzles is significantly improved. The piecewise linear regression model can construct a linear relationship based on different level intervals, thereby effectively fitting the level change trend within different intervals. The introduction of locally weighted regression further enhances the model's adaptability to local data changes. By dynamically adjusting the weights of data points within each interval, the impact of outlier measurement points on the overall fitting result is reduced. This dual regression mechanism works together to make the corrected level value closer to the true value, improving the measurement accuracy.

[0082] (2) Enhance the robustness and adaptability of the model

[0083] The piecewise linear regression model employs a dynamic bin-division strategy to divide the liquid level intervals, automatically determining the optimal number of intervals based on actual conditions, thus avoiding the subjectivity and uncertainty caused by manual division. Simultaneously, locally weighted regression dynamically adjusts the data weights within each interval through feature extraction and real-time weight calculation of historical measurement data, enhancing the model's local adaptability. This design allows the model to maintain high stability and prediction accuracy even under complex conditions such as varying operating conditions and environmental noise levels, improving its robustness and adaptability.

[0084] (3) Achieve real-time correction and efficient application

[0085] This solution also features real-time correction capabilities. In practical applications, accurate liquid level values ​​can be quickly obtained simply by acquiring liquid level sensor data in real time and correcting the data according to the established error correction model. This real-time correction mechanism not only reflects liquid level changes promptly but also effectively reduces resource waste and safety hazards caused by measurement errors. Furthermore, the algorithm design in this solution is simple and clear, easy to implement and apply, lowering the technical threshold and cost, and facilitating its widespread promotion and application in the field of coal mine dust suppression.

[0086] (4) Promote the intelligent development of coal mine dust suppression technology

[0087] The implementation and application of this solution will strongly promote the intelligent development of coal mine dust suppression technology. By accurately measuring and calibrating the water distribution and liquid level of the nozzles, it can be ensured that the dust suppression system sprays only as needed, neither too much nor too little, thereby improving the dust suppression effect and effectively reducing the dust concentration in the coal mine working environment. This not only helps improve the working environment and protect the health of workers, but also improves resource utilization efficiency and reduces operating costs. At the same time, the intelligent features of the solution also provide strong support for the automation, remote monitoring and management of coal mine dust suppression systems, promoting the transformation, upgrading and sustainable development of the coal mining industry.

[0088] In summary, the liquid level measurement correction scheme for water distribution in mine nozzles based on piecewise linear regression and local weighted regression of the present invention has shown significant beneficial technical effects in improving the accuracy of liquid level measurement, enhancing the robustness and adaptability of the model, realizing real-time correction and efficient application, and promoting the intelligent development of coal mine dust suppression technology.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for measuring and correcting the water distribution level in mine nozzles based on piecewise linear regression and local weighted regression, characterized in that: The method includes the following steps: S1. Acquire basic data at the mine nozzle to be calibrated. The acquired basic data includes the measured values ​​obtained by the mine nozzle water distribution level sensor and the standard values ​​obtained by the standard equipment. In step S1, the multiple measurement data obtained by the level sensor are recorded as measured values. The standard liquid level value obtained through standard equipment is recorded as the standard value. During the data acquisition process, the measured values and standard value The actual liquid level is acquired synchronously and dynamically adjusted according to a preset step size, with the adjustment range covering the historical range of liquid level distribution in the mining nozzles. The data is preprocessed, and the sliding window mean method is used to smooth the liquid level fluctuation noise while retaining effective trend information. S2. Establish a piecewise linear regression model based on the obtained measured values ​​and standard values; in step S2, first determine the range of measured values ​​based on the maximum and minimum values ​​of the measured standard values. The measurement range is divided into sections using a dynamic binning strategy. A range; Then, for each of the divided intervals Construct a piecewise linear regression model, where the slope of the linear regression model is set for each interval. and offset Then the piecewise linear regression model for each interval is expressed as: ,in, These are real-time measurements, and they are the independent variables of the model; These are the preliminary revised predicted values, which are the dependent variables of the model; Determine the slope of the linear regression model within each interval. and offset At this time, Mahalanobis distance weights are introduced to reduce the impact of outlier measurement points on the fitting. The weight calculation formula is as follows: in For interval The average liquid level of the standard value, This indicates the solution to the covariance matrix. Indicates the interval The first One measurement value; Then, the parameters are solved using matrix operations: in, This is a matrix of measured values. This is a weighted diagonal matrix. This represents the true error vector. S3. Introduce local weighted regression to dynamically adjust the weights of each segment interval; S4. Combining the weighting function and piecewise linear regression, establish the final error correction model.

2. The method for measuring and correcting the water distribution level of mine nozzles based on piecewise linear regression and local weighted regression according to claim 1, characterized in that: The water distribution level sensor for mining nozzles refers to the level sensor actually installed in the dust suppression system of a mine; standard equipment refers to equipment used separately for standard level measurement.

3. The method for measuring and correcting the water distribution level of mine nozzles based on piecewise linear regression and local weighted regression according to claim 2, characterized in that: In step S3, let each interval have Historical average measurement data points First, extract the interval. Historical average measurement data points feature: In the above process, Representing an interval of Historical average measurement data points The mean, Representing an interval of Historical average measurement data points The variance; Therefore, based on the real-time measurement data values ​​within the current interval Determine the interval The real-time weight is expressed as: In the formula, It is a positive number that guarantees the denominator is not zero. That is, the interval Real-time local weights.

4. The method for measuring and correcting the water distribution level of a mine nozzle based on piecewise linear regression and local weighted regression according to claim 3, characterized in that: In step S4, the final error correction model is established as follows: Therefore, the corrected liquid level value is: Real-time acquisition of liquid level sensor data, followed by correction of the liquid level sensor data based on the final error correction model.

Images