Coal sample assay fraud behavior analysis method and system

Abstract

The invention relates to the technical field of coal sample detection, in particular to a coal sample assay fraud behavior analysis method and system, and the method comprises the steps: building a plurality of working areas according to a coal sample assay process, and constructing a primary monitoring model of each working area; presetting a plurality of characteristic indexes, obtaining characteristic index parameters in each working area according to a preset adjustment time node, and judging whether a correction instruction is generated or not; acquiring monitoring data in each working area, and generating a risk evaluation value of each working area according to all the monitoring data; judging whether an early warning instruction is generated or not according to all the risk evaluation values; a plurality of working areas are constructed according to a coal sample assay process, corresponding monitoring points are set according to equipment parameters and working contents in different working areas, image monitoring data in the working areas are collected in real time, whether illegal behaviors exist or not is identified according to a corresponding primary monitoring model, and early warning is performed on fraud behaviors in time. The authenticity of a coal sample test result is ensured.

Description

Technical Field

[0001] The present application relates to the technical field of coal sample detection, and in particular to a method and system for analyzing fraudulent behavior in coal sample testing. Background Art

[0002] Coal is the main raw material of thermal power plants, and the cost of coal accounts for more than 70% of the production cost of power generation enterprises. Coal testing at the factory is an important part of coal quality acceptance, the main basis for checking whether the incoming coal meets the contract requirements, and an important basis for the reasonable storage and economical burning of incoming coal.

[0003] At present, the method generally used by power plants is equipment testing and manual reading. This method has the following drawbacks: there is human fraud, the testing process is unsupervised, it does not follow the relevant national standards, and there is a phenomenon of simplifying the testing process due to laziness, which affects the fairness and justice of the test results. Therefore, the existing laboratory testing methods can no longer meet the requirements of power plants for coal quality management in the current market environment. Summary of the invention

[0004] The purpose of this application is: to solve the above-mentioned technical problems, this application provides a coal sample testing fraud analysis method and system, aiming to achieve real-time monitoring of the coal sample testing process, timely warn of illegal and fraudulent behavior, and improve the accuracy of coal sample testing.

[0005] In some embodiments of the present application, multiple working areas are constructed according to the coal sampling test process, and corresponding monitoring points are set according to the equipment parameters and work content in different working areas. Image monitoring data in the working areas are collected in real time, and whether there are any violations is identified based on the corresponding first-level monitoring model. Fraudulent behavior is warned in a timely manner to ensure the authenticity of the coal sampling test results.

[0006] In some embodiments of the present application, multiple characteristic indicators are constructed based on factors such as the supplier of coal samples and process operators, and whether to modify the primary monitoring model is determined based on the real-time parameters of each characteristic indicator, thereby improving the efficiency of judging violations and ensuring timely warning of fraudulent behavior.

[0007] In some embodiments of the present application, a method for analyzing fraudulent behavior in coal sample testing is provided, comprising: Establish multiple working areas according to the coal sample testing process, and construct a primary monitoring model for each working area; Preset multiple characteristic indicators, obtain characteristic indicator parameters in each working area according to the preset adjustment time node, and determine whether to generate a correction instruction; Obtain monitoring data in each work area and generate risk assessment values ​​for each work area based on all monitoring data; Judge whether to generate a warning instruction according to all risk evaluation values; Among them, when establishing multiple working areas, it includes: Establish a working area sequence A, A=(a1, a2…ai…an), where ai is the i-th working area; n is the number of working areas.

[0008] In some embodiments of the present application, when constructing the first-level monitoring model of each working area, it includes: Set ai as the target working area in sequence according to the working area sequence A; Obtain the working process parameters and historical parameters of the target working area; Set various violations in the target working area according to the historical parameters, and establish a training data packet according to all violations; Iteratively optimize the training data packet to generate a behavior recognition model for the target working area; Set the data collection frequency of the target working area according to the working process data, and set multiple monitoring time nodes according to the data collection frequency; Generate the first-level monitoring model of the target working area according to the data collection frequency and the behavior recognition model; Generate the first-level monitoring models of each working area in sequence; Establish a first-level monitoring model sequence B, B=(b1, b2…bi…bn), where bi is the first-level monitoring model of the i-th working area.

[0009] In some embodiments of the present application, when judging whether to generate a correction instruction, it includes: Obtain the characteristic index parameters of the target working area at the current adjustment time node, and generate a correction evaluation value f; f= (ηi*ki); Among them, θ is the number of characteristic indexes; ηi is the influence factor of the i-th characteristic index; ki is the reference value of the i-th characteristic index of the target working area at the current adjustment time node; Preset the first correction evaluation value threshold F1; If f<F1, at the current adjustment time node, the target working area does not generate a correction instruction, and set the compensation coefficient r = 1; If f≧F1, at the current adjustment time node, the target working area generates a first-level correction instruction, and set the compensation coefficient r according to the correction evaluation value f; Judge in sequence whether each working area generates a correction instruction at the current adjustment time node.

[0010] In some embodiments of the present application, when setting the compensation coefficient r according to the correction evaluation value f, it includes: Preset a first corrected evaluation value interval (F1, F2), a second corrected evaluation value interval (F2, F3 and a third corrected evaluation value interval (F3, F4); If f is in the preset first compensation evaluation value interval, the compensation coefficient r is set to the preset first compensation coefficient r1, that is, r=r1; If f is in the preset second compensation evaluation value interval, the compensation coefficient r is set to the preset second compensation coefficient r2, ie r=r2; If f is in the preset first compensation evaluation value interval, the compensation coefficient r is set to the preset third compensation coefficient r3, that is, r=r3; and 1 <r1<r2<r3。

[0011] In some embodiments of the present application, when generating the risk assessment value of each work area according to all monitoring data, it includes: According to the working area sequence A, set ai as the working area to be monitored in sequence; Obtain monitoring data of the work area to be monitored at the current monitoring time node; Preprocess the monitoring data according to the behavior recognition model of the work area to be monitored; Generate potential risk values ​​for each violation based on the preprocessing results; Establish a potential risk value sequence C, C = (c1, c2…ci…cm), where ci is the potential risk value of the i-th violation in the work area to be monitored at the current monitoring time node; m is the number of violation types in the work area to be monitored; Generate the risk assessment value g of the work area to be monitored at the current monitoring time node according to the potential risk value sequence C; Update the risk assessment value g according to the monitoring time nodes of the work area to be monitored.

[0012] In some embodiments of the present application, when generating the risk assessment value g of the work area to be monitored at the current monitoring time node, it includes: Get the compensation coefficient r' of the monitoring time node of the working area to be monitored; H=e1*Q1*[ βi*ci]+e2*Q2*[ βi*Y(i)(ci-c')]; g=r'*H; Among them, e1 is the preset first weight coefficient; e2 is the preset second weight coefficient; Q1 is the preset first fixed coefficient; Q2 is the preset second fixed coefficient; βi is the influencing factor of the i-th violation in the working area to be monitored; c' is the preset potential risk value threshold; Y(i) is the selection coefficient; if (ci-c')<0, Y(i)=0; if (ci-c')<0, Y(i)=1.

[0013] In some embodiments of the present application, when determining whether to generate a warning instruction, it includes: Preset a first risk evaluation value threshold G1 and a second risk evaluation value threshold G2; If g < G1, no warning instruction is generated for the work area to be monitored at the current monitoring time node; If G1 < g < G2, a first-level warning instruction is generated for the work area to be monitored at the current monitoring time node; If g > G2, a second-level warning instruction is generated for the work area to be monitored at the current monitoring time node.

[0014] In some embodiments of the present application, a coal sample testing fraud behavior analysis system is provided, including: A central control unit for establishing multiple work areas according to the coal sample testing process and constructing a first-level monitoring model for each work area; A correction unit for presetting multiple characteristic indicators, obtaining the characteristic indicator parameters in each work area according to the preset adjustment time node, and determining whether to generate a correction instruction; A monitoring unit for collecting monitoring data in each work area; The central control unit includes: A first processing module for generating a risk evaluation value for each work area according to all the monitoring data; A first warning module for determining whether to generate a warning instruction according to all the risk evaluation values; A second processing module for establishing a work area sequence A, A = (a1, a2... ai... an), where ai is the i-th work area; n is the number of work areas; A third processing module for constructing a first-level monitoring model for each work area.

[0015] In some embodiments of the present application, the third processing module is further used for: Sequentially setting ai as the target work area according to the work area sequence A; Obtaining the work process parameters and historical parameters of the target work area; Setting various violation behaviors in the target work area according to the historical parameters, and establishing a training data packet according to all the violation behaviors; Iteratively optimizing the training data packet to generate a behavior recognition model for the target work area; Setting the data collection frequency of the target work area according to the work process data, and setting multiple monitoring time nodes according to the data collection frequency; Generating a first-level monitoring model for the target work area according to the data collection frequency and the behavior recognition model; Sequentially generating first-level monitoring models for each work area; Establish a first-level monitoring model sequence B, B = (b1, b2…bi…bn), where bi is the first-level monitoring model of the i-th working area.

[0016] In some embodiments of the present application, the correction unit is further configured to: Obtain the characteristic index parameters of the target working area at the current adjustment time node, and generate a correction evaluation value f; f = (ηi * ki); Where θ is the number of characteristic indexes; ηi is the influence factor of the i-th characteristic index; ki is the reference value of the i-th characteristic index of the target working area at the current adjustment time node; Preset the first correction evaluation value threshold F1; If f < F1, at the current adjustment time node, the target working area does not generate a correction instruction, and set the compensation coefficient r = 1; If f ≥ F1, at the current adjustment time node, the target working area generates a first-level correction instruction, and set the compensation coefficient r according to the correction evaluation value f; Sequentially determine whether each working area generates a correction instruction at the current adjustment time node.

[0017] Compared with the prior art, the beneficial effects of a coal sample testing fraud behavior analysis method and system in an embodiment of the present application are as follows: Construct multiple working areas according to the coal sample testing process, set corresponding monitoring points according to the equipment parameters and work content in different working areas, collect image monitoring data in the working areas in real time, and identify whether there are violations according to the corresponding first-level monitoring model, and give early warnings of fraud behaviors in a timely manner to ensure the authenticity of coal sample testing results.

[0018] Construct multiple characteristic indexes based on factors such as the supplier of coal samples and process operators, and judge whether to correct the first-level monitoring model according to the real-time parameters of each characteristic index, so as to improve the judgment efficiency of violations and ensure timely early warnings of fraud behaviors. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of a coal sample testing fraud behavior analysis method in a preferred embodiment of an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following combines the drawings and embodiments to further describe the specific implementation manners of the present application in detail. The following embodiments are used to illustrate the present application, but are not used to limit the scope of the present application.

[0021] In the description of the present application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0022] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0023] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0024] like Figure 1 As shown, a coal sample testing fraud analysis method according to a preferred embodiment of the present application is characterized by comprising: S101: Establish multiple working areas according to the coal sample testing process, and construct a primary monitoring model for each working area; S102: Preset multiple characteristic indicators, obtain characteristic indicator parameters in each working area according to the preset adjustment time node, and determine whether to generate a correction instruction; S103: Acquire monitoring data in each work area, and generate a risk assessment value for each work area based on all monitoring data; S104: Determine whether to generate a warning instruction based on all risk assessment values; Among them, when establishing multiple work areas, it includes: Establish a working area sequence A, A=(a1,a2…ai…an), where ai is the i-th working area and n is the number of working areas.

[0025] Specifically, according to the workflow of coal sample analysis, multiple sub-steps are generated and multiple working areas are constructed according to each sub-step, and the working areas include but are not limited to: weighing area, sampling area, sample preparation area, iron removal area and other working areas.

[0026] Specifically, when constructing the first-level monitoring model for each work area, it includes: According to the working area sequence A, set ai as the target working area in sequence; Obtain workflow parameters and historical parameters of the target work area; Set multiple violations in the target work area based on historical parameters, and create a training data package based on all violations; Iteratively optimize the training data package to generate a behavior recognition model for the target work area; Set the data collection frequency of the target work area according to the workflow data, and set multiple monitoring time nodes according to the data collection frequency; Generate a primary monitoring model for the target work area based on the data collection frequency and behavior recognition model; Generate the first-level monitoring model of each working area in turn; Establish a first-level monitoring model series B, B=(b1, b2…bi…bn), where bi is the first-level monitoring model of the i-th working area.

[0027] Specifically, according to the different work contents in each work area, the possible violations are grouped. For example, the violations in the measurement and weighing area include but are not limited to the area: there is an angular difference in the truck scale, and weighing in the wrong position will lead to measurement deviation. It is hoped that video monitoring can be used to determine whether the parking position meets the standards; when weighing heavy and light vehicles, the number of people on the vehicle is inconsistent; after the vehicle is weighed, during the measurement process of the truck scale system, people approach, weigh, or intervene in the truck scale instrument. In the sampling area, detect and identify normal bagging and bag explosion (sample bag damage), and warn of bag explosion; when the sampler enters and exits the sampling room, detect whether the sample maker is carrying a sample bag; ensure that the sample bag is in the video screen throughout the process, and check whether the sealing action is completed under the camera to ensure that the entire sample collection process is completed under monitoring.

[0028] Specifically, by traversing the historical parameters of each work area and combining it with expert analysis data, corresponding violation behaviors are generated, and corresponding identification standards are formulated to set the corresponding behavior recognition model.

[0029] Specifically, the corresponding monitoring data collection frequency is set according to the historical fraud data of each work area. The higher the data collection frequency, the greater the possibility of fraud in the current monitoring area, and the smaller the time interval between corresponding adjacent monitoring time nodes.

[0030] It is understandable that in the above embodiments, multiple working areas are constructed according to the coal sample testing process, and corresponding monitoring points are set according to the equipment parameters and work content in different working areas. The image monitoring data in the working areas is collected in real time, and whether there are violations is identified according to the corresponding primary monitoring model, and fraud behaviors are warned in time to ensure the authenticity of the coal sample testing results.

[0031] In the preferred embodiment of the embodiment of the present application, when judging whether to generate a correction instruction, it includes: Obtain the characteristic index parameters of the target working area at the current adjustment time node, and generate a correction evaluation value f; f = (ηi * ki); Where θ is the number of characteristic indexes; ηi is the influence factor of the i-th characteristic index; ki is the reference value of the i-th characteristic index of the target working area at the current adjustment time node; Preset the first correction evaluation value threshold F1; If f < F1, at the current adjustment time node, the target working area does not generate a correction instruction, and set the compensation coefficient r = 1; If f ≥ F1, at the current adjustment time node, the target working area generates a primary correction instruction, and set the compensation coefficient r according to the correction evaluation value f; Judge in turn whether each working area generates a correction instruction at the current adjustment time node.

[0032] Specifically, the characteristic indexes include but are not limited to parameters such as the current quantity of coal samples, coal sample suppliers, and operators. The larger the correction evaluation value, the greater the possibility of fraud behavior in the current target working area.

[0033] Specifically, the values of each characteristic index are set according to historical parameters, and the value ranges of each characteristic index are the same. For example, the historical working parameters of each operator can be obtained according to historical data. The more violations and frauds occur, the larger the corresponding reference value. The larger the reference values of each characteristic index, the greater the possibility of violations and frauds in the parameters represented by the current characteristic index.

[0034] Specifically, the primary correction instruction refers to dynamically adjusting the compensation coefficient to correct the real-time data collected in the target working area, improve the accuracy of fraud behavior identification, and at the same time shorten the time interval between adjacent monitoring time nodes in the target working area.

[0035] Specifically, when setting the compensation coefficient r according to the correction evaluation value f, it includes: Preset a first corrected evaluation value interval (F1, F2), a second corrected evaluation value interval (F2, F3 and a third corrected evaluation value interval (F3, F4); If f is in the preset first compensation evaluation value interval, the compensation coefficient r is set to the preset first compensation coefficient r1, that is, r=r1; If f is in the preset second compensation evaluation value interval, the compensation coefficient r is set to the preset second compensation coefficient r2, ie r=r2; If f is in the preset first compensation evaluation value interval, the compensation coefficient r is set to the preset third compensation coefficient r3, that is, r=r3; and 1 <r1<r2<r3。

[0036] It can be understood that in the above embodiment, multiple characteristic indicators are constructed based on factors such as the supplier of coal samples and process operators, and whether to correct the first-level monitoring model is determined according to the real-time parameters of each characteristic indicator, thereby improving the efficiency of judging violations and ensuring timely warning of fraudulent behavior.

[0037] In a preferred embodiment of the present application, when generating the risk assessment value of each work area according to all monitoring data, it includes: According to the working area sequence A, set ai as the working area to be monitored in sequence; Obtain monitoring data of the work area to be monitored at the current monitoring time node; Preprocess the monitoring data according to the behavior recognition model of the work area to be monitored; Generate potential risk values ​​for each violation based on the preprocessing results; Establish a potential risk value sequence C, C = (c1, c2…ci…cm), where ci is the potential risk value of the i-th violation in the work area to be monitored at the current monitoring time node; m is the number of violation types in the work area to be monitored; Generate the risk assessment value g of the work area to be monitored at the current monitoring time node according to the potential risk value sequence C; Update the risk assessment value g according to the monitoring time nodes of the work area to be monitored.

[0038] Specifically, the collected real-time image data is analyzed according to the behavior recognition model of the work area to be monitored, the occurrence probability of each violation is generated, and the corresponding potential risk value is set according to its occurrence probability. The greater the occurrence probability, the greater the corresponding potential risk value.

[0039] Specifically, when generating the risk assessment value g of the work area to be monitored at the current monitoring time node, it includes: Get the compensation coefficient r' of the monitoring time node of the working area to be monitored; H=e1*Q1*[ βi * ci]+e2 * Q2 * βi * Y(i)(ci - c')]; g = r' * H; Wherein, e1 is a preset first weight coefficient; e2 is a preset second weight coefficient; Q1 is a preset first fixed coefficient; Q2 is a preset second fixed coefficient; βi is the influence factor of the i-th type of violation behavior in the work area to be monitored; c' is the preset potential risk value threshold; Y(i) is a selection coefficient; if (ci - c') < 0, Y(i) = 0; if (ci - c') ≥ 0, Y(i) = 1.

[0040] Specifically, all parameters in the model are normalized by presetting the first fixed coefficient and the second fixed coefficient, so that each parameter is within the same value range.

[0041] Specifically, the influence factor of the violation behavior can be set according to the interference degree of the violation behavior on the coal sample detection. The greater the influence on the accuracy of the coal sample detection result, the greater the corresponding influence factor.

[0042] Specifically, when judging whether to generate a warning instruction, it includes: Presetting a first risk evaluation value threshold G1 and a second risk evaluation value threshold G2; If g < G1, no warning instruction is generated for the work area to be monitored at the current monitoring time node; If G1 < g < G2, a first-level warning instruction is generated for the work area to be monitored at the current monitoring time node; If g > G2, a second-level warning instruction is generated for the work area to be monitored at the current monitoring time node.

[0043] Specifically, the first-level warning instruction means that there is a violation behavior in the current work area to be monitored, but it will not have a great impact on the detection result. A reminder message needs to be sent in time to correct the corresponding violation behavior. The second-level warning instruction means that there is a fraud behavior in the current monitoring area, which needs to be stopped immediately, and the corresponding detection result needs to be excluded to ensure the authenticity of the coal sample detection result.

[0044] Based on another preferred embodiment of a method for analyzing coal sample testing fraud behavior in any one of the above preferred embodiments, in this preferred embodiment, a system for analyzing coal sample testing fraud behavior is provided, including: A central control unit, configured to establish multiple work areas according to the coal sample testing process and construct a first-level monitoring model for each work area; A correction unit, configured to preset multiple characteristic indicators, obtain the characteristic indicator parameters in each work area according to the preset adjustment time node, and judge whether to generate a correction instruction; A monitoring unit, used to collect monitoring data in each working area; The central control unit includes: The first processing module is used to generate a risk assessment value for each work area based on all monitoring data; The first warning module is used to determine whether to generate a warning instruction based on all risk assessment values; The second processing module is used to establish a working area sequence A, A=(a1, a2…ai…an), where ai is the i-th working area; n is the number of working areas; The third processing module is used to construct a primary monitoring model for each working area.

[0045] Specifically, the monitoring submodule is preferably a high-definition camera, which can achieve full coverage of the working area by setting up multiple monitoring points in each working area, and timely identify various violations and frauds.

[0046] In a preferred embodiment of the present application, the third processing module is also used for: According to the working area sequence A, set ai as the target working area in sequence; Obtain workflow parameters and historical parameters of the target work area; Set multiple violations in the target work area based on historical parameters, and create a training data package based on all violations; Iteratively optimize the training data package to generate a behavior recognition model for the target work area; Set the data collection frequency of the target work area according to the workflow data, and set multiple monitoring time nodes according to the data collection frequency; Generate a primary monitoring model for the target work area based on the data collection frequency and behavior recognition model; Generate the first-level monitoring model of each working area in turn; Establish a first-level monitoring model series B, B=(b1, b2…bi…bn), where bi is the first-level monitoring model of the i-th working area.

[0047] In a preferred embodiment of the present application, the correction unit is also used for: Obtain characteristic index parameters of the target working area at the current adjustment time node and generate a modified evaluation value f; f= (ηi*ki); Among them, θ is the number of characteristic indicators; ηi is the influencing factor of the i-th characteristic indicator; ki is the reference value of the i-th characteristic indicator of the target working area at the current adjustment time node; Preset the first modified evaluation value threshold F1; If f < F1 at the current adjustment time node, no correction instruction is generated for the target working area, and the compensation coefficient r is set to 1; If f ≥ F1 at the current adjustment time node, a first-level correction instruction is generated for the target working area, and the compensation coefficient r is set according to the correction evaluation value f; Judge in turn whether correction instructions are generated for each working area at the current adjustment time node.

[0048] According to the first concept of the present application, multiple working areas are constructed according to the coal sample testing process, corresponding monitoring points are set according to the equipment parameters and work content in different working areas, image monitoring data in the working areas is collected in real time, and whether there are violations is identified according to the corresponding first-level monitoring model, and fraud behavior is warned in time to ensure the authenticity of the coal sample testing results.

[0049] According to the second concept of the present application, multiple characteristic indicators are constructed based on factors such as the supplier of coal samples and process operators, and whether to correct the first-level monitoring model is judged according to the real-time parameters of each characteristic indicator, so as to improve the judgment efficiency of violations and ensure timely warning of fraud behavior.

[0050] The above are only the preferred embodiments of the present application. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present application, several improvements and replacements can still be made, and these improvements and replacements should also be regarded as the protection scope of the present application.

Claims

1. A method for analyzing fraudulent behavior in coal sample testing, characterized in that: Including: Establish multiple working areas according to the coal sample testing process, and construct a first-level monitoring model for each working area; Preset multiple characteristic indicators, obtain the characteristic indicator parameters in each working area according to the preset adjustment time node, and determine whether to generate a correction instruction; Obtain the monitoring data in each working area, and generate a risk evaluation value for each working area according to all the monitoring data; Judge whether to generate a warning instruction according to all the risk evaluation values; Among them, when establishing multiple working areas, it includes: Establish a working area sequence A, A = (a1, a2…ai…an), where ai is the i-th working area; n is the number of working areas.

2. The coal sample testing fraud analysis method according to claim 1, characterized in that: When constructing the first-level monitoring model for each working area, it includes: Set ai as the target working area in sequence according to the working area sequence A; Obtain the working process parameters and historical parameters of the target working area; Set various types of violations in the target working area according to the historical parameters, and establish a training data packet according to all the violations; Iteratively optimize the training data packet to generate a behavior recognition model for the target working area; Set the data collection frequency of the target working area according to the working process data, and set multiple monitoring time nodes according to the data collection frequency; Generate the first-level monitoring model of the target working area according to the data collection frequency and the behavior recognition model; Generate the first-level monitoring models of each working area in sequence; Establish a first-level monitoring model sequence B, B = (b1, b2…bi…bn), where bi is the first-level monitoring model of the i-th working area.

3. The coal sample testing fraud analysis method according to claim 2, characterized in that: When judging whether to generate a correction instruction, it includes: Obtain the characteristic indicator parameters of the target working area at the current adjustment time node, and generate a correction evaluation value f; f= (ηi*ki); Among them, θ is the number of characteristic indicators; ηi is the influence factor of the i-th characteristic indicator; ki is the reference value of the i-th characteristic indicator of the target working area at the current adjustment time node; Preset a first correction evaluation value threshold F1; If f < F1, at the current adjustment time node, the target working area does not generate a correction instruction, and set the compensation coefficient r = 1; If f ≥ F1, at the current adjustment time node, the target working area generates a first-level correction instruction, and set the compensation coefficient r according to the correction evaluation value f; Judge in sequence whether each working area generates a correction instruction at the current adjustment time node.

4. The method for analyzing fraudulent behavior in coal sample testing according to claim 3, characterized in that: When setting the compensation coefficient r according to the correction evaluation value f, it includes: Preset a first correction evaluation value interval (F1, F2), a second correction evaluation value interval (F2, F3), and a third correction evaluation value interval (F3, F4); If f is in the preset first compensation evaluation value interval, set the compensation coefficient r as the preset first compensation coefficient r1, that is, r = r1; If f is in the preset second compensation evaluation value interval, set the compensation coefficient r as the preset second compensation coefficient r2, that is, r = r2; If f is in the preset first compensation evaluation value interval, set the compensation coefficient r as the preset third compensation coefficient r3, that is, r = r3; and 1 < r1 < r2 < r3.

5. The method for analyzing fraudulent behavior in coal sample testing according to claim 3, characterized in that: When generating the risk evaluation value of each working area according to all the monitoring data, it includes: Set ai as the working area to be monitored in sequence according to the working area sequence A; Obtain the monitoring data of the work area to be monitored at the current monitoring time node; Preprocess the monitoring data according to the behavior recognition model of the work area to be monitored; Generate the potential risk values of each violation behavior according to the preprocessing results; Establish a sequence C of potential risk values, C = (c1, c2…ci…cm), where ci is the potential risk value of the i-th violation behavior in the work area to be monitored at the current monitoring time node; m is the number of types of violation behaviors in the work area to be monitored; Generate the risk evaluation value g of the work area to be monitored at the current monitoring time node according to the sequence C of potential risk values; Update the risk evaluation value g according to the monitoring time node of the work area to be monitored.

6. The method for analyzing fraudulent behavior in coal sample testing according to claim 5, characterized in that: When generating the risk evaluation value g of the work area to be monitored at the current monitoring time node, it includes: Obtain the compensation coefficient r' of the monitoring time node of the work area to be monitored; H=e1*Q1*[ βi*ci]+e2*Q2*[ βi*Y(i)(ci-c')]; g = r' * H; Where, e1 is the preset first weight coefficient; e2 is the preset second weight coefficient; Q1 is the preset first fixed coefficient; Q2 is the preset second fixed coefficient; βi is the influence factor of the i-th violation behavior in the work area to be monitored; c' is the preset potential risk value threshold; Y(i) is the selection coefficient; if (ci - c') < 0, Y(i) = 0; if (ci - c') < 0, Y(i) = 1.

7. The method for analyzing fraudulent behavior in coal sample testing according to claim 6, characterized in that: When judging whether to generate a warning instruction, it includes: Preset the first risk evaluation value threshold G1 and the second risk evaluation value threshold G2; If g < G1, no warning instruction is generated for the work area to be monitored at the current monitoring time node; If G1 < g < G2, a first-level warning instruction is generated for the work area to be monitored at the current monitoring time node; If g > G2, a second-level warning instruction is generated for the work area to be monitored at the current monitoring time node.

8. A coal sample testing fraud analysis system, using the coal sample testing fraud analysis method according to any one of claims 1 to 7, characterized in that: It includes: A central control unit, used to establish multiple work areas according to the coal sample testing process and construct the first-level monitoring models of each work area; A correction unit, used to preset multiple characteristic indicators, obtain the characteristic indicator parameters in each work area according to the preset adjustment time node, and judge whether to generate a correction instruction; A monitoring unit, used to collect the monitoring data in each work area; The central control unit includes: A first processing module, used to generate the risk evaluation values of each work area according to all the monitoring data; A first warning module, used to judge whether to generate a warning instruction according to all the risk evaluation values; A second processing module, used to establish a sequence A of work areas, A = (a1, a2…ai…an), where ai is the i-th work area; n is the number of work areas; A third processing module, used to construct the first-level monitoring models of each work area.

9. The coal sample testing fraud analysis system according to claim 8, characterized in that: The third processing module is also used for: Sequentially set ai as the target work area according to the sequence A of work areas; Obtain the work process parameters and historical parameters of the target work area; Set various violation behaviors in the target work area according to the historical parameters, and establish a training data packet according to all the violation behaviors; Iteratively optimize the training data packet to generate the behavior recognition model of the target work area; Set the data collection frequency of the target work area according to the work process data, and set multiple monitoring time nodes according to the data collection frequency; Generate a primary monitoring model for the target working area according to the data collection frequency and the behavior recognition model; Generate the primary monitoring models for each working area in sequence; Establish a sequence B of primary monitoring models, B = (b1, b2…bi…bn), where bi is the primary monitoring model of the i-th working area.

10. The coal sample testing fraud analysis system according to claim 9, characterized in that: The correction unit is further configured to: Obtain the characteristic index parameters of the target working area at the current adjustment time node and generate a correction evaluation value f; f= (ηi*ki); where θ is the number of characteristic indexes; ηi is the influence factor of the i-th characteristic index; ki is the reference value of the i-th characteristic index of the target working area at the current adjustment time node; Preset a first correction evaluation value threshold F1; If f < F1, at the current adjustment time node, the target working area does not generate a correction instruction and sets the compensation coefficient r = 1; If f ≥ F1, at the current adjustment time node, the target working area generates a primary correction instruction and sets the compensation coefficient r according to the correction evaluation value f; Judge in sequence whether each working area generates a correction instruction at the current adjustment time node.

Images