Multi-dimensional adaptive coal ash content on-line detection system and method

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a system which comprises a signal processing module, a signal feature reconstruction module, a signal feature optimization module, a coal quality ash content judgment module and an ash content value verification module, the signal processing module is used for carrying out noise suppression processing on a radiation signal of coal to obtain a preprocessing signal of the coal; the signal feature reconstruction module is used for performing adaptive feature reconstruction on the preprocessed signal to obtain an optimized feature signal of the preprocessed signal; the signal feature optimization module is used for establishing a dynamic coupling relationship between the environmental parameters and the optimized feature signal, and performing real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain an environmental adaptive signal of the optimized feature signal; according to the invention, the precision and the result credibility of coal ash content online detection can be improved, and the detection system is ensured to continuously output stable and reliable detection results in an industrial online detection scene.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a multidimensional adaptive online coal ash content detection system and method. Background Technology

[0002] In the field of online coal ash content detection, existing technologies have significant shortcomings in noise suppression of coal radiation signals. They often only achieve simple interference filtering and cannot accurately suppress the complex and variable noise sources in coal detection scenarios. This results in preprocessed signals still carrying a significant amount of interference, making it difficult to meet the accuracy requirements of subsequent ash content detection. Furthermore, existing technologies lack effective adaptive adjustment capabilities in the signal feature processing stage, making it difficult to optimize feature extraction based on differences in signal quality. This results in a lack of reliable data foundation for subsequent ash content analysis based on signal features, directly affecting the accuracy of ash content detection results.

[0003] Existing online coal ash content detection technologies suffer from poor environmental adaptability and a lack of result verification. In real-world industrial scenarios, environmental parameters (such as ambient humidity, coal thickness, and conveyor belt speed) are constantly changing. However, current technologies cannot dynamically correlate these environmental parameters with the detection signal. When environmental parameters fluctuate, the detection signal is prone to drift, and the lack of a real-time compensation mechanism leads to significant deviations in the detection results as the environment changes. Furthermore, after obtaining the ash content value, existing technologies typically lack a dedicated reliability verification process, making it impossible to determine whether the ash content value is within a reasonable range. This can easily lead to the output of abnormal detection values ​​as valid results, making it difficult to guarantee the reliability of the detection results and failing to meet the actual needs of industrial production for high precision and high stability in online coal ash content detection. Summary of the Invention

[0004] This invention provides a multidimensional adaptive online coal ash content detection system and method to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this invention provides a multidimensional adaptive online coal ash content detection system. The system includes a signal processing module, a signal feature reconstruction module, a signal feature optimization module, a coal ash content judgment module, and an ash content verification module, wherein:

[0006] The signal processing module is used to perform noise suppression processing on the radiation signal of the coal to obtain the preprocessed signal of the coal.

[0007] The signal feature reconstruction module is used to perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal;

[0008] The signal feature optimization module is used to establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and to perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain an environmental adaptive signal of the optimized feature signal.

[0009] The coal ash content judgment module is used to perform ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal to obtain the ash content value of the coal.

[0010] The ash content verification module is used to verify the reliability of the ash content value and obtain the ash content test result of the coal.

[0011] In a preferred embodiment, when performing noise suppression processing on the radiation signal of the coal, the signal processing module is specifically used for:

[0012] The radiation signal is subjected to anti-interference processing to obtain the radiation signal anti-interference signal;

[0013] The anti-interference signal is smoothed to obtain a smoothed signal of the anti-interference signal;

[0014] The smoothed signal is subjected to noise reduction processing to obtain the preprocessed signal of the coal.

[0015] In a preferred embodiment, when performing adaptive feature reconstruction on the preprocessed signal, the signal feature reconstruction module is specifically used for:

[0016] The preprocessed signal is evaluated for signal quality to obtain the signal quality index of the preprocessed signal.

[0017] Based on the signal quality index, the preprocessed signal is decomposed into multi-scale features to obtain the feature components of the preprocessed signal.

[0018] Adaptive filtering is performed on the feature components to obtain the effective feature components of the preprocessed signal;

[0019] The effective feature components are weighted, fused, and reconstructed to obtain the optimized feature signal.

[0020] In a preferred embodiment, when performing adaptive filtering of the feature components, the signal feature reconstruction module is specifically used for:

[0021] The stability of the signal quality index is analyzed to obtain the stability evaluation results of the signal quality index.

[0022] The stability evaluation results of the indicators are mapped using rules to obtain a parameter adjustment strategy for feature selection.

[0023] Based on the parameter adjustment strategy, the threshold and weight of the filtering rules are configured to obtain the filtering rules for the feature components;

[0024] The feature components are iteratively filtered based on the filtering rules to obtain the effective feature components.

[0025] In a preferred embodiment, when the signal feature optimization module performs the process of establishing a dynamic coupling relationship between environmental parameters and the optimized feature signal, it is specifically used for:

[0026] Under standard operating conditions, the optimized feature signal is collected as a reference signal, and environmental parameters are collected as reference values.

[0027] Under varying operating conditions, the real-time values ​​of the optimized feature signal and the environmental parameters are collected. The real-time value of the optimized feature signal is compared with the reference signal to obtain the signal drift of the optimized feature signal.

[0028] The real-time values ​​of the environmental parameters are combined with the corresponding signal drift to obtain a training sample set;

[0029] The training sample set is input into a preset initial neural network, and the connection weights of the initial neural network are adjusted to obtain a nonlinear mapping model between the environmental parameters and the signal drift.

[0030] When the amount of incremental sample data reaches the preset batch size, or when the statistical variance of the signal drift exceeds the stability threshold, the connection weights of the nonlinear mapping model are updated to obtain the dynamically updated model.

[0031] In a preferred embodiment, when the signal feature optimization module performs real-time compensation of the optimized feature signal based on the dynamic coupling relationship, it is specifically used for:

[0032] Obtain the current environmental parameters and input them into the nonlinear mapping model to obtain the compensation amount of the optimized feature signal, wherein the compensation amount is calculated using the following formula:

[0033]

[0034] The compensation amount is applied to the optimized feature signal to obtain the environment adaptive signal, and the calculation formula for the environment adaptive signal is as follows:

[0035] ;

[0036] In the formula, Indicates the current belt speed. Indicates the standard belt speed; Indicates the current coal thickness. Indicates the thickness of standard coal; Indicates the current ambient humidity. Indicates standard ambient humidity; , , These represent the influence coefficients of belt speed, coal thickness, and ambient humidity, respectively, which are output in real time by the nonlinear mapping model based on the current environmental parameters. This represents the optimized feature signal; This refers to the environmental adaptive signal.

[0037] In a preferred embodiment, when the coal ash content determination module performs an ash content fusion determination of the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for:

[0038] Based on the historical database, the ash content of the environmental adaptive signal is initially predicted to obtain the signal ash content value of the coal.

[0039] Obtain the coal quality parameters of the current batch of coal, perform a coupled analysis of moisture and calorific value on the coal quality parameters, and obtain the compensated ash content value of the coal.

[0040] The confidence levels of the signal gray value and the compensated gray value are evaluated to obtain the signal confidence level of the signal gray value and the parameter confidence level of the compensated gray value.

[0041] Based on the signal confidence level and the parameter confidence level, the signal ash value and the compensated ash value are weighted and fused to obtain the ash value of the coal, wherein the final ash value is calculated using the following formula:

[0042]

[0043] In the formula, The confidence level of the signal. The gray value of the signal. The confidence level of the parameter. The compensation ash value is given.

[0044] In a preferred embodiment, when the coal ash content determination module performs a confidence assessment of the signal ash content value and the compensated ash content value, it is specifically used for:

[0045] Variance calculation is performed on the fluctuation data of the environmental adaptive signal within a preset time window to obtain the signal fluctuation quantization value;

[0046] The signal fluctuation quantization value is compared with a preset stability threshold to obtain the signal reliability factor;

[0047] The ratio of moisture value to calorific value in the coal quality parameters is calculated to obtain the actual ratio of moisture to calorific value.

[0048] The difference between the actual ratio of moisture to calorific value and the parameter range of standard coal type is calculated to obtain the parameter deviation.

[0049] The parameter deviation is input into a preset rationality mapping table for query and matching to obtain the parameter rationality factor;

[0050] Retrieve historical consistency records of current coal source information from the database;

[0051] The signal reliability factor, the parameter rationality factor, and the historical detection consistency record are fused together to obtain the signal confidence level and the parameter confidence level.

[0052] In a preferred embodiment, when the ash content verification module performs an ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for:

[0053] The ash content value is input into a preset reasonable ash content range for range comparison;

[0054] When the ash content value is within the reasonable range, the ash content value is output as the ash content detection result;

[0055] When the ash content value exceeds the reasonable range, an anomaly flag is output as the ash content detection result.

[0056] To address the above problems, the present invention also provides a multidimensional adaptive online detection method for coal ash content, the method comprising:

[0057] S1. Perform noise suppression processing on the radiation signal of the coal to obtain the preprocessed signal of the coal;

[0058] S2. Perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal;

[0059] S3. Establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain the environmental adaptive signal of the optimized feature signal;

[0060] S4. Based on the coal quality parameters and the environmental adaptive signal, perform ash content fusion judgment on the coal to obtain the ash content value of the coal.

[0061] S5. Verify the reliability of the ash content value to obtain the ash content test result of the coal.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] 1. Noise suppression processing of coal radiation signals through the signal processing module effectively improves the purity of the preprocessed signal, providing high-quality data support for subsequent ash content detection. Adaptive feature reconstruction of the preprocessed signal using the signal feature reconstruction module accurately extracts optimized feature signals that better meet the needs of ash content detection. Establishing a dynamic coupling relationship between environmental parameters and optimized feature signals and implementing real-time compensation through the signal feature optimization module allows the environmental adaptive signal to accurately match the characteristics of the actual coal detection scenario. The ash content fusion judgment module, which integrates coal quality parameters and environmental adaptive signals, improves the accuracy of ash content calculation. Finally, the ash content verification module verifies the credibility of the ash content value, ensuring that the final output ash content detection results have high validity, significantly improving the accuracy and reliability of online coal ash content detection.

[0064] 2. Further enhance the stability and scenario adaptability of online coal ash content detection. During signal processing, targeted noise suppression further reduces the impact of complex interference on the detection signal, ensuring stable signal quality. The signal feature reconstruction stage, through signal quality assessment and multi-scale feature decomposition, accurately selects effective feature components, improving the representativeness of the optimized feature signal. The signal feature optimization stage compensates based on dynamically updated coupling relationships, adapting to changes in environmental parameters in real time and avoiding detection deviations caused by environmental fluctuations. The coal ash content judgment stage combines confidence assessment with weighted fusion, fully leveraging the synergistic effect of signals and parameters to further optimize the reliability of ash content values. Reasonable range verification in the ash content value validation stage effectively eliminates abnormal detection results, ensuring the detection system continuously outputs stable and reliable detection results in industrial online detection scenarios, meeting the high standards required for coal ash content detection in actual production.

[0065] Figure caption attached

[0066] Figure 1 This is a system architecture diagram of a multidimensional adaptive online coal ash content detection system provided in an embodiment of the present invention;

[0067] Figure 2 This is a schematic flowchart of a multidimensional adaptive online coal ash content detection method provided in an embodiment of the present invention.

[0068] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments belong to some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “said” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0071] Depending on the context, the word "if" or "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0072] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0073] In practice, the server-side equipment deployed in a multi-dimensional adaptive online coal ash content detection system may consist of one or more devices. This multi-dimensional adaptive online coal ash content detection system can be implemented as: a business instance, a virtual machine, or hardware devices. For example, this multi-dimensional adaptive online coal ash content detection system can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, this multi-dimensional adaptive online coal ash content detection system can be understood as software deployed on a cloud node, used to provide multi-dimensional adaptive online coal ash content detection to various user terminals. Alternatively, this multi-dimensional adaptive online coal ash content detection system can also be implemented as a virtual machine deployed on one or more devices in a cloud node. This virtual machine contains application software for managing various user terminals. Alternatively, this multi-dimensional adaptive online coal ash content detection system can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide multi-dimensional adaptive online coal ash content detection to various user terminals.

[0074] In terms of implementation, the multidimensional adaptive online coal ash content detection system and the user terminal are mutually adaptive. That is, if the multidimensional adaptive online coal ash content detection system is implemented as an application installed on a cloud service platform, then the user terminal is a client that establishes a communication connection with the application; or if the multidimensional adaptive online coal ash content detection system is implemented as a website, then the user terminal is implemented as a webpage; or if the multidimensional adaptive online coal ash content detection system is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.

[0075] like Figure 1 The figure shown is a system architecture diagram of a multi-dimensional adaptive online coal ash content detection system provided in an embodiment of the present invention.

[0076] The multidimensional adaptive online coal ash content detection system 100 of this invention can be set up in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed into a website. Depending on the functions implemented, the multidimensional adaptive online coal ash content detection system 100 may include a signal processing module 101, a signal feature reconstruction module 102, a signal feature optimization module 103, a coal ash content judgment module 104, and an ash content verification module 105. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device's processor and can perform a fixed function, stored in the electronic device's memory.

[0077] In this embodiment of the invention, in the multidimensional adaptive online coal ash content detection system, each of the above modules can be implemented independently and can call other modules. Here, "calling" can be understood as one module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the multidimensional adaptive online coal ash content detection system provided by this embodiment of the invention, the applicable scope of the multidimensional adaptive online coal ash content detection system architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the multidimensional adaptive online coal ash content detection system. In practical applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0078] The following describes the components and workflow of the multidimensional adaptive online coal ash content detection system, using specific embodiments as examples:

[0079] The signal processing module 101 is used to perform noise suppression processing on the radiation signal of the coal to obtain the preprocessed signal of the coal.

[0080] In this embodiment of the invention, when the signal processing module 101 performs noise suppression processing on the radiation signal of the coal, it is specifically used for:

[0081] The radiation signal is subjected to anti-interference processing to obtain the radiation signal anti-interference signal;

[0082] The anti-interference signal is smoothed to obtain a smoothed signal of the anti-interference signal;

[0083] The smoothed signal is subjected to noise reduction processing to obtain the preprocessed signal of the coal.

[0084] Specifically, a metal shield is used to completely enclose the radiation signal acquisition component, blocking electromagnetic signals from the external environment from entering the acquisition area. At the same time, a conductive shielding layer is wrapped around the signal transmission line to filter out noise mixed in during transmission. Through these two shielding measures, external interference components in the radiation signal are removed, resulting in an anti-interference radiation signal.

[0085] Furthermore, a series of consecutive anti-interference signal sampling points are selected as a group of processing units. Starting from the first sampling point, the sampling point and a fixed number of adjacent sampling points are extracted together. The average value of the signal values ​​of these sampling points is calculated, and the average value is used to replace the value of the original sampling point. All anti-interference signal sampling points are processed in this way to obtain a smoothed anti-interference signal.

[0086] Furthermore, a fixed noise threshold range is set according to the normal value range of the coal radiation signal. Each signal value of the smoothed signal is read one by one, and it is determined whether the value is within the set noise threshold range. If it is, the signal value is directly removed; if it is not, the signal value is retained. After the judgment and screening of all smoothed signal values ​​are completed, the preprocessed signal of coal is obtained.

[0087] In summary, the signal processing module 101 employs a layered noise suppression process of "interference suppression processing - smoothing processing - noise cancellation processing" to specifically address the problems of existing technologies that can only simply filter interference and still contain a significant amount of noise in the pre-processed signal. First, it wraps the acquisition components with a metal shield and the transmission lines with a conductive shielding layer to effectively isolate external electromagnetic interference and transmission noise. Then, it calculates the average value of the sampling points using a sliding window to smooth the signal and reduce signal fluctuations. Finally, it filters and removes noise values ​​according to a threshold.

[0088] In summary, this processing significantly improves the purity of the preprocessed signal, providing high-quality data support for subsequent signal feature reconstruction, accurate extraction and optimization of feature signals, and accurate calculation of coal ash content. At the same time, it reduces the impact of complex interference on the detection signal, ensures stable signal quality, and avoids deviations in subsequent detection steps due to signal interference, thus laying the foundation for improving the overall accuracy of online coal ash content detection.

[0089] The signal feature reconstruction module 102 is used to perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal.

[0090] In this embodiment of the invention, when the signal feature reconstruction module 102 performs adaptive feature reconstruction on the preprocessed signal, it is specifically used for:

[0091] The preprocessed signal is evaluated for signal quality to obtain the signal quality index of the preprocessed signal.

[0092] Based on the signal quality index, the preprocessed signal is decomposed into multi-scale features to obtain the feature components of the preprocessed signal.

[0093] Adaptive filtering is performed on the feature components to obtain the effective feature components of the preprocessed signal;

[0094] The effective feature components are weighted, fused, and reconstructed to obtain the optimized feature signal.

[0095] Specifically, the standard amplitude range of the preprocessed signal is selected as a reference. The actual amplitude of the preprocessed signal is read segment by segment. The actual amplitude of each segment is compared with the standard amplitude range. The number of amplitudes exceeding the standard range in each segment is counted. The proportion of the number of exceeding the standard range to the total number of amplitudes in that segment is used as the evaluation basis to obtain the signal quality index of the preprocessed signal.

[0096] Furthermore, three fixed-size sliding windows of small, medium and large are prepared. Each window is moved sequentially over the preprocessed signal. After each movement, a signal segment within the window is captured. After each scale window completes all movements, a set of signal segments of that scale is formed. The sets of signal segments of all scales together constitute the feature components of the preprocessed signal.

[0097] Furthermore, based on the signal quality index, a threshold for retaining feature components is set. Each feature component's signal segment is read one by one, and it is determined whether the percentage of amplitudes in the segment that meet the standard amplitude range reaches the threshold. If the threshold is reached, the feature component is retained; otherwise, it is discarded. The final retained feature components are the effective feature components of the preprocessed signal.

[0098] Furthermore, weights are assigned according to the proportion of effective feature component signal segments that conform to the standard amplitude. The higher the proportion, the greater the weight. Each effective feature component is multiplied by its weight, and then all the multiplication results are superimposed. The superimposed signal is the optimized feature signal.

[0099] In this embodiment of the invention, when the signal feature reconstruction module 102 performs adaptive filtering of the feature components, it is specifically used for:

[0100] The stability of the signal quality index is analyzed to obtain the stability evaluation results of the signal quality index.

[0101] The stability evaluation results of the indicators are mapped using rules to obtain a parameter adjustment strategy for feature selection.

[0102] Based on the parameter adjustment strategy, the threshold and weight of the filtering rules are configured to obtain the filtering rules for the feature components;

[0103] The feature components are iteratively filtered based on the filtering rules to obtain the effective feature components.

[0104] Specifically, the signal quality index is divided into multiple continuous segments at fixed time intervals. The difference between the maximum and minimum values ​​of the index in each segment is calculated. The differences of all segments are compared. If all differences are within a preset small fluctuation range, the index is considered stable. If any difference exceeds this range, the index is considered unstable. This yields the stability evaluation result of the signal quality index.

[0105] Furthermore, parameter adjustment strategies are preset for two evaluation results: stable and unstable. When the indicator stability evaluation result is stable, it is mapped to a parameter adjustment strategy of "fine-tuning the screening threshold and controlling the weight deviation within a small range". When the evaluation result is unstable, it is mapped to a parameter adjustment strategy of "moderately lowering the screening threshold and biasing the weight towards the component with small signal fluctuations". Thus, the parameter adjustment strategy for feature screening is obtained.

[0106] Furthermore, specific screening thresholds are set according to the threshold requirements in the parameter adjustment strategy, and the weight ratios of different feature components are allocated according to the weight requirements in the strategy to ensure that both the thresholds and weights strictly comply with the strategy, thereby obtaining the screening rules for feature components.

[0107] Furthermore, all feature components are initially screened using screening rules, retaining those that meet the threshold and have the required weight. Then, the same rules are used to screen the initially retained components a second time, removing those that do not meet the requirements in the second screening. This process is repeated until the results of two consecutive screenings are consistent, and finally, the effective feature components are obtained.

[0108] In summary, the signal feature reconstruction module 102 addresses the shortcomings of existing technologies, such as lack of adaptive adjustment and insufficient targeting of feature extraction, through an adaptive processing flow of "signal quality assessment - multi-scale feature decomposition - adaptive screening - weighted fusion reconstruction". First, it assesses signal quality based on the proportion of deviation between the preprocessed signal amplitude and the standard range, setting the direction for subsequent processing. Then, it uses a multi-scale sliding window to extract signal segments, comprehensively capturing features from different dimensions and avoiding the loss of key information at a single scale. Next, it dynamically adjusts the screening rules based on signal quality stability, accurately retaining effective feature components that meet the threshold. Finally, it fuses the effective components according to their quality-assigned weights, obtaining an optimized feature signal that meets the requirements of grayscale detection.

[0109] In summary, this module enhances the representativeness and effectiveness of optimized feature signals, providing a reliable foundation for the subsequent signal feature optimization module 103 to establish environmental coupling relationships and perform real-time compensation, thus reducing environmental compensation bias. Simultaneously, it provides high-quality feature data for the coal ash content judgment module 104, aiding in accurate ash content calculation. Furthermore, its adaptive characteristics can adapt to preprocessed signals of varying quality, preventing feature failure due to signal quality differences, further ensuring the stability of the detection system, and providing crucial support for improving the overall accuracy and adaptability of online coal ash content detection.

[0110] The signal feature optimization module 103 is used to establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and to perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain an environmental adaptive signal of the optimized feature signal.

[0111] In this embodiment of the invention, when the signal feature optimization module 103 performs the dynamic coupling relationship between environmental parameters and the optimized feature signal, it is specifically used for:

[0112] Under standard operating conditions, the optimized feature signal is collected as a reference signal, and environmental parameters are collected as reference values.

[0113] Under varying operating conditions, the real-time values ​​of the optimized feature signal and the environmental parameters are collected. The real-time value of the optimized feature signal is compared with the reference signal to obtain the signal drift of the optimized feature signal.

[0114] The real-time values ​​of the environmental parameters are combined with the corresponding signal drift to obtain a training sample set;

[0115] The training sample set is input into a preset initial neural network, and the connection weights of the initial neural network are adjusted to obtain a nonlinear mapping model between the environmental parameters and the signal drift.

[0116] When the amount of incremental sample data reaches the preset batch size, or when the statistical variance of the signal drift exceeds the stability threshold, the connection weights of the nonlinear mapping model are updated to obtain the dynamically updated model.

[0117] Specifically, the environmental conditions are fixed as standard operating conditions with constant temperature, constant humidity and no external electromagnetic interference. Signal acquisition equipment is used to continuously collect and record optimized characteristic signals. The obtained signals are the reference signals. At the same time, environmental sensors are used to collect and record environmental parameters such as temperature and humidity. The obtained parameters are the reference values.

[0118] Furthermore, the environmental conditions in the standard operating condition are changed to form a variable operating condition. Under the variable operating condition, the optimized feature signal is collected and recorded as a real-time value using the same equipment as the reference signal. The environmental parameters are collected and recorded as real-time values ​​using the same sensor as the reference value. The real-time value of each optimized feature signal is subtracted from the signal value of the reference signal at the corresponding time. The difference obtained is the signal drift of the optimized feature signal.

[0119] Furthermore, each set of corresponding real-time environmental parameter values ​​is paired with the signal drift. For example, the real-time temperature value, real-time humidity value, and signal drift at a certain moment are combined to form a sample. Multiple such samples are aggregated to obtain the training sample set.

[0120] Furthermore, the input layer of the initial neural network is determined to be the environmental parameters and the output layer to be the signal drift. The environmental parameters in the training sample set are input into the input layer, and the predicted drift output by the output layer is compared with the actual signal drift in the sample. The connection weights between the layers of the neural network are adjusted according to the difference between the two. The adjustment is repeated until the difference is stable, and a nonlinear mapping model between the environmental parameters and the signal drift is obtained.

[0121] Furthermore, the batch size of incremental samples and the stability threshold of signal drift are pre-set. New real-time values ​​of environmental parameters and corresponding signal drift are continuously collected as incremental samples. When the number of incremental samples reaches the batch size, or when the statistical variance of signal drift over a period of time (first calculate the average of these drifts, then calculate the sum of squares of the differences between each drift and the average, and divide by the total number of drifts) exceeds the stability threshold, the connection weights of the nonlinear mapping model are readjusted using incremental samples to obtain the dynamically updated model.

[0122] In this embodiment of the invention, when the signal feature optimization module 103 performs real-time compensation of the optimized feature signal based on the dynamic coupling relationship, it is specifically used for:

[0123] Obtain the current environmental parameters and input them into the nonlinear mapping model to obtain the compensation amount of the optimized feature signal, wherein the compensation amount is calculated using the following formula:

[0124]

[0125] The compensation amount is applied to the optimized feature signal to obtain the environment adaptive signal, and the calculation formula for the environment adaptive signal is as follows:

[0126] ;

[0127] In the formula, Indicates the current belt speed. Indicates the standard belt speed; Indicates the current coal thickness. Indicates the thickness of standard coal; Indicates the current ambient humidity. Indicates standard ambient humidity; , , These represent the influence coefficients of belt speed, coal thickness, and ambient humidity, respectively, which are output in real time by the nonlinear mapping model based on the current environmental parameters. This represents the optimized feature signal; This refers to the environmental adaptive signal.

[0128] Specifically, the current belt speed Under varying operating conditions, the standard belt speed is used and collected. The same equipment collects real-time environmental parameters including belt speed data; standard belt speed. These are belt speed parameters recorded when environmental parameters are collected as reference values ​​under standard operating conditions of constant temperature, constant humidity, and no external electromagnetic interference.

[0129] Current coal thickness Under varying operating conditions, this is achieved by comparing the thickness of the standard coal sample. The same detection equipment collects environmental parameters in real time, including coal quantity and thickness data; standard coal quantity and thickness. It is the coal quantity and thickness parameter recorded when environmental parameters are collected as reference values ​​under standard operating conditions.

[0130] Current ambient humidity Under varying operating conditions, the standard ambient humidity is utilized and collected. Humidity data from environmental parameters collected in real time by the same environmental sensor; standard ambient humidity. It is the humidity parameter recorded when environmental parameters are collected as reference values ​​under standard operating conditions.

[0131] , , The training sample set is input into a preset initial neural network. After adjusting the connection weights of the initial neural network to obtain a nonlinear mapping model, the nonlinear mapping model outputs the influence coefficients of belt speed, coal thickness and environmental humidity in real time according to the input current environmental parameters.

[0132] Optimize feature signals It is the product obtained by the signal feature reconstruction module 102 after performing adaptive feature reconstruction on the preprocessed signal.

[0133] Compensation amount The calculation formula is to square the difference between the current belt speed and the standard belt speed, and then multiply it by the belt speed influence coefficient output by the nonlinear mapping model. Then, the ratio of the current coal thickness to the standard coal thickness is logarithmically calculated and multiplied by the coal thickness influence coefficient output by the nonlinear mapping model. Simultaneously, the difference between the current ambient humidity and the standard ambient humidity is cubed and then multiplied by the ambient humidity influence coefficient output by the nonlinear mapping model. Finally, the results of these three calculations are added together to comprehensively consider the differences between the current three environmental parameters and the standard parameters, as well as their respective degrees of influence, and to calculate the compensation amount that needs to be adjusted to optimize the characteristic signal. .

[0134] Environmental Adaptive Signals The calculation formula applies the compensation amount to the optimized feature signal through multiplication, ultimately obtaining an environmental adaptive signal that can adapt to the current environmental conditions. .

[0135] When the current belt speed Deviation from standard belt speed At any time, regardless Greater than or less than The square of the difference between the two will increase, multiplied by the belt speed influence coefficient. The resulting value increases accordingly, leading to a larger compensation amount. Increase; when the current belt speed Approximately standard belt speed When the square of the difference between the two decreases, multiplied by The result obtained later is reduced, and the compensation amount is reduced. It decreases accordingly.

[0136] When the current coal thickness Thickness greater than standard coal quantity When the ratio of the two is greater than 1, the logarithm of the ratio is positive, and it is multiplied by the coal quantity thickness influence coefficient. The result obtained was positive, and the compensation amount was... Consequently, the thickness of the current coal layer increases; Thickness less than standard coal quantity When the ratio is less than 1, the logarithm of the ratio is negative, and multiplied by... The result obtained was negative, and the compensation amount was... It decreases accordingly.

[0137] When the current ambient humidity Deviation from standard ambient humidity At any time, regardless Greater than or less than The cube of the difference between the two will increase, multiplied by the environmental humidity influence coefficient. The resulting value increases accordingly, leading to a larger compensation amount. Increase; when the current ambient humidity Approximately standard ambient humidity When the cube of the difference between the two decreases, multiplied by The result obtained later is reduced, and the compensation amount is reduced. It decreases accordingly.

[0138] When compensation amount When the signal increases, the environment adapts to the signal. Will follow It increases with the increase of the compensation amount; When the signal decreases, the environment adapts to the signal. Will follow It decreases as it decreases.

[0139] In summary, the signal feature optimization module 103, through a process of "establishing a dynamic coupling relationship - real-time compensation," can solve the problems of poor environmental adaptability and signal drift caused by environmental fluctuations without compensation in existing technologies. It first collects benchmark signals and environmental reference values ​​under standard operating conditions, then collects real-time values ​​under varying operating conditions to calculate signal drift. It uses samples to train a neural network to build a nonlinear mapping model and dynamically updates it. This allows for real-time correlation between environmental parameters such as belt speed, coal thickness, and humidity and signal drift, avoiding the shortcomings of static models that cannot adapt to environmental changes.

[0140] In summary, based on the coefficients output by the model, the compensation amount is accurately calculated using formulas and applied to optimize the feature signal to obtain an environmentally adaptive signal. This can offset the impact of environmental fluctuations on the signal and avoid signal drift leading to subsequent detection deviations. This module provides the coal ash content judgment module 104 with a high-quality signal adapted to actual working conditions, reducing ash content calculation errors. At the same time, the dynamically updated characteristics improve the system's adaptability to changing industrial scenarios, ensuring signal stability and reliability, and providing key support for improving overall detection accuracy and stability.

[0141] The coal ash content judgment module 104 is used to perform ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal to obtain the ash content value of the coal.

[0142] In this embodiment of the invention, when the coal ash content judgment module 104 performs ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for:

[0143] Based on the historical database, the ash content of the environmental adaptive signal is initially predicted to obtain the signal ash content value of the coal.

[0144] Obtain the coal quality parameters of the current batch of coal, perform a coupled analysis of moisture and calorific value on the coal quality parameters, and obtain the compensated ash content value of the coal.

[0145] The confidence levels of the signal gray value and the compensated gray value are evaluated to obtain the signal confidence level of the signal gray value and the parameter confidence level of the compensated gray value.

[0146] Based on the signal confidence level and the parameter confidence level, the signal ash value and the compensated ash value are weighted and fused to obtain the ash value of the coal, wherein the final ash value is calculated using the following formula:

[0147]

[0148] In the formula, The confidence level of the signal. The gray value of the signal. The confidence level of the parameter. The compensation ash value is given.

[0149] Specifically, the historical database stores environmental adaptive signals and corresponding measured ash values ​​of different batches of coal. The current environmental adaptive signal is compared with the environmental adaptive signals in the historical database one by one. The historical signal with the smallest difference from the current signal value is found, and the measured ash value corresponding to the historical signal is extracted to obtain the signal ash value of the coal.

[0150] Furthermore, by collecting the moisture content and calorific value data of the current batch of coal using coal quality testing equipment as coal quality parameters, the standard moisture content and standard calorific value of this type of coal are first determined. The difference between the current moisture content and the standard moisture content, and the difference between the current calorific value and the standard calorific value are calculated. Based on these two differences, the basic ash content of this type of coal (the average value calculated from the measured ash content of similar coals in the past) is adjusted. The larger the difference in moisture content and the larger the difference in calorific value, the greater the adjustment range of the basic ash content, thus obtaining the compensated ash content value of the coal.

[0151] Furthermore, the numerical overlap ratio between the current environmental adaptive signal and the matching signal in the historical database is calculated. This ratio is the signal confidence level of the ash content value. The higher the ratio, the higher the signal confidence level. The operating status of the detection equipment during the current coal quality parameter acquisition process is checked. If the equipment is normal and the deviation of the moisture and calorific value data collected three times in a row is less than a fixed range, the confidence level is high. Otherwise, the confidence level is reduced according to the degree of deviation to obtain the parameter confidence level of the compensated ash content value.

[0152] Furthermore, the signal confidence level and the parameter confidence level are added together to obtain the total confidence level. The signal confidence level is divided by the total confidence level to obtain the signal weight, and the parameter confidence level is divided by the total confidence level to obtain the parameter weight. The signal ash value is multiplied by the signal weight, and the compensated ash value is multiplied by the parameter weight. The two products are added together to obtain the ash value of the coal.

[0153] Specifically, signal confidence The confidence level is obtained when evaluating the signal gray value and the compensated gray value. The evaluation process is as follows: First, determine the matching degree standard between the environmental adaptive signal and the corresponding gray value in the historical database. Then, compare the currently obtained signal gray value with the matching degree standard and count the proportion of times the signal gray value meets the matching degree requirement out of the total number of evaluations. This proportion is the signal confidence level. .

[0154] Furthermore, the signal gray value This is based on a preliminary ash content prediction of the environmental adaptive signal using a historical database. Specifically, it involves extracting historical environmental adaptive signals with values ​​similar to the current environmental adaptive signal from the historical database, retrieving the corresponding ash content data from these historical signals, and taking the average of these ash content data as the signal ash content value of the current coal. .

[0155] Furthermore, parameter confidence The confidence level is obtained when assessing the confidence level of the signal ash content and the compensated ash content. The assessment process is as follows: An allowable error range is set for the coupled analysis results of coal quality parameters moisture and calorific value. The currently obtained compensated ash content value is compared with the allowable error range. The proportion of times the compensated ash content value falls within the allowable error range out of the total number of assessments is counted. This proportion is the parameter confidence level. .

[0156] Furthermore, compensate for ash value This is obtained by performing a coupled analysis of moisture and calorific value on the coal quality parameters of the current batch of coal. Specifically, the moisture content and calorific value data of the current batch of coal are first measured, and the moisture content and calorific value data are correlated in a fixed order. Then, the ash content data corresponding to the same correlation in historical data is searched, and the ash content data is taken as the compensation ash value. .

[0157] Furthermore, the significance of this formula lies in obtaining the ash content value of coal through weighted fusion. The specific calculation process is as follows: first, the signal confidence level is... With signal gray value Multiply to obtain the weighted gray result on the signal side; then calculate the parameter confidence level. Compensation ash value Multiply the two weighted gray values ​​to obtain the weighted gray value result on the parameter side; then add the two weighted gray value results to obtain the total weighted gray value; finally, divide the total weighted gray value by the signal confidence level. With parameter confidence The sum of the ash content and the ash content of the coal is the ash content value. By combining the reliability of the signal ash content value and the compensated ash content value, a more accurate ash content judgment result can be obtained.

[0158] Furthermore, when the signal confidence level Increase, and parameter confidence. and compensation ash value Signal gray content When both remain constant, the proportion of the weighted ash content of the signal side in the total weighted ash content increases, and the final coal ash content value will shift towards the signal ash content value. near.

[0159] Furthermore, when the parameter confidence level Increase, and signal confidence and compensation ash value Signal gray content When all parameters remain constant, the proportion of the weighted ash content in the total weighted ash content increases, and the final coal ash content value will shift towards the compensated ash content value. near.

[0160] Furthermore, when the signal gray value Increase, and signal confidence Parameter confidence Compensation ash value When all remain constant, the weighted ash content result on the signal side increases, the total weighted ash content increases accordingly, and the final coal ash content value also increases.

[0161] Furthermore, when compensating for ash value Increase, and signal confidence Parameter confidence Signal gray content When all parameters remain unchanged, the weighted ash content result on the parameter side increases, the total weighted ash content increases accordingly, and the final coal ash content value also increases.

[0162] In an embodiment of the invention, the coal ash content determination module 104, when performing a confidence assessment of the signal ash content value and the compensated ash content value, is specifically used for:

[0163] Variance calculation is performed on the fluctuation data of the environmental adaptive signal within a preset time window to obtain the signal fluctuation quantization value;

[0164] The signal fluctuation quantization value is compared with a preset stability threshold to obtain the signal reliability factor;

[0165] The ratio of moisture value to calorific value in the coal quality parameters is calculated to obtain the actual ratio of moisture to calorific value.

[0166] The difference between the actual ratio of moisture to calorific value and the parameter range of standard coal type is calculated to obtain the parameter deviation.

[0167] The parameter deviation is input into a preset rationality mapping table for query and matching to obtain the parameter rationality factor;

[0168] Retrieve historical consistency records of current coal source information from the database;

[0169] The signal reliability factor, the parameter rationality factor, and the historical detection consistency record are fused together to obtain the signal confidence level and the parameter confidence level.

[0170] Specifically, a fixed preset time window is set, and all environmental adaptive signal data within the window are continuously collected. First, the average value of these data is calculated, and then the difference between each data point and the average value is calculated and squared. All squared results are added together and divided by the total number of data points to obtain the signal fluctuation quantization value.

[0171] Furthermore, a fixed stability threshold is preset. If the signal fluctuation quantization value is less than the threshold, a preset corresponding table is queried to match a high signal reliability factor; if it is greater than or equal to the threshold, a low signal reliability factor is matched to obtain the signal reliability factor.

[0172] Furthermore, the moisture value and calorific value are extracted from the coal quality parameters of the current batch of coal, and the moisture value is directly divided by the calorific value to obtain the actual ratio of moisture to calorific value.

[0173] Furthermore, the range of moisture-to-calorific-value ratios for standard coal types is obtained, the median value of this range is calculated, and the median value is subtracted from the actual moisture-to-calorific-value ratio to obtain the parameter deviation.

[0174] Furthermore, a rationality mapping table is preset, and the corresponding value is found in the table according to the parameter deviation to obtain the parameter rationality factor.

[0175] Furthermore, historical testing data of the current coal source can be retrieved from the database. For example, the percentage of times the ash content deviation was less than 3% in the past 8 test results can be used as a record of historical testing consistency.

[0176] Furthermore, the signal reliability factor, parameter rationality factor, and historical record are set to account for 40%, 30%, and 30% of the weights, respectively. After multiplying each of the three by their respective weights, they are added together. A total of 0.7 or above is a confidence level of 0.9, 0.4-0.7 is a confidence level of 0.6, and below is a confidence level of 0.3. This yields the signal confidence level and the parameter confidence level, respectively.

[0177] In summary, the coal ash content judgment module 104 solves the problem of low accuracy in existing technologies that rely solely on signals or parameters, through a judgment logic of "dual-source data fusion + confidence weighting". First, based on a historical database, it obtains signal ash content values ​​by matching environmental adaptive signals (data optimized by previous modules), ensuring reference reliability through historical data. Then, combined with the current batch of coal quality parameters, it obtains compensated ash content values ​​through moisture and calorific value coupling analysis, closely matching real-time coal quality characteristics. Furthermore, it performs confidence assessments on the two types of ash content values ​​through signal fluctuation quantification, parameter deviation analysis, and historical consistency records, ensuring reasonable weight allocation, and finally, weighted fusion yields the final ash content value.

[0178] In summary, this module fully leverages the synergistic advantages of optimized signals and real-time parameters, avoiding the impact of single data deviations on results and significantly improving the accuracy of ash content values. At the same time, it provides a reliable initial ash content value for the subsequent ash content verification module 105, reduces the probability of misjudging anomalies, further ensures the reliability of the detection system's output results, and provides core support for meeting the high standards of industrial production in terms of overall online coal ash content detection accuracy.

[0179] The ash content verification module 105 is used to verify the credibility of the ash content value and obtain the ash content test result of the coal.

[0180] In this embodiment of the invention, when the ash content verification module 105 performs an ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for:

[0181] The ash content value is input into a preset reasonable ash content range for range comparison;

[0182] When the ash content value is within the reasonable range, the ash content value is output as the ash content detection result;

[0183] When the ash content value exceeds the reasonable range, an anomaly flag is output as the ash content detection result.

[0184] Specifically, the preset reasonable ash content range is a normal ash content value range determined by statistical analysis of a large amount of historical ash content test data of different coal types. The ash content value obtained by the coal quality ash content judgment module 104 is compared with the upper and lower limits of the reasonable range to confirm whether the ash content value falls between the upper and lower limits.

[0185] Furthermore, when the comparison confirms that the gray value is greater than or equal to the lower limit of the reasonable range and less than or equal to the upper limit of the reasonable range, that is, the gray value is within the reasonable range, the gray value is directly output. The gray value output at this time is the gray detection result.

[0186] Furthermore, when the comparison reveals that the gray value is less than the lower limit of the reasonable range or greater than the upper limit of the reasonable range, i.e., the gray value exceeds the reasonable range, a preset specific abnormality label (such as a prompt character in a fixed format) is generated and output. At this time, the output abnormality label is the gray detection result.

[0187] In summary, the ash content verification module 105, through its core operation of "comparing the ash content value with a preset reasonable range," can specifically address the problems of existing technologies lacking reliable verification of ash content results and easily outputting abnormal detection values. Its preset reasonable ash content range is generated based on statistical analysis of historical ash content detection data from a large number of different coal types, ensuring that the verification basis is scientifically reliable. By comparing the ash content value output by the coal ash content judgment module 104 with the reasonable range, it can accurately identify abnormal values—outputting valid results when the ash content value is within the range, and outputting an abnormality flag when it exceeds the range, directly filtering out unreliable data.

[0188] In summary, this module, acting as the "final checkpoint" in the testing process, prevents abnormal ash content values ​​from being used as valid results in industrial production, ensuring the reliability of test results. Simultaneously, anomaly identification can indicate potential deviations in preceding stages (such as signal processing and feature optimization), providing direction for troubleshooting. Furthermore, its verification logic is simple and efficient, requiring no additional complex calculations while significantly improving the overall effectiveness of the test results. This meets the high standards of industrial production for coal ash content testing—"high reliability and no false positives"—providing credible data support for subsequent production decisions.

[0189] Reference Figure 2 The diagram shown is a schematic flowchart of a multidimensional adaptive online coal ash content detection method provided in an embodiment of the present invention. In this embodiment, the multidimensional adaptive online coal ash content detection method includes:

[0190] S1. Perform noise suppression processing on the radiation signal of the coal to obtain the preprocessed signal of the coal;

[0191] S2. Perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal;

[0192] S3. Establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain the environmental adaptive signal of the optimized feature signal;

[0193] S4. Based on the coal quality parameters and the environmental adaptive signal, perform ash content fusion judgment on the coal to obtain the ash content value of the coal.

[0194] S5. Verify the reliability of the ash content value to obtain the ash content test result of the coal.

[0195] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0196] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0197] 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 technical solutions of the present invention.

Claims

1. A multidimensional adaptive online coal ash content detection system, characterized in that, The system includes a signal processing module, a signal feature reconstruction module, a signal feature optimization module, a coal ash content determination module, and an ash content verification module, wherein: The signal processing module is used to perform noise suppression processing on the radiation signal of coal to obtain the preprocessed signal of coal. The signal feature reconstruction module is used to perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal; The signal feature optimization module is used to establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and to perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain an environmental adaptive signal of the optimized feature signal. The coal ash content judgment module is used to perform ash content fusion judgment on coal based on coal quality parameters and the environmental adaptive signal to obtain the ash content value of coal. The ash content verification module is used to verify the reliability of the ash content value and obtain the ash content test results of the coal.

2. The multidimensional adaptive online coal ash content detection system as described in claim 1, characterized in that, When performing noise suppression processing on the radiation signal of the coal, the signal processing module is specifically used for: The radiation signal is subjected to anti-interference processing to obtain the radiation signal anti-interference signal; The anti-interference signal is smoothed to obtain a smoothed signal of the anti-interference signal; The smoothed signal is subjected to noise reduction processing to obtain the preprocessed signal of the coal.

3. The multidimensional adaptive online coal ash content detection system as described in claim 1, characterized in that, When performing adaptive feature reconstruction on the preprocessed signal, the signal feature reconstruction module is specifically used for: The preprocessed signal is evaluated for signal quality to obtain the signal quality index of the preprocessed signal. Based on the signal quality index, the preprocessed signal is decomposed into multi-scale features to obtain the feature components of the preprocessed signal. The feature components are adaptively filtered to obtain the effective feature components of the preprocessed signal; The effective feature components are weighted, fused, and reconstructed to obtain the optimized feature signal of the preprocessed signal.

4. The multidimensional adaptive online coal ash content detection system as described in claim 3, characterized in that, When performing adaptive filtering of the feature components, the signal feature reconstruction module is specifically used for: The stability of the signal quality index is analyzed to obtain the stability evaluation results of the signal quality index. The stability evaluation results of the indicators are mapped using rules to obtain a parameter adjustment strategy for feature selection. Based on the parameter adjustment strategy, the threshold and weight of the filtering rules are configured to obtain the filtering rules for the feature components; The feature components are iteratively filtered based on the filtering rules to obtain the effective feature components.

5. The multidimensional adaptive online coal ash content detection system as described in claim 1, characterized in that, When the signal feature optimization module establishes a dynamic coupling relationship between environmental parameters and the optimized feature signal, it is specifically used for: Under standard operating conditions, the optimized feature signal is collected as a reference signal, and environmental parameters are collected as reference values. Under varying operating conditions, the real-time values ​​of the optimized feature signal and the environmental parameters are collected. The real-time value of the optimized feature signal is compared with the reference signal to obtain the signal drift of the optimized feature signal. The real-time values ​​of the environmental parameters are combined with the corresponding signal drift to obtain a training sample set; The training sample set is input into a preset initial neural network, and the connection weights of the initial neural network are adjusted to obtain a nonlinear mapping model between the environmental parameters and the signal drift. When the amount of incremental sample data reaches the preset batch size, or when the statistical variance of the signal drift exceeds the stability threshold, the connection weights of the nonlinear mapping model are updated to obtain the dynamically updated model.

6. The multidimensional adaptive online coal ash content detection system as described in claim 5, characterized in that, When the signal feature optimization module performs real-time compensation of the optimized feature signal based on the dynamic coupling relationship, it is specifically used for: Obtain the current environmental parameters and input them into the nonlinear mapping model to obtain the compensation amount of the optimized feature signal, wherein the compensation amount is calculated using the following formula: The compensation amount is applied to the optimized feature signal to obtain the environment adaptive signal, and the calculation formula for the environment adaptive signal is as follows: ； In the formula, Indicates the current belt speed. Indicates the standard belt speed. Indicates the current coal thickness. Indicates the thickness of the standard coal quantity. Indicates the current ambient humidity. Indicates standard ambient humidity. , , These represent the influence coefficients of belt speed, coal thickness, and ambient humidity, respectively, output in real time by the nonlinear mapping model based on current environmental parameters. This represents the optimized feature signal. This refers to the environmental adaptive signal.

7. The multidimensional adaptive online coal ash content detection system as described in claim 1, characterized in that, When the coal ash content determination module performs an ash content fusion determination on the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for: Based on the historical database, the ash content of the environmental adaptive signal is initially predicted to obtain the signal ash content value of the coal. Obtain the coal quality parameters of the current batch of coal, perform a coupled analysis of moisture and calorific value on the coal quality parameters, and obtain the compensated ash content value of the coal. The confidence levels of the signal gray value and the compensated gray value are evaluated to obtain the signal confidence level of the signal gray value and the parameter confidence level of the compensated gray value. Based on the signal confidence level and the parameter confidence level, the signal ash value and the compensated ash value are weighted and fused to obtain the ash value of the coal, wherein the final ash value is calculated using the following formula: In the formula, The confidence level of the signal. The gray value of the signal. The confidence level of the parameter. The compensation ash value is given.

8. The multidimensional adaptive online coal ash content detection system as described in claim 7, characterized in that, When performing a confidence assessment of the signal ash content value and the compensated ash content value, the coal ash content determination module is specifically used for: Variance calculation is performed on the fluctuation data of the environmental adaptive signal within a preset time window to obtain the signal fluctuation quantization value; The signal fluctuation quantization value is compared with a preset stability threshold to obtain the signal reliability factor; The ratio of moisture value to calorific value in the coal quality parameters is calculated to obtain the actual ratio of moisture to calorific value. The difference between the actual ratio of moisture to calorific value and the parameter range of standard coal type is calculated to obtain the parameter deviation. The parameter deviation is input into a preset rationality mapping table for query and matching to obtain the parameter rationality factor; Retrieve historical consistency records of current coal source information from the database; The signal reliability factor, the parameter rationality factor, and the historical detection consistency record are fused together to obtain the signal confidence level and the parameter confidence level.

9. The multidimensional adaptive online coal ash content detection system as described in claim 1, characterized in that, When the ash content verification module performs an ash content fusion judgment on the coal based on coal quality parameters and the environmental adaptive signal, it is specifically used for: The ash content value is input into a preset reasonable ash content range for range comparison; When the ash content value is within the reasonable range, the ash content value is output as the ash content detection result; When the ash content value exceeds the reasonable range, an anomaly flag is output as the ash content detection result.

10. A multidimensional adaptive online detection method for coal ash content, characterized in that, The method includes: S1. Perform noise suppression processing on the radiation signal of the coal to obtain the preprocessed signal of the coal; S2. Perform adaptive feature reconstruction on the preprocessed signal to obtain the optimized feature signal of the preprocessed signal; S3. Establish a dynamic coupling relationship between environmental parameters and the optimized feature signal, and perform real-time compensation on the optimized feature signal based on the dynamic coupling relationship to obtain the environmental adaptive signal of the optimized feature signal; S4. Based on the coal quality parameters and the environmental adaptive signal, perform ash content fusion judgment on the coal to obtain the ash content value of the coal. S5. Verify the reliability of the ash content value to obtain the ash content test result of the coal.

Images