A coal quality detection method and system based on multi-sensor fusion

Abstract

The present application proposes a multi-sensor fusion coal quality detection method and system, which belongs to the field of coal quality detection technology. The target coal is detected by a gap sensor to obtain gap sensing parameters and coal characteristic parameters; the coal gap superposition analysis is performed based on the gap sensing parameters, and the gap error parameters are calculated in combination with the coal characteristic parameters; the density parameters of the target coal are collected by a weight sensor, and the density parameters are corrected according to the coal gap parameters and gap error parameters to obtain the corrected density parameters; the ash content prediction accuracy is configured according to the gap parameters and error parameters, and the ash content is predicted based on the corrected density parameters to obtain the final coal quality detection results. The present application solves the problem in the prior art that the coal gap leads to inaccurate density measurement, which in turn leads to inaccurate ash content detection, and achieves the effect of improving the accuracy of coal quality ash content detection through multi-sensor fusion and gap correction.

Description

Technical Field

[0001] The present invention relates to the technical field of coal quality detection, and in particular to a multi-sensor fusion coal quality detection method and system. Background Art

[0002] Currently, the commonly used methods for coal ash content testing in industry primarily include laboratory chemical analysis and density-based online testing. While laboratory chemical analysis offers high accuracy, it is cumbersome and time-consuming, making it difficult to meet the demand for real-time testing during coal production, processing, and use. Density-based online testing, on the other hand, measures coal density and establishes a correlation model between density and ash content for prediction. It offers advantages such as rapid testing and real-time monitoring, leading to its widespread use in industrial settings.

[0003] However, existing density-based online ash content measurement methods have significant shortcomings. Because coal inevitably forms large gaps during transportation, stacking, and testing, these gaps can cause the measured coal volume to be greater than the actual volume, resulting in an underestimation of the calculated density. This "false decrease" in coal density directly affects the accuracy of ash content prediction, especially when large variations in coal gaps lead to significant prediction errors. Summary of the Invention

[0004] The present invention aims to solve the technical problem in the prior art that coal gaps lead to inaccurate density measurement, which in turn causes inaccurate ash detection, and provides a multi-sensor fusion coal quality detection method and system to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a multi-sensor fusion coal quality detection method, comprising: detecting target coal through a gap sensor, obtaining gap sensing parameters, and collecting coal characteristic parameters of the target coal; performing coal gap superposition analysis based on the gap sensing parameters, calculating and obtaining coal gap parameters and a first gap error parameter, combining the coal characteristic parameters, indexing and calculating to obtain a second gap error parameter, and combining the first gap error parameter, calculating and obtaining a gap error parameter; sensing, collecting, and processing the density parameter of the target coal through a weight sensor, performing correction processing on the density parameter based on the coal gap parameter and the gap error parameter, and obtaining a corrected density parameter; processing and obtaining an ash content prediction accuracy parameter based on the coal gap parameter and the gap error parameter, performing ash content prediction based on the corrected density parameter, and obtaining a predicted ash content as a coal quality detection result.

[0007] In the second aspect, the present invention provides a multi-sensor fusion coal quality detection system, including: a parameter acquisition module, used to detect the target coal through a gap sensor, obtain gap sensing parameters, and collect coal characteristic parameters of the target coal; a gap analysis module, used to perform coal gap superposition analysis based on the gap sensing parameters, calculate and obtain coal gap parameters and a first gap error parameter, combine the coal characteristic parameters, index and calculate to obtain the second gap error parameter, and combine the first gap error parameter to calculate and obtain the gap error parameter; a density correction module, used to sense, collect and process the density parameters of the target coal through a weight sensor, perform correction processing on the density parameters according to the coal gap parameters and gap error parameters, and obtain a corrected density parameter; an ash prediction module, used to process and obtain ash prediction accuracy parameters according to the coal gap parameters and gap error parameters, perform ash prediction according to the corrected density parameters, and obtain predicted ash as a coal quality detection result.

[0008] The beneficial effects of the present invention are:

[0009] The target coal is detected using a gap sensor to obtain gap sensing parameters. Simultaneously, the target coal's characteristic parameters are collected, providing a preliminary understanding of the coal gap situation. Characteristic information such as coal size, type, and origin is also recorded, laying the data foundation for subsequent gap analysis and density correction. Based on the acquired gap sensing parameters, a coal gap overlay analysis is performed to calculate the coal gap parameters and the first gap error parameter. Then, combined with the coal characteristic parameters, an index calculation is performed to obtain the second gap error parameter. The first and second gap error parameters are then combined to calculate the final gap error parameter, enabling precise, multi-dimensional analysis of coal gaps and ensuring the accuracy and reliability of the gap parameters. A weight sensor is used to sense, collect, and process the density parameters of the target coal. Based on the acquired coal gap parameters and gap error parameters, the initially measured density parameters are corrected to obtain the corrected density parameters, enabling accurate density calculation and effectively eliminating the interference of gaps on density measurement. According to the coal gap parameters and gap error parameters, the ash content prediction accuracy parameters are processed and obtained, and then the ash content is predicted according to the corrected density parameters to obtain the predicted ash content as the coal quality detection result, realizing the dynamic adjustment of the ash content prediction accuracy and improving the accuracy of ash content prediction.

[0010] Through the above technical solution, the present application realizes the precise measurement and correction of coal gaps, effectively solves the problem of inaccurate density measurement caused by gaps in traditional coal quality testing, and improves the accuracy and reliability of coal ash content testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1A schematic flow chart of a multi-sensor fusion coal quality detection method provided by the present invention;

[0012] Figure 2 This is a structural schematic diagram of a multi-sensor fusion coal quality detection system provided by the present invention.

[0013] The reference numerals are as follows:

[0014] Parameter acquisition module 11, gap analysis module 12, density correction module 13, ash content prediction module 14. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0016] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0017] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0018] Example 1, as Figure 1 As shown, the embodiment of the present invention provides a multi-sensor fusion coal quality detection, including:

[0019] S1. Detect target coal through a gap sensor, obtain gap sensing parameters, and collect coal characteristic parameters of the target coal.

[0020] Specifically, the target coal is first scanned and inspected using a gap sensor. The target coal refers to the coal pile requiring quality inspection, such as coal in transit or stored in a storage bin. The gap sensor can employ various methods, including ultrasonic, X-ray, or optical sensors. By transmitting signals to the target coal and receiving reflected or transmitted signals, it accurately captures the distribution of voids between coal particles and obtains information about the coal's internal structure. Subsequently, data analysis is performed based on this internal structure information, and gap sensing parameters are output. For example, by performing spectral analysis, energy attenuation calculation, and binary image processing on the reflected or transmitted signals, gap characteristics are extracted, a void distribution model is established, and information about the coal's internal structure is obtained. These gap sensing parameters include numerical indicators such as void distribution density, average void size, and void area ratio, reflecting the degree of looseness of the target coal.

[0021] At the same time, coal characteristic parameters of the target coal are collected. These include information such as coal size, coal type, and origin. Coal size can be measured using a laser measuring device or image recognition to obtain coal particle size distribution parameters and calculate the average value as the coal size. Information such as coal type and origin can be obtained from the coal data management system. These coal characteristic parameters provide data support for subsequent gap analysis and correction, as coals of different sizes and types have different natural gap characteristics.

[0022] By obtaining gap sensing parameters and coal characteristic parameters, a data foundation is laid for subsequent gap analysis and density correction, providing support for accurate coal ash content detection.

[0023] S2. Perform coal gap superposition analysis based on the gap sensing parameters to calculate and obtain coal gap parameters and a first gap error parameter. Combined with the coal characteristic parameters, index calculation is performed to obtain a second gap error parameter. Combined with the first gap error parameter, a gap error parameter is calculated and obtained.

[0024] Specifically, the acquired gap sensor parameters are first input into a pre-built coal gap identifier sequence for processing. This coal gap identifier sequence consists of multiple trained gap identification models, each built based on a different dataset and algorithm, capable of identifying coal gap characteristics from different perspectives. By inputting the gap sensor parameters, each identifier outputs the corresponding identified coal gap parameters, forming an identified coal gap parameter sequence. For example, if there are n coal gap identifiers, n identified coal gap parameters will be obtained, forming an identified coal gap parameter sequence. Subsequently, the mean of the identified coal gap parameter sequence is calculated to obtain the coal gap parameter. This coal gap parameter represents the combined judgment results of multiple coal gap identifiers. To evaluate the reliability of this coal gap parameter, multiple identified coal gap parameters are randomly selected from the identified coal gap parameter sequence, and the average absolute deviation from the coal gap parameter (i.e., the mean) is calculated as the first gap error parameter. This first gap error parameter reflects the consistency of the results of different identifiers; a smaller value indicates a more reliable recognition result.

[0025] Next, based on the acquired coal characteristic parameters, the gap parameters of multiple coals of the same family are indexed in the coal testing database. This refers to historical test samples with similar characteristics (such as size, type, and origin) to the target coal. By comparing the gap parameters of these coals of the same family with the currently acquired coal gap parameters, the mean absolute deviation is calculated to obtain a second gap error parameter. The second gap error parameter reflects the degree of difference between the target coal and the historical samples of the same family. Subsequently, a comprehensive gap error parameter is calculated based on the first and second gap error parameters through weighted averaging or other methods. This gap error parameter takes into account both the internal consistency of the current test and the degree of conformity with historical data. This allows for a more comprehensive assessment of the accuracy of the gap parameters, providing a reliable error estimate for subsequent density correction and ash content prediction.

[0026] This multi-dimensional gap analysis method can effectively reduce errors in the gap identification process and improve the accuracy of coal quality detection.

[0027] S3. Using a weight sensor, sense, collect, and process the density parameter of the target coal, and perform correction processing on the density parameter according to the coal gap parameter and gap error parameter to obtain a corrected density parameter.

[0028] Specifically, the target coal is first weighed using a weight sensor to obtain its weight parameters. This weight sensor can be an industrial electronic scale, a conveyor belt weighing system, or other large-scale weighing equipment capable of accurately weighing the coal. Simultaneously, the detected volume of the target coal is determined using a volume measurement device. This volume measurement device can obtain the coal's volume information using technologies such as laser scanning, 3D imaging, or physical volumetric methods. Based on the obtained weight and volume information, the density parameter of the target coal is obtained by calculating the weight-to-volume ratio.

[0029] However, due to the large number of gaps between the particles in the target coal, the density parameter obtained by direct measurement is usually lower than the true density of the coal itself, which is also the main reason why the traditional coal quality detection method is not accurate. Therefore, the density parameter is corrected using the obtained coal gap parameter and gap error parameter. Specifically, the corrected coal gap parameter is calculated based on the coal gap parameter and the gap error parameter. The corrected coal gap parameter more accurately reflects the proportion of voids in the target coal. Subsequently, the density parameter is corrected based on the corrected coal gap parameter. For example, the density parameter can be corrected by the formula: corrected density parameter = density parameter / (1-corrected coal gap parameter) to obtain a corrected density parameter. In this way, the influence of the gap factor on the density measurement can be eliminated, and a corrected density parameter that is closer to the actual material density of the coal can be obtained.

[0030] By eliminating the influence of gap factors, more accurate density information can be provided, significantly improving the accuracy of the final ash content prediction. The above-mentioned density parameter processing method based on gap correction effectively solves the density measurement error caused by gaps in traditional coal quality testing.

[0031] S4. Process and obtain an ash content prediction accuracy parameter based on the coal gap parameter and gap error parameter, and perform ash content prediction based on the corrected density parameter to obtain a predicted ash content as a coal quality detection result.

[0032] Specifically, the obtained coal gap parameters and gap error parameters are first calculated through mathematical model processing to obtain the corrected coal gap parameter, which is used as the ash content prediction accuracy parameter. This ash content prediction accuracy parameter is used to determine the number of predictors to be called during the subsequent ash content prediction process to ensure the accuracy of the prediction results. A larger corrected coal gap parameter indicates more gaps in the target coal and higher measurement uncertainty. In this case, the number of predictors called is automatically increased to improve the reliability of the prediction results.

[0033] Then, a pre-built ash content predictor sequence was used to achieve ash content prediction. This ash content predictor sequence consists of multiple ash content prediction models (i.e., ash content predictors) trained using different training datasets and algorithms. Each ash content predictor outputs a corresponding ash content prediction value based on a corrected density parameter. Based on the previously obtained ash content prediction accuracy parameter, the number of ash content predictors required was dynamically calculated. When the gap is large and the measurement uncertainty is high, more ash content predictors are called for comprehensive judgment. When the gap is small and the measurement certainty is high, fewer ash content predictors are called for accurate prediction.

[0034] Subsequently, the calculated number of ash content predictors is randomly selected from the ash content predictor sequence, and the obtained corrected density parameters are input into each of these ash content predictors to obtain multiple independent ash content prediction values. These independent ash content prediction values ​​are then averaged to obtain the final predicted ash content, which is output as the coal quality test result.

[0035] By dynamically adjusting prediction accuracy based on gap parameters, combined with an ash content prediction mechanism that corrects density parameters, we can effectively address various gap differences and improve the accuracy and reliability of ash content prediction. Compared with traditional single-prediction methods, the multi-predictor fusion mechanism can better handle the uncertainties in coal testing and provide more accurate ash content data support for coal quality testing.

[0036] Furthermore, the target coal is detected by a gap sensor to obtain gap sensing parameters, and coal characteristic parameters of the target coal are collected, including:

[0037] S11. Detecting target coal through a gap sensor to obtain gap sensing parameters;

[0038] S12. Collect coal characteristic parameters of the target coal, wherein the coal characteristic parameters include coal size and coal type.

[0039] In one feasible implementation, the target coal is first detected using a gap sensor. This gap sensor can be of various types, such as ultrasonic, X-ray, or optical sensors. It transmits a corresponding signal (such as ultrasound, X-rays, or light) to illuminate the target coal and receives the reflected or transmitted signal. The signal received by the gap sensor contains information about the internal structure of the coal, particularly the distribution of voids between coal particles. The received signal is processed and analyzed, such as through spectrum analysis of ultrasonic signals, density analysis of X-ray images, and brightness and darkness analysis of optical images, to obtain gap sensing parameters. These gap sensing parameters include values ​​such as void distribution density, void size distribution, and void ratio, reflecting the porosity and internal structural characteristics of the target coal.

[0040] Next, characteristic coal parameters of the target coal are collected, including coal size and coal category. Coal size refers to the size of coal particles and can be measured through physical screening, laser particle size analyzers, or image recognition systems. Coal size is expressed as average particle size or particle size distribution range, such as coarse coal (>50mm), medium coal (25-50mm), fine coal (13-25mm), and fine coal (<13mm). Coal size has a direct impact on interstitial formation: larger coal size results in larger interstitials, while smaller coal size results in denser packing and smaller interstitials. Coal category refers to the classification of the target coal, including different types such as anthracite, bituminous coal, and lignite, or classifications based on origin, volatile matter content, and fixed carbon content. Different coal categories have different physical properties and chemical compositions. These differences affect the shape, surface characteristics, and packing patterns of coal particles, which in turn influence interstitial formation. Coal category can be quickly identified through pre-labeling, database lookup, or spectral characteristics of the coal.

[0041] Furthermore, according to the gap sensing parameters, a coal gap superposition analysis is performed to calculate and obtain the coal gap parameter and the first gap error parameter, including:

[0042] S21, inputting the gap sensing parameters into a pre-built coal gap identifier sequence, and identifying and outputting a coal gap parameter sequence;

[0043] S22. Calculate the mean of the identified coal gap parameter sequence to obtain a coal gap parameter;

[0044] S23. Randomly select a plurality of identified coal gap parameters in the identified coal gap parameter sequence, and calculate the average absolute deviation amplitude from the coal gap parameters as a first gap error parameter.

[0045] In a preferred embodiment, first, the obtained gap sensing parameters are input into a pre-constructed coal gap identifier sequence. The coal gap identifier sequence is a model set composed of multiple independent coal gap identifiers, each of which is constructed based on different algorithm principles or training data, and can identify coal gap characteristics from different angles. For example, the coal gap identifier sequence contains multiple coal gap identifiers constructed based on different machine learning algorithms such as neural networks, support vector machines, and random forests. Each coal gap identifier receives the same gap sensing parameters as input, but due to differences in model structure and training data, it will output its own recognition results to form a series of identified coal gap parameters, that is, an identified coal gap parameter sequence. This multi-model parallel recognition method can analyze coal gaps from multiple angles and reduce the one-sidedness and errors that may be caused by a single model.

[0046] The obtained coal gap parameter sequence is then statistically processed, and the mean of all identification results in the sequence is calculated to obtain a comprehensive coal gap parameter. This coal gap parameter represents the collective judgment results of multiple coal gap identifiers and can more accurately reflect the actual gap situation of the target coal. Using mean calculation can effectively reduce the influence of accidental errors of individual coal gap identifiers and improve the reliability of the gap parameter.

[0047] To assess the reliability of the coal gap parameters, multiple identification results are randomly selected from the coal gap parameter sequence. The mean absolute deviation between these results and the previously obtained coal gap parameter (i.e., the mean) is calculated as the first gap error parameter. This is calculated by taking the absolute value of the difference between each randomly selected identified coal gap parameter and the coal gap parameter (mean), and then averaging these absolute values ​​to obtain the first gap error parameter. This first gap error parameter reflects the degree of dispersion of the results from different identifiers. Smaller values ​​indicate more consistent identification results from the coal gap identifiers and higher reliability of the coal gap parameter. Larger values ​​indicate greater dispersion of the identification results and lower reliability.

[0048] Through the above steps, multi-angle identification and comprehensive evaluation of coal gaps are achieved, which not only obtains accurate coal gap parameters, but also quantifies the reliability of the identification results through the first gap error parameter, providing a more reliable data basis for subsequent density correction and ash content prediction, thereby effectively improving the accuracy and stability of coal gap identification.

[0049] Furthermore, the step of constructing the coal gap identifier sequence includes:

[0050] S211. Based on historical data of coal gap identification, a sample gap sensor parameter set is collected, and coal gap parameters under different sample gap sensor parameters are tested and collected, and the obtained sample coal gap parameter set is marked;

[0051] S212, using machine learning to build multiple coal gap identifiers;

[0052] S213. Randomly divide the sample gap sensor parameter set and the sample coal gap parameter set to obtain multiple coal gap recognition training data, train the multiple coal gap identifiers respectively until the accuracy meets the requirements, and obtain a coal gap identifier sequence.

[0053] In a preferred embodiment, data collection and labeling are first performed to prepare a data foundation for subsequent machine learning training. Specifically, based on historical data on coal gap identification, a large number of sample gap sensor parameters are collected to form a sample gap sensor parameter set. These sample gap sensor parameters come from coal samples of different types, sizes, and stacking states, and have good representativeness and diversity. At the same time, these coal samples are laboratory tested, and the actual gap parameters of each sample are accurately measured using professional coal gap measurement methods (such as immersion method, gas displacement method, image analysis method, etc.), and labeled to obtain a sample coal gap parameter set that corresponds one-to-one with the sample gap sensor parameters. These two data sets constitute the input features and output labels of machine learning, providing basic data support for subsequent model training.

[0054] Then, using machine learning techniques, multiple coal gap identifiers with different structures are constructed. These identifiers can be based on different machine learning algorithms, such as artificial neural networks, support vector machines, decision trees, random forests, and gradient boosting trees. Each algorithm has its own unique characteristics. By constructing coal gap identifiers based on multiple algorithms, the mapping relationship between gap sensing parameters and actual coal gap parameters can be learned from different perspectives, improving the robustness and accuracy of overall identification. Furthermore, even when using the same type of algorithm, multiple differentiated coal gap identifiers can be constructed by adjusting the network structure and hyperparameters, further enhancing model diversity.

[0055] Next, the previously obtained sample gap sensor parameter sets and sample coal gap parameter sets are randomly partitioned to obtain multiple different sets of coal gap identification training data. This random partitioning can utilize data sampling techniques such as cross-validation and bootstrapping to ensure that each set of training data has both commonalities and differences. Subsequently, the constructed multiple coal gap identifiers are trained using these different coal gap identification training data, adjusting model parameters until each coal gap identifier achieves the preset accuracy requirement on its respective validation dataset. This results in multiple coal gap identifiers with qualified performance, forming a coal gap identifier sequence.

[0056] The coal gap identifier sequence, constructed through a data-driven approach, exhibits strong generalization and robustness, enabling accurate identification of gap parameters for different types of coal. Furthermore, because each coal gap identifier is built based on a different algorithm or training data, they complement each other to a certain degree, enabling analysis of coal gaps from multiple perspectives, improving identification accuracy and reliability, and providing intelligent, high-precision gap identification capabilities for coal quality testing.

[0057] Furthermore, combining the coal characteristic parameters, indexing and calculating to obtain a second gap error parameter, and combining the first gap error parameter, calculating and obtaining a gap error parameter, including:

[0058] S24. According to the coal characteristic parameters, obtain a plurality of coal gap parameters of a plurality of coals of the same family by indexing in a coal detection database, wherein the coal detection database includes sample detection data of a plurality of sample coals;

[0059] S25. Calculate, based on the plurality of coal gap parameters of the same family, an average absolute deviation from the coal gap parameter as a second gap error parameter;

[0060] S26. Calculate and obtain a gap error parameter based on the first gap error parameter and the second gap error parameter.

[0061] In a preferred embodiment, first, the acquired coal characteristic parameters (such as coal size, coal category, etc.) are used as search conditions to query coal samples with similar characteristics, i.e., coal of the same family, in a pre-established coal detection database. The so-called coal of the same family refers to historical detection samples that are similar or close to the target coal currently being detected in terms of size, category, origin, and other characteristics. The coal detection database is a collection of coal detection records accumulated over a long period of time, which contains a large amount of sample detection data of coal of different types and sources, including key indicators such as gap parameters, density parameters, and ash content. By matching and retrieving characteristic parameters, multiple coal samples of the same family can be found in the coal detection database, and the gap parameters of these samples can be extracted to obtain multiple gap parameters of coal of the same family.

[0062] Then, using the obtained gap parameters of multiple coals of the same family, the average absolute deviation amplitude between the obtained coal gap parameters is calculated as the second gap error parameter. The specific calculation method is to take the absolute value of the difference between each gap parameter of the same family coal and the current coal gap parameter, and then calculate the average of these absolute values ​​to obtain the second gap error parameter. The second gap error parameter reflects the degree of difference between the gap parameters of the currently identified coal and historical coals of the same type. If the second gap error parameter is small, it indicates that the gap characteristics of the currently identified coal gap parameters and historical coals of the same type are highly consistent, and the recognition reliability is good; if the second gap error parameter is large, it may indicate that there is an abnormality in the gap parameters of the currently identified coal, or that the target coal has special characteristics and requires further analysis and correction.

[0063] Subsequently, a comprehensive calculation is performed based on the obtained first and second gap error parameters to obtain the final gap error parameter. This calculation can employ various methods, such as weighted averaging, maximum value selection, and weighted fusion, to combine the results of internal consistency verification (the first gap error parameter) and external empirical verification (the second gap error parameter) to achieve a more comprehensive error assessment. For example, weight coefficients can be set based on the reliability or importance of the two error parameters, and a weighted average calculation can be performed. Alternatively, the larger of the two values ​​can be selected for a more conservative error estimation strategy. Through this comprehensive process, the final gap error parameter is obtained, which takes into account both the internal consistency of the current identification process and the degree of consistency with historical empirical data, enabling a more comprehensive and accurate assessment of the reliability of coal gap parameters.

[0064] By introducing historical data comparison and multi-source error fusion, the accuracy and reliability of coal gap parameters are effectively improved, laying a solid foundation for subsequent density correction and ash content prediction, thereby effectively dealing with the uncertainty and complexity in coal detection.

[0065] Furthermore, the density parameter of the target coal is sensed, collected and processed by a weight sensor, and the density parameter is corrected according to the coal gap parameter and the gap error parameter to obtain a corrected density parameter, including:

[0066] S31. Using a weight sensor, sense and collect the weight parameter of the target coal, detect the detection volume of the target coal, and calculate and obtain the density parameter;

[0067] S32. Calculate and obtain a corrected coal gap parameter based on the coal gap parameter and the gap error parameter;

[0068] S33. Perform correction calculation processing on the density parameter according to the corrected coal gap parameter to obtain a corrected density parameter.

[0069] In a preferred embodiment, the target coal is first weighed using a weight sensor to obtain the weight parameters of the target coal. The weight sensor can be an industrial electronic scale, a conveyor belt weighing system, a pressure sensor, or other equipment. A suitable weighing device is selected based on the detection scenario and the amount of coal. Simultaneously, the detection volume of the target coal is detected and collected. The detection volume can be measured in a variety of ways, such as by scanning the surface contour of the coal pile with a three-dimensional laser scanning system and calculating the volume based on the bottom geometry; by taking images of the coal pile from multiple angles and estimating the volume using three-dimensional reconstruction technology; or for small batches of coal, the volume can be directly measured using a standard container. After obtaining the weight parameters and detection volume, the density parameter of the target coal is obtained by calculating their ratio: density parameter = weight parameter / detection volume. However, due to the presence of gaps between coal particles, the obtained density parameter is often lower than the actual density of the coal material and requires further correction.

[0070] Then, the corrected coal gap parameters are calculated using the obtained coal gap parameters and gap error parameters. The corrected coal gap parameters are optimized adjustments to the original coal gap parameters, taking into account the reliability and uncertainty of the gap parameters. For example, a table lookup method combined with interpolation calculation is used to obtain the corrected coal gap parameters. Specifically, a correction table is first established. This table, constructed based on a large amount of experimental data, contains correction coefficients corresponding to different gap error parameter value ranges. The rows of the correction table represent coal gap parameter values ​​for different intervals, and the columns represent gap error parameter values ​​for different intervals. The values ​​in the correction table are the correction coefficients under the corresponding conditions. Then, based on the obtained coal gap parameters and gap error parameters, the closest parameter is searched in the correction table to obtain the corresponding correction coefficient. If the actual gap error parameter value falls between two adjacent discrete points in the correction table, a more accurate correction coefficient is calculated using linear interpolation. The original coal gap parameters are then multiplied by the obtained correction coefficient to calculate the corrected coal gap parameters. This correction process yields more reliable corrected coal gap parameters, providing accurate basic data for subsequent density calibration.

[0071] The calculated density parameter is then corrected using the obtained corrected coal gap parameter to obtain a corrected density parameter. Due to the presence of gaps, the directly measured coal weight-to-volume ratio (density parameter) is lower than the actual density of the coal material. By correcting the density parameter, the effect of gaps on volume is accounted for, subtracting the voids from the measured volume and calculating only the volume of the solid coal, resulting in a more accurate density value. Specifically, the density parameter is divided by the proportion of solid coal (i.e., 1 minus the corrected coal gap parameter) to obtain the corrected density parameter. This correction effectively eliminates the effect of gaps on density measurement, resulting in a corrected density parameter that is closer to the actual density of the coal material.

[0072] The above steps effectively address the problem of under-measured density due to gaps in traditional coal quality testing. The corrected density parameter more accurately reflects the actual density characteristics of the coal, providing reliable data for subsequent ash content prediction.

[0073] Furthermore, according to the coal gap parameter and gap error parameter, an ash content prediction accuracy parameter is obtained, and according to the corrected density parameter, an ash content prediction is performed to obtain a predicted ash content as a coal quality test result, including:

[0074] S41. Calculate and obtain a corrected coal gap parameter based on the coal gap parameter and the gap error parameter, as an ash content prediction accuracy parameter;

[0075] S42. Based on the pre-built ash content predictor sequence, the ash content prediction accuracy parameter and the number of ash content predictors in the ash content predictor sequence are used to calculate the number of calls.

[0076] S43. Randomly select a number of ash content predictors to be called, output the corrected density parameters respectively, obtain multiple individual predicted ash contents, calculate the mean, and obtain the predicted ash contents.

[0077] In a preferred embodiment, first, based on the obtained coal gap parameter and gap error parameter, a processing method similar to S32 is adopted to calculate and obtain the corrected coal gap parameter, and use it as the ash prediction accuracy parameter. The corrected coal gap parameter here is not only used for density correction, but also plays the role of controlling the ash prediction accuracy. The size of the corrected coal gap parameter directly reflects the complexity of the target coal and the uncertainty of the measurement. The larger the gap or the larger the gap measurement error, the more complex the coal structure and the higher the difficulty of ash prediction. At this time, a more accurate prediction strategy is required. By using the corrected coal gap parameter as the ash prediction accuracy parameter, the prediction strategy can be dynamically adjusted according to the characteristics of the target coal to achieve adaptive prediction.

[0078] Then, based on a pre-built ash predictor sequence, the number of ash predictors to be called is dynamically calculated according to the ash prediction accuracy parameter. The ash predictor sequence consists of multiple independent ash predictors, each trained to predict the corresponding coal ash content based on an input corrected density parameter. For the ash predictor sequence, a training dataset is first constructed by collecting sample-corrected density parameters and laboratory-measured ash contents from a large number of historical coal samples. Multiple ash predictors are then trained using different machine learning algorithms (such as linear regression, polynomial regression, support vector regression, neural networks, and random forests), or using the same algorithm with different parameter configurations, to form an ash predictor sequence. These ash predictors have different characteristics and can analyze the relationship between density and ash from different perspectives. For example, a base number of ash predictors is set and then adjusted based on the ash prediction accuracy parameter to determine the number of ash predictors to be called. When the ash prediction accuracy parameter is large (i.e., the coal gap is large or the gap measurement uncertainty is high), the number of calls is increased to improve the prediction stability by integrating more ash predictors; when the ash prediction accuracy parameter is small, fewer ash predictors can be used to complete the task, improving processing efficiency.

[0079] Next, a calculated number of ash predictors are randomly selected from the ash predictor sequence, and the corrected density parameters are input into each of these predictors to obtain multiple independent ash prediction results, known as individual predicted ash contents. Each ash predictor processes the input corrected density parameters based on its own model characteristics and training data, outputting a predicted ash content as the individual predicted ash content. The mean of these individual predicted ash contents is then calculated to obtain the final predicted ash content, which is then output as the coal quality test result. The collective judgment of multiple ash predictors can reduce the bias and fluctuations that may be introduced by a single model, improving the stability and reliability of the prediction results. Furthermore, the random selection of ash predictors increases the randomness of the prediction process, preventing the ensemble results from being overly influenced by specific predictors, further improving the robustness of the prediction. Furthermore, by dynamically adjusting the number of calls, the prediction strategy can be adaptively selected based on the characteristics of different target coals, ensuring accuracy while optimizing computing resource utilization.

[0080] Example 2, as Figure 2 As shown, based on the same inventive concept of a multi-sensor fusion coal quality detection method provided in Example 1, an embodiment of the present invention further provides a multi-sensor fusion coal quality detection system, including:

[0081] The parameter acquisition module 11 is used to detect the target coal through the gap sensor, obtain the gap sensing parameters, and collect the coal characteristic parameters of the target coal;

[0082] a gap analysis module 12 configured to perform coal gap superposition analysis based on the gap sensing parameters, calculate and obtain a coal gap parameter and a first gap error parameter, combine the coal characteristic parameters, perform index calculation to obtain a second gap error parameter, and calculate and obtain a gap error parameter based on the first gap error parameter;

[0083] The density correction module 13 is configured to sense, collect, and process the density parameter of the target coal using a weight sensor, and perform correction processing on the density parameter according to the coal gap parameter and gap error parameter to obtain a corrected density parameter;

[0084] The ash content prediction module 14 is used to process and obtain the ash content prediction accuracy parameter according to the coal gap parameter and the gap error parameter, and perform ash content prediction according to the corrected density parameter to obtain the predicted ash content as the coal quality detection result.

[0085] Furthermore, the parameter acquisition module 11 includes the following execution steps:

[0086] Through the gap sensor, the target coal is detected to obtain the gap sensing parameters;

[0087] The coal characteristic parameters of the target coal are collected, wherein the coal characteristic parameters include coal size and coal category.

[0088] Furthermore, the gap analysis module 12 includes the following execution steps:

[0089] Input the gap sensing parameters into a pre-built coal gap identifier sequence, and identify the output to obtain an identified coal gap parameter sequence;

[0090] Calculating the mean of the identified coal gap parameter sequence to obtain the coal gap parameter;

[0091] A plurality of identified coal gap parameters are randomly selected from the identified coal gap parameter sequence, and an average absolute deviation amplitude from the coal gap parameters is calculated as a first gap error parameter.

[0092] Furthermore, the step of constructing the coal gap identifier sequence includes:

[0093] According to the historical data of coal gap identification, a sample gap sensor parameter set is collected, and the gap parameters of coal under different sample gap sensor parameters are tested and collected, and the sample coal gap parameter set is marked;

[0094] Using machine learning, we built multiple coal gap identifiers;

[0095] The sample gap sensor parameter set and the sample coal gap parameter set are randomly divided to obtain multiple coal gap recognition training data, and the multiple coal gap identifiers are trained respectively until the accuracy meets the requirements to obtain a coal gap identifier sequence.

[0096] Furthermore, the gap analysis module 12 further includes the following execution steps:

[0097] According to the coal characteristic parameters, a plurality of coal gap parameters of a plurality of coals of the same family are obtained by indexing in a coal detection database, wherein the coal detection database includes sample detection data of a plurality of sample coals;

[0098] Calculating, based on the plurality of coal gap parameters of the same family, an average absolute deviation amplitude from the coal gap parameter as a second gap error parameter;

[0099] A gap error parameter is obtained by calculation according to the first gap error parameter and the second gap error parameter.

[0100] Furthermore, the density correction module 13 includes the following execution steps:

[0101] The weight parameter of the target coal is sensed and collected by a weight sensor, and the detection volume of the target coal is detected and collected, and the density parameter is calculated;

[0102] Calculating and obtaining a corrected coal gap parameter according to the coal gap parameter and the gap error parameter;

[0103] According to the corrected coal gap parameter, the density parameter is corrected and calculated to obtain a corrected density parameter.

[0104] Furthermore, the ash content prediction module 14 includes the following execution steps:

[0105] According to the coal gap parameter and the gap error parameter, a corrected coal gap parameter is calculated and obtained as an ash content prediction accuracy parameter;

[0106] Based on a pre-built ash content predictor sequence, the ash content prediction accuracy parameter and the number of ash content predictors in the ash content predictor sequence are used to calculate the number of calls;

[0107] A number of ash content predictors are randomly selected for calling, and the corrected density parameters are outputted respectively to obtain a plurality of individually predicted ash contents, and the mean is calculated to obtain the predicted ash contents.

[0108] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0109] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0110] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0111] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0112] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0113] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0114] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A coal quality detection method based on multi-sensor fusion, characterized in that: The method comprises: Detecting target coal through a gap sensor, obtaining gap sensing parameters, and collecting coal characteristic parameters of the target coal; Perform coal gap superposition analysis based on the gap sensing parameters to calculate and obtain coal gap parameters and a first gap error parameter. Combined with the coal characteristic parameters, index calculation is performed to obtain a second gap error parameter. Combined with the first gap error parameter, a gap error parameter is calculated. The density parameter of the target coal is sensed, collected and processed by a weight sensor, and the density parameter is corrected according to the coal gap parameter and the gap error parameter to obtain a corrected density parameter; According to the coal gap parameter and gap error parameter, an ash content prediction accuracy parameter is obtained by processing, and according to the corrected density parameter, an ash content prediction is performed to obtain a predicted ash content as a coal quality test result; The method includes sensing, collecting, and processing the density parameter of the target coal by a weight sensor, and performing correction processing on the density parameter according to the coal gap parameter and the gap error parameter to obtain the corrected density parameter, including: The weight parameter of the target coal is sensed and collected by a weight sensor, and the detection volume of the target coal is detected and collected, and the density parameter is calculated; Calculating and obtaining a corrected coal gap parameter according to the coal gap parameter and the gap error parameter; According to the corrected coal gap parameter, the density parameter is corrected and calculated to obtain a corrected density parameter; The method includes: obtaining an ash content prediction accuracy parameter based on the coal gap parameter and the gap error parameter; performing ash content prediction based on the corrected density parameter; and obtaining a predicted ash content as a coal quality test result. According to the coal gap parameter and the gap error parameter, a corrected coal gap parameter is calculated and obtained as an ash content prediction accuracy parameter; Based on a pre-built ash content predictor sequence, the ash content prediction accuracy parameter and the number of ash content predictors in the ash content predictor sequence are used to calculate the number of calls; A number of ash content predictors are randomly selected for calling, and the corrected density parameters are outputted respectively to obtain a plurality of individually predicted ash contents, and the mean is calculated to obtain the predicted ash contents.

2. The multi-sensor fusion coal quality detection method according to claim 1, characterized in that: The target coal is detected by the gap sensor to obtain gap sensing parameters and collect coal characteristic parameters of the target coal, including: Through the gap sensor, the target coal is detected to obtain the gap sensing parameters; The coal characteristic parameters of the target coal are collected, wherein the coal characteristic parameters include coal size and coal category.

3. The coal quality detection method based on multi-sensor fusion according to claim 1, characterized in that: According to the gap sensing parameters, coal gap superposition analysis is performed to calculate and obtain coal gap parameters and first gap error parameters, including: Input the gap sensing parameters into a pre-built coal gap identifier sequence, and identify the output to obtain an identified coal gap parameter sequence; Calculating the mean of the identified coal gap parameter sequence to obtain the coal gap parameter; A plurality of identified coal gap parameters are randomly selected from the identified coal gap parameter sequence, and an average absolute deviation amplitude from the coal gap parameters is calculated as a first gap error parameter.

4. The coal quality detection method based on multi-sensor fusion according to claim 3 is characterized in that: The steps of constructing the coal gap identifier sequence include: According to the historical data of coal gap identification, a sample gap sensor parameter set is collected, and the gap parameters of coal under different sample gap sensor parameters are tested and collected, and the sample coal gap parameter set is marked; Using machine learning, we built multiple coal gap identifiers; The sample gap sensor parameter set and the sample coal gap parameter set are randomly divided to obtain multiple coal gap recognition training data, and the multiple coal gap identifiers are trained respectively until the accuracy meets the requirements to obtain a coal gap identifier sequence.

5. The coal quality detection method based on multi-sensor fusion according to claim 1, characterized in that: In combination with the coal characteristic parameters, index calculation is performed to obtain a second gap error parameter, and in combination with the first gap error parameter, calculation is performed to obtain a gap error parameter, including: According to the coal characteristic parameters, a plurality of coal gap parameters of a plurality of coals of the same family are obtained by indexing in a coal detection database, wherein the coal detection database includes sample detection data of a plurality of sample coals; Calculating, based on the plurality of coal gap parameters of the same family, an average absolute deviation amplitude from the coal gap parameter as a second gap error parameter; A gap error parameter is obtained by calculation according to the first gap error parameter and the second gap error parameter.

6. A multi-sensor fusion coal quality detection system, characterized in that: A system for implementing the multi-sensor fusion coal quality detection method according to any one of claims 1 to 5, comprising: A parameter acquisition module is used to detect target coal through a gap sensor, obtain gap sensing parameters, and collect coal characteristic parameters of the target coal; a gap analysis module, configured to perform coal gap superposition analysis based on the gap sensing parameters, calculate and obtain a coal gap parameter and a first gap error parameter, combine the coal characteristic parameters, perform index calculation to obtain a second gap error parameter, and calculate and obtain a gap error parameter based on the first gap error parameter; a density correction module, configured to sense, collect, and process the density parameter of the target coal using a weight sensor, and perform correction processing on the density parameter according to the coal gap parameter and the gap error parameter to obtain a corrected density parameter; The ash content prediction module is used to process and obtain the ash content prediction accuracy parameter according to the coal gap parameter and the gap error parameter, and to perform ash content prediction according to the corrected density parameter to obtain the predicted ash content as the coal quality detection result.

Images