Deviation traceability and quantitative evaluation method and device in coal sample preparation process

Abstract

The invention relates to the technical field of coal quality detection and process control, and discloses a coal sample preparation process deviation traceability and quantitative evaluation method and device, and the method achieves the precise definition of an error source through the standardization of a disassembly sample preparation process, and the construction of a deviation transmission directed acyclic graph and a Bayesian model. Error source self-made sample preparation, sampling and test links can be clearly distinguished, and specific problem procedures in the sample preparation process can be accurately positioned. Meanwhile, the deviation contribution degree is quantitatively calculated based on real-time data and probabilistic reasoning, a traditional subjective qualitative judgment mode depending on an engineer to observe the appearance of a sample is thoroughly abandoned, and a quantitative conclusion of the influence of deviation on a test result is accurately output.

Description

Technical Field

[0001] This invention relates to the field of coal quality testing and process control technology, specifically to a method and apparatus for tracing and quantitatively evaluating deviations in the coal sample preparation process. Background Technology

[0002] Coal sample preparation is a crucial link in the coal quality testing system, connecting sampling and analysis. Through a series of operations such as crushing, mixing, reducing, and drying, the raw coal sample is prepared into a representative analytical specimen, providing a foundation for subsequent testing of key quality indicators such as calorific value, sulfur content, and ash content. The accuracy of coal quality indicators is directly related to the fairness of trade settlements and the scientific nature of coal blending in production. Deviations in sample preparation can be transmitted and amplified to the final test results, potentially leading to trade disputes, or causing problems such as incorrect coal blending decisions and increased energy consumption. Therefore, the standardization and reliability of coal sample preparation are essential for the efficient operation of the entire coal industry chain and are a prerequisite for ensuring the accuracy of coal quality testing.

[0003] Therefore, the quality assessment of sample preparation typically employs a results-oriented control approach, recording only basic data such as raw coal sample weight and preparation time, and inferring the quality of the sample preparation from the final test results. However, this method lacks traceability and quantitative assessment capabilities. Existing records cannot fully reconstruct the specific operational status of each stage of crushing and reduction processes. When test results are abnormal, it is difficult to accurately determine whether the error originates from the sample preparation, sampling, or testing process. Even if the sample preparation process is identified, the specific problematic process cannot be pinpointed. Furthermore, relying on engineers to observe the appearance characteristics of the sample, such as particle size and moisture content, for qualitative judgment is highly subjective and cannot provide quantitative conclusions on the impact of deviations on test results. Summary of the Invention

[0004] This invention provides a method and apparatus for tracing and quantifying deviations in the coal sample preparation process. It addresses the problems of difficulty in accurately determining whether the source of error is in the sample preparation, sampling, or testing stages, and even if the sample preparation stage is identified, the specific problematic process cannot be located. Furthermore, it relies on engineers to observe the appearance characteristics of the sample, such as particle size and humidity, for qualitative judgment, which is highly subjective and cannot provide quantitative conclusions on the impact of deviations on the test results.

[0005] In a first aspect, the present invention provides a method for tracing and quantitatively evaluating deviations in the coal sampling process, the method comprising: Input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node. A directed acyclic graph is constructed by setting the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node. A conditional probability table is assigned to each node of the directed acyclic graph, and a sample preparation bias Bayesian model is constructed based on all the nodes and their corresponding conditional probability tables. The real-time coal sample preparation process data is input into the sample preparation deviation Bayesian model, and the real-time coal sample preparation process is evaluated through the sample preparation deviation Bayesian model to obtain the evaluation results.

[0006] This invention achieves precise identification of error sources by standardizing the sample preparation process, constructing a directed acyclic graph of deviation propagation, and using a Bayesian model. It can clearly distinguish whether the error originates from sample preparation, sampling, or testing, and accurately pinpoint the specific problematic step in the sample preparation process. Simultaneously, relying on real-time data and probabilistic reasoning to quantify the contribution of deviations, it completely abandons the traditional subjective qualitative judgment mode that relies on engineers' observation of sample appearance, and accurately outputs quantitative conclusions on the impact of deviations on test results.

[0007] In one optional implementation, the step of inputting coal sampling process data from a preset coal field operation procedure into a preset sampling process model, and outputting multiple sampling process nodes and the deviation influencing factors of each sampling process node, includes: Obtain on-site coal mining operation procedures; Input the coal sample preparation process data from the coal field operation procedure into the preset sample preparation process model. By analyzing the entire coal sample preparation process data using the aforementioned sample preparation process model, a series of continuous sample preparation process nodes are obtained. The factors influencing deviations from the ideal state at each of the sample preparation process steps were analyzed.

[0008] This invention achieves standardized breakdown of the entire sample preparation process by analyzing on-site coal operation procedures and inputting a sample preparation process model. It accurately outputs continuous sample preparation process nodes and identifies the deviation influencing factors at each node. This overcomes the limitations of traditional sample preparation processes, which are often vaguely broken down and have fragmented causes for deviation. It lays a structured foundation for subsequently constructing a deviation transmission network and tracing the source of errors.

[0009] In one optional implementation, the construction of a directed acyclic graph, with the deviation influencing factors of each of the sample preparation process nodes as parent nodes and each of the coal sample preparation process nodes as child nodes, includes: Set all deviation influencing factors of each sample preparation process node as parent nodes, and set each sample preparation process node as child nodes; A directed acyclic graph is obtained by connecting the parent node and its corresponding child nodes with directed arrows.

[0010] This invention constructs a directed acyclic graph (DAG) with directed arrows, setting the deviation influencing factors of sample preparation process nodes as parent nodes and the corresponding process nodes as child nodes. This clearly outlines the transmission path of deviation from the source factors to the process result. By establishing a clear association between dispersed deviation influencing factors and process nodes, the transmission logic of deviations in each stage is presented intuitively, improving the logic and traceability of deviation control in the sample preparation process.

[0011] In one optional implementation, assigning conditional probability tables to each node of the directed acyclic graph and constructing a sample preparation bias Bayesian model based on all the nodes and their corresponding conditional probability tables includes: Define the node state of each node in the directed acyclic graph; Based on the node state combinations of each parent node and the node state probability distribution of the corresponding child nodes, a conditional probability table for each node is constructed. Based on historical operating data, equipment design tolerances, experimental test data, and expert experience, the conditional probability parameters in the conditional probability table for each node are determined. A sample preparation bias Bayesian model is constructed using each node and its corresponding conditional probability table.

[0012] This invention constructs a Bayesian model of sample preparation deviation by defining node states, building conditional probability tables, and fusing multi-source data to determine probability parameters. It transforms the causal relationships of deviations in the sample preparation process into quantified probabilistic dependencies, laying the core model foundation for subsequent probabilistic inference using real-time input data. This model achieves a leap from qualitative description to quantitative modeling of sample preparation deviations, providing scientific algorithmic support for the quantitative assessment and precise tracing of deviation impacts, and significantly improving the scientific rigor and accuracy of sample preparation quality control.

[0013] In one optional implementation, the step of inputting real-time coal sampling process data into the sampling deviation Bayesian model, and evaluating the real-time coal sampling process using the sampling deviation Bayesian model to obtain evaluation results includes: Acquire real-time current values, equipment status parameters, operating parameters, and intermediate product quality parameters during the real-time coal sampling process; The real-time current value, the equipment status parameters, the operating parameters, and the intermediate product quality parameters are processed and input into the sample preparation deviation Bayesian model. The sample preparation deviation Bayesian model loads each node and its corresponding conditional probability table, and transforms the real-time current value, equipment status parameters, operating parameters and intermediate product quality parameters after data processing into discrete state evidence that can be identified by the sample preparation deviation Bayesian model. The sample preparation deviation Bayesian model is compiled using a preset Bayesian algorithm, and all the discrete state evidence is input into the compiled sample preparation deviation Bayesian model for global probability propagation. The probability distribution of unknown nodes in the compiled sample preparation bias Bayesian model is updated based on the propagation results. Extract the posterior probability distribution of all nodes in the updated sample preparation deviation Bayesian model, and then select the posterior probability distribution of the preset target ash content deviation nodes. Extract the posterior probability value corresponding to the state of each node from the posterior probability distribution of the target ash deviation node and select the maximum posterior probability value; Based on the maximum posterior probability value, the optimal state of the target ash deviation node and the deviation evaluation information are determined; The prior probability value and posterior probability value of each node are compared, and the node corresponding to the key deviation influencing factor is determined based on the comparison result. The difference between the prior probability value and the posterior probability value of each node is calculated to obtain the contribution of each node. An evaluation result is generated by using the deviation assessment information of the target ash deviation node, the nodes corresponding to the key deviation influencing factors, and the contribution of each node.

[0014] This invention collects multi-dimensional real-time sample preparation data, processes it into discrete state evidence recognizable by the model, and uses a Bayesian algorithm to complete model compilation and global probability propagation, accurately updating the node probability distribution. By extracting the posterior probability of target ash deviation nodes and selecting the maximum probability value to determine deviation assessment information, it combines prior and posterior probabilities to identify key deviation sources and calculate their contribution, ultimately generating a quantitative assessment result. This achieves real-time quantitative assessment and precise source tracing of sample preparation deviations, abandoning traditional subjective judgment and clearly defining the degree of deviation impact and core causes.

[0015] In an optional implementation, the method further includes: The evaluation results are sent to the operation interface or centralized control system to generate an inspection work order that matches the evaluation results; Based on the inspection work order, predictive maintenance operations are performed.

[0016] This invention promotes a shift in sample preparation equipment maintenance from reactive to predictive maintenance by synchronizing quantitative evaluation results to the user interface or centralized control system and generating matching inspection work orders. Based on the work orders, it accurately identifies the deviation-related links requiring repair, avoiding blind troubleshooting, shortening maintenance response time, and reducing the risk of unplanned downtime.

[0017] Secondly, the present invention provides a device for tracing and quantifying deviations in coal sample preparation processes, the device comprising: The input module is used to input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node. The construction module is used to construct a directed acyclic graph by setting the deviation influencing factors of each of the sample preparation process nodes as parent nodes and each of the coal sample preparation process nodes as child nodes. The allocation module is used to assign conditional probability tables to each node of the directed acyclic graph, and to construct a sample preparation bias Bayesian model based on all the nodes and the corresponding conditional probability tables. The evaluation module is used to input real-time coal sampling process data into the sampling deviation Bayesian model, evaluate the real-time coal sampling process through the sampling deviation Bayesian model, and obtain the evaluation result.

[0018] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the coal sampling process deviation tracing and quantitative evaluation method described in the first aspect or any corresponding embodiment.

[0019] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for tracing and quantifying deviations in the coal sample preparation process described in the first aspect or any corresponding embodiment.

[0020] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the method for tracing and quantifying deviations in the coal sample preparation process described in the first aspect or any corresponding embodiment. Attached Figure Description

[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the first type of the method for tracing and quantitatively evaluating deviations in the coal sampling process according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the second process of the method for tracing and quantitatively evaluating deviations in the coal sampling process according to an embodiment of the present invention. Figure 3This is a schematic diagram of a directed acyclic graph according to an embodiment of the present invention; Figure 4 This is a structural block diagram of a coal sample preparation process deviation tracing and quantitative evaluation device according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0023] 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 are only some embodiments of the present invention, not all embodiments. 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.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0025] This invention provides a method for tracing and quantitatively assessing deviations in coal sample preparation. By standardizing and decomposing the sample preparation process, constructing a directed acyclic graph of deviation propagation, and using a Bayesian model, the method achieves precise identification of error sources. It can clearly distinguish whether the error originates from the sample preparation, sampling, or testing stages, and accurately pinpoint the specific problematic steps in the sample preparation process. Simultaneously, relying on real-time data and probabilistic reasoning, it quantitatively calculates the contribution of deviations, thereby completely abandoning the traditional subjective qualitative judgment mode that relies on engineers' observation of sample appearance, and accurately outputting quantitative conclusions on the impact of deviations on test results.

[0026] According to an embodiment of the present invention, an embodiment of a method for tracing and quantitatively evaluating deviations in coal sampling process is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] This embodiment provides a method for tracing and quantitatively evaluating deviations in the coal sample preparation process. Figure 1 This is a flowchart of the method for tracing and quantitatively evaluating deviations in the coal sampling process according to an embodiment of the present invention, as shown below. Figure 1 As shown, the process includes the following steps: Step S101: Input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node.

[0028] It should be noted that the coal field operation procedures refer to the standardized documents developed at the coal sample preparation site to regulate operational behavior and ensure sample quality and safety.

[0029] Coal sample preparation process data refers to various types of data related to the sample preparation process generated based on coal field operation procedures.

[0030] A sample preparation process model refers to a structured model that is pre-constructed based on relevant coal sample preparation standards, industry norms, and the actual sample preparation process logic.

[0031] Sample preparation process nodes refer to the continuous units formed after the entire sample preparation process is broken down into standardized steps. Each node corresponds to a specific sample preparation operation (such as primary crushing, primary reduction, drying, etc.).

[0032] Deviation influencing factors refer to various variables that may cause the output results of a sample preparation process node to deviate from the preset ideal state.

[0033] In this embodiment of the invention, coal sample preparation process data, such as the full-process operation requirements, process parameter thresholds, and quality control standards specified in the preset coal field operation procedures, are input into a sample preparation process model pre-constructed based on a preset standard (such as GB474-2008 standard). This model automatically decomposes the entire sample preparation process and outputs multiple continuous sample preparation process nodes with clear input and output boundaries by parsing the process logic and constraints in the operation procedures. At the same time, combined with the sample preparation process principle and historical deviation cases, it simultaneously identifies the deviation factors that may cause the output results to deviate from the ideal state in actual operation of each sample preparation process node.

[0034] Step S102: Construct a directed acyclic graph by setting the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node.

[0035] It should be noted that, in a directed acyclic graph, the parent node represents the source factor of sample preparation deviation, that is, the deviation influencing factors of each sample preparation process node (such as the wear of crusher blades, feeding speed, etc.), which is the "cause" that causes the deviation of the child node (process node).

[0036] In a directed acyclic graph, a child node refers to a sample preparation process unit that is affected by the parent node. These child nodes (such as primary crushing, primary reduction, etc.) are the "results" of deviations caused by the parent node (the factor affecting deviation).

[0037] A directed acyclic graph (DAG) is a structured graph consisting of nodes (parent nodes, child nodes) and directed arrows.

[0038] In this embodiment of the invention, the deviation influencing factors corresponding to each sample preparation process node are taken as parent nodes, and each sample preparation process node itself is taken as a child node. According to the transmission logic of deviation from influencing factors to process results, each parent node and its corresponding child node are connected one by one with directed arrows to ensure that the arrow direction always points from the deviation cause to the deviation result, and that all nodes and connection relationships do not form a loop. Finally, a directed acyclic graph that can clearly characterize the generation and transmission path of deviation in the sample preparation process is constructed.

[0039] Step S103: Assign conditional probability tables to each node of the directed acyclic graph, and construct a sample preparation bias Bayesian model based on all nodes and their corresponding conditional probability tables.

[0040] It should be noted that the conditional probability table (CPT) is a table used to quantitatively describe the probabilistic dependencies between nodes in a Bayesian network.

[0041] A sample-biased Bayesian model refers to a probabilistic inference model that uses a directed acyclic graph as its structural framework and a conditional probability table as its probability dependency basis.

[0042] In this embodiment of the invention, for each node in the directed acyclic graph, its discrete state is first defined (e.g., the parent node "tool wear degree" is set to normal, light wear, and severe wear, and the child node "primary breakage" is set to qualified, coarse, and fine). Then, based on the dependency relationship between each child node and its parent node, a conditional probability table for each node is constructed. This table quantitatively describes the probability distribution of each state corresponding to each state when the parent node is in different state combinations. Subsequently, historical operating data, equipment design tolerances, experimental data, and expert experience are integrated to determine the probability parameters in the table. Finally, all nodes in the directed acyclic graph, the directed connection relationships between nodes, and the conditional probability tables corresponding to each node are integrated to construct a sample preparation deviation Bayesian model that can realize deviation quantification reasoning.

[0043] Step S104: Input the real-time coal sample preparation process data into the sample preparation deviation Bayesian model, evaluate the real-time coal sample preparation process through the sample preparation deviation Bayesian model, and obtain the evaluation results.

[0044] It should be noted that real-time coal sample preparation process data refers to various dynamic data related to sample quality collected in real time during the sample preparation process.

[0045] The evaluation result refers to the comprehensive conclusion output by the model after inference, which includes quantified deviation information (such as the range and confidence interval of ash deviation), key deviation influencing factors (such as tool wear), and the ranking of deviation contribution.

[0046] In this embodiment of the invention, data such as equipment status parameters, operating parameters, intermediate product quality parameters, and real-time current values ​​during the coal sample preparation process are collected in real time. After cleaning and normalization, these data are transformed into discrete state evidence that can be identified by the Bayesian model of sample preparation deviation. After inputting this evidence into the model, the model loads a preset directed acyclic graph structure and conditional probability tables for each node. Through a preset Bayesian inference algorithm, global probability propagation is completed, the probability distribution of unknown nodes is updated, and then the posterior probability distribution of the target ash deviation node and the probability change of each deviation source node are extracted. Deviation assessment information, key deviation influencing factors, and contribution are determined, and finally, an assessment result containing quantitative deviation results and source tracing conclusions is obtained.

[0047] This embodiment provides a method for tracing and quantitatively evaluating deviations in the coal sample preparation process. Figure 2 This is a flowchart of the method for tracing and quantitatively evaluating deviations in the coal sampling process according to an embodiment of the present invention, as shown below. Figure 2 As shown, the process includes the following steps: Step S201: Input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node.

[0048] Specifically, step S201 includes: Step S2011: Obtain the coal field operation procedures.

[0049] In this embodiment of the invention, the coal sample preparation process, i.e., the coal field operation procedure, is obtained according to a preset standard (such as GB474-2008).

[0050] Step S2012: Input the coal sample preparation process data from the coal field operation procedure into the preset sample preparation process model.

[0051] In this embodiment of the invention, the coal sampling process data, including the sampling steps, process parameter thresholds, operation specifications, and quality control standards recorded in the coal field operation procedures, are organized according to a preset data format and input into a preset sampling process model. The sampling process model will parse and decompose the input data based on the built-in coal sampling process logic and preset standard specifications.

[0052] Step S2013: Analyze the entire coal sample preparation process data through the sample preparation process model to obtain a series of continuous sample preparation process nodes.

[0053] It should be noted that the entire sample preparation process refers to the complete process from receiving the raw coal, through a series of operations such as crushing, reducing, drying, and grinding, to finally preparing an analytical sample that meets the testing requirements.

[0054] In this embodiment of the invention, the entire process of coal sample preparation is analyzed and decomposed using a sample preparation process model, breaking it down into a series of continuous sample preparation process nodes (such as primary crushing → primary reduction → secondary crushing → drying → secondary reduction...).

[0055] Step S2014: Analyze the factors that cause deviations from the ideal state at each sample preparation process node.

[0056] It should be noted that the ideal state refers to the output state that should be achieved when each sample preparation process node is operated in accordance with the preset coal sample preparation standards (such as GB474-2008 "Methods for Preparing Coal Samples") and the requirements of on-site operating procedures.

[0057] Deviation influencing factors refer to various variables that may cause the output results of sample preparation process nodes to deviate from the ideal state.

[0058] In this embodiment of the invention, by combining the sample preparation process principle, equipment operating characteristics and historical sample preparation deviation cases, a targeted analysis is conducted on each sample preparation process node to identify various deviation factors that may cause the output result of the node to deviate from the preset ideal state (i.e., the process output state that meets the requirements of the coal sample preparation standard) during actual operation.

[0059] Step S202: Construct a directed acyclic graph by setting the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node.

[0060] In some optional implementations, step S202 above includes: Step S2021: Set all deviation influencing factors of each sample preparation process node as parent nodes, and set each sample preparation process node as child nodes.

[0061] In this embodiment of the invention, all deviation influencing factors of each sample preparation process node are set as parent nodes, and each sample preparation process node is set as a child node. At the same time, the causal relationship between the parent node and the corresponding child node is clarified, that is, the deviation influencing factors represented by the parent node are the causes that cause the output results of the child node (sample preparation process node) to deviate from the ideal state.

[0062] Step S2022: Connect the parent node and its corresponding child nodes with directed arrows to obtain a directed acyclic graph.

[0063] It should be noted that a directed arrow is a symbol used to clarify the direction of deviation transmission between parent and child nodes, and its direction directly reflects the causal logic of "deviation influencing factors → sample preparation process nodes".

[0064] In embodiments of the present invention, such as Figure 3As shown, directed arrows are used to connect the parent node and the corresponding child node. The arrow direction is set strictly according to the transmission logic of deviation from cause to result. That is, the arrow always points from the parent node (the factor affecting deviation) to the corresponding child node (the sample preparation process node). At the same time, it is ensured that all nodes and connections do not form a reverse loop. Finally, a directed acyclic graph that can clearly and intuitively represent the generation and transmission path of deviation in the sample preparation process is obtained.

[0065] First, based on pre-defined standards (such as GB474-2008) and on-site operating procedures, the entire coal sample preparation process is decomposed into a series of continuous process units (sample preparation process nodes) with clearly defined inputs and outputs. These units constitute the nodes of a directed acyclic graph (DAG). For example, a typical process might include: raw coal sample → primary crushing → primary fractionation → secondary crushing → drying → secondary fractionation → analytical sample.

[0066] Deviation Factor Identification: For each sample preparation step, analyze the "deviation influencing factors" that may cause the output result to deviate from the ideal state. These factors constitute the "parent node" of that step. For example: The "Particle size after primary crushing" node may have parent nodes such as "crusher cutter wear", "feeding speed", and "initial particle size of raw coal".

[0067] The "Representative after primary reduction" node: its parent node may include "reducer cutting speed", "reduction ratio", "particle size after primary crushing" (because particle size affects the uniformity of mixing), etc.

[0068] Construct a directed acyclic graph (DAG): Connect nodes with directed arrows, where the arrow direction represents the direction of deviation propagation, i.e., from the "cause" (parent node) to the "effect" (child node). For example, an arrow pointing from "crusher cutter wear" to "particle size after primary crushing" indicates that cutter wear is the cause of particle size deviation. Ultimately, all nodes and edges form a complete network, clearly depicting all possible paths of deviation from its source (equipment status, operating parameters) to its final result (analytical deviation).

[0069] Step S203: Assign conditional probability tables to each node of the directed acyclic graph, and construct a sample preparation bias Bayesian model based on all nodes and their corresponding conditional probability tables.

[0070] Specifically, step S203 includes: Step S2031: Define the node state of each node in the directed acyclic graph.

[0071] It should be noted that node state refers to the result of discretizing and classifying the attributes of each node in a directed acyclic graph.

[0072] In this embodiment of the invention, the attributes of each node in the directed acyclic graph are discretized and classified. For example, the parent node "tool wear degree" can be defined as "normal", "slight wear" or "severe wear", and the child node "primary breakage" can be defined as "qualified", "coarse" or "fine".

[0073] Step S2032: Based on the node state combination of each parent node and the node state probability distribution of the corresponding child node, construct the conditional probability table for each node.

[0074] It should be noted that the combination of node states of a parent node refers to the permutation and combination of the different node states of all parent nodes (deviation influencing factors) corresponding to a certain child node.

[0075] The node state probability distribution refers to the set of probability values ​​of a child node (sample preparation process node) in each discrete state under the influence of a specific combination of parent node states.

[0076] In this embodiment of the invention, a conditional probability table is constructed for each child node in the directed acyclic graph. The table clearly lists all possible combinations of the parent nodes of the child node, and at the same time, it marks the probability value of the child node in each discrete state under each combination, ensuring that the sum of the probabilities in each row is 1 to meet the requirements of the probability axiom. In this way, the influence of the parent node state on the child node state is quantitatively characterized.

[0077] Step S2033: Based on historical operating data, equipment design tolerances, experimental test data, and expert experience, determine the conditional probability parameters in the conditional probability table for each node.

[0078] It should be noted that historical operational data refers to various practical data recorded during past coal sample preparation processes.

[0079] Equipment design tolerances refer to the allowable fluctuation range of parameters specified during the design of sample preparation equipment. They reflect the performance boundaries of normal equipment operation and can be used to correct statistically obtained probability parameters.

[0080] Experimental test data refers to data obtained through specially designed sample preparation experiments, which are usually used to simulate the output state of the sample preparation process under different combinations of deviation influencing factors.

[0081] Expert experience refers to the judgment and understanding of sample preparation deviation patterns and parameter correlations formed by technical personnel with rich practical and analytical experience in the field of coal sample preparation based on long-term industry practice.

[0082] In this embodiment of the invention, each node is a random variable, and its node state needs to be explicitly defined. Node states are typically discrete categories to facilitate the construction of a conditional probability table (CPT). For example: The "Crusher Tool Wear" node can be defined as [Normal, Slight Wear, Severe Wear].

[0083] The "Particle size after crushing" node can be defined as [qualified, coarse, fine].

[0084] The “Target Ash Deviation” node can be defined as a range such as [-0.5%~0%, 0%~+0.5%, +0.5%~+1.0%].

[0085] Taking the "crushed particle size" node as an example, its parent nodes are "tool wear" and "feeding speed".

[0086] "Feeding speed" status: [Normal, Too fast]; "Tool Wear" status: [Normal, Wear]; "Particle size after crushing" status: [Acceptable, slightly coarse]; Its CPT may look like this (values ​​are for example):

[0087] Historical operating data learning: By collecting a large amount of process data (such as current and speed) and corresponding result data (such as laboratory granularity analysis results) under normal and abnormal operating conditions, machine learning algorithms (such as the expectation-maximization algorithm) are used to automatically learn and fit CPT parameters (i.e., conditional probability parameters) from the data.

[0088] Equipment design tolerances: For equipment parameters, probabilities can be set based on their design accuracy and tolerance range. For example, if the speed control accuracy is ±2%, then the probability that the speed fluctuation within ±2% is a "normal" state is extremely high.

[0089] Experimental test data: Experiments were conducted by deliberately setting specific working conditions (such as adjusting worn tools and changing the feed rate), recording inputs and outputs, and directly statistically analyzing the state frequencies of the results under different conditions as an estimate of CPT.

[0090] Expert experience: When data is insufficient, domain experts make estimates based on their knowledge and experience. For example, an expert can determine that "when the feed rate is too fast and the cutting tools are worn, the probability of producing coarse particles is very high," thus providing a qualitative probability estimate (such as "high, medium, or low"), which is then converted into specific numerical values.

[0091] Step S2034: Construct a Bayesian model of sample preparation bias using each node and its corresponding conditional probability table.

[0092] In this embodiment of the invention, each node and its corresponding conditional probability table are used, combined with the deviation transmission logic between nodes represented by the directed acyclic graph constructed in step S2022. The nodes are used as the basic building blocks of the model, and the conditional probability table is used as the quantitative carrier of the probability dependency between nodes. This is integrated into the Bayesian inference framework, and finally a sample preparation deviation Bayesian model that can realize the source tracing and quantitative evaluation of sample preparation deviation is constructed.

[0093] Step S204: Input the real-time coal sample preparation process data into the sample preparation deviation Bayesian model, evaluate the real-time coal sample preparation process through the sample preparation deviation Bayesian model, and obtain the evaluation results.

[0094] In some optional implementations, step S204 above includes: Step S2041: Obtain the real-time current value, equipment status parameters, operating parameters, and intermediate product quality parameters of the real-time coal sample preparation process.

[0095] It should be noted that the real-time coal sample preparation process refers to the dynamic, real-time sample preparation process that begins with the receipt of raw coal and proceeds through a series of sample preparation operations such as crushing, reduction, and drying until the analytical sample is prepared.

[0096] Real-time current value refers to the current data monitored in real time during the operation of the sample preparation equipment.

[0097] Equipment status parameters refer to the vibration frequency / amplitude, drive motor power, and bearing temperature of each stage of crusher and divider; and the working status of the heating element and fan speed of the drying oven.

[0098] Operating parameters refer to the feeder frequency (which directly controls the feeding speed), the actual reduction ratio (calculated by the reducer's operating cycle and the feeding speed), the set value and actual value of the drying temperature, and the drying time.

[0099] Intermediate product quality parameters refer to the moisture values ​​measured by an online moisture meter before and after drying, the material flow fluctuations recorded by an online belt scale, and the foreign matter contamination events detected by a metal detector.

[0100] In this embodiment of the invention, real-time coal sample preparation process data, such as real-time current value, equipment status parameters, operating parameters, and intermediate product quality parameters, are obtained through an online image recognition system.

[0101] Step S2042: Process the real-time current value, equipment status parameters, operating parameters, and intermediate product quality parameters and input them into the sample preparation deviation Bayesian model.

[0102] It should be noted that data processing refers to preprocessing operations such as denoising real-time current values, equipment status parameters, operating parameters, and intermediate product quality parameters.

[0103] In this embodiment of the invention, taking the real-time current value of the crusher motor as an example, the real-time current value of the crusher motor is classified into "current too high", "normal" or "current too low" states, and then used as evidence of the "crusher status" node.

[0104] Using an online image recognition system installed on the conveyor belt, images of the crushed coal flow are analyzed in real time to estimate its particle size distribution (e.g., the proportion of >6mm), and the results are used as evidence for the "crushed particle size" node.

[0105] Specifically, the process for estimating granularity distribution in an online image recognition system is as follows: 1) Image acquisition: An industrial camera and light source are installed above the discharge belt of the crusher to continuously capture images of the falling coal flow.

[0106] 2) Image preprocessing: Denoise the acquired image, enhance contrast, and segment it to separate the coal block from the background.

[0107] 3) Feature extraction: Use computer vision algorithms (such as edge detection and connected component analysis) to identify the outline of each coal block in the image and calculate its equivalent diameter or projected area.

[0108] 4) Particle size distribution calculation: Statistically analyze the size of all identified coal blocks in the current frame or within a certain period, and generate a particle size distribution histogram. For example, calculate the percentage of coal blocks larger than 6mm out of the total number of coal blocks.

[0109] 5) State Classification: Compare the calculated percentage (e.g., >6mm = 20%) with the preset standard (e.g., the standard requires <10%), and classify it into a predefined node state, such as "severely coarse". This classified state serves as evidence input to the sample preparation bias Bayesian model for the "particle size after fragmentation" node.

[0110] It also records the actual reduction ratio as evidence of the "reducer performance" node.

[0111] Specifically, the reduction ratio is calculated as follows: The reduction ratio typically refers to the ratio of the retained sample mass to the total sample mass. For a cutting-type divider, its theoretical reduction ratio is determined by the cutter opening width and speed. The actual reduction ratio is calculated as follows: Actual reduction ratio = Retained sample mass / (Retained sample mass + Discarded sample mass); In a practical system, the mass of retained and discarded samples can be obtained in real time by weighing them using a precision scale integrated on the corresponding belt.

[0112] The specific content that was determined to be evidence is as follows: 1) Compare the calculated actual reduction ratio with the theoretical value or the standard allowable range (e.g., 1 / 8 ± 0.05).

[0113] 2) Based on the magnitude of the deviation, map it to the discrete states of the "divider efficiency" node. For example: If the deviation is within ±0.02, the status is "normal".

[0114] If the deviation is within ±0.05, the status is "slight deviation".

[0115] If the deviation exceeds ±0.05, the status is "severe deviation".

[0116] This defined state is used as an evidence variable and input into the sample preparation bias Bayesian model.

[0117] Step S2043: Load each node and its corresponding conditional probability table through the sample preparation deviation Bayesian model, and transform the real-time current value, equipment status parameters, operating parameters and intermediate product quality parameters after data processing into discrete state evidence that can be identified by the sample preparation deviation Bayesian model.

[0118] It should be noted that discrete state evidence refers to the input information of the model node preset discrete state obtained after the real-time coal sampling process data is converted. For example, the "real-time current value 3.2A" is matched and converted into the "normal" state of the node "equipment load status".

[0119] In this embodiment of the invention, a complete probabilistic reasoning foundation is constructed by loading predefined nodes and corresponding conditional probability tables through a sample preparation deviation Bayesian model. At the same time, the real-time current value, equipment status parameters, operating parameters and intermediate product quality parameters after cleaning and normalization are matched and transformed according to the preset discrete state division standard of each node. The continuous or original parameter data is transformed into discrete state evidence that can be recognized by the sample preparation deviation Bayesian model, ensuring that the data input to the model can be accurately associated with the corresponding nodes.

[0120] Step S2044: The sample preparation deviation Bayesian model is compiled using a preset Bayesian algorithm, and all discrete state evidence is input into the compiled sample preparation deviation Bayesian model for global probability propagation.

[0121] It should be noted that the preset Bayesian algorithm refers to the core algorithm selected in advance for probability calculation and inference of the Bayesian model. In this embodiment, the connection tree algorithm is preferred.

[0122] Compilation refers to the process of preprocessing the directed acyclic graph structure of the model using Bayesian algorithms. The core purpose is to transform the originally complex network structure into a form that is suitable for efficient probability calculation (such as a connection tree).

[0123] Global probability propagation refers to the process of spreading and transmitting discrete state evidence across the entire compiled model.

[0124] In this embodiment of the invention, a preset Bayesian algorithm (such as the connection tree algorithm) is used to compile the sample preparation bias Bayesian model, transforming the directed acyclic graph structure of the model into a tree structure for fitting probability calculation to improve inference efficiency. Then, all discrete state evidence is input into the compiled sample preparation bias Bayesian model. Based on the loaded conditional probability table and the compiled inference structure, the model completes the global probability propagation of evidence among all nodes, realizing the update of the prior probability to the posterior probability of each unknown node.

[0125] Step S2045: Update the probability distribution of unknown nodes in the compiled sample preparation deviation Bayesian model based on the propagation results.

[0126] It should be noted that the propagation result refers to the change data of the probability of all nodes in the model after the discrete state evidence is input into the compiled sample bias Bayesian model and then propagated through global probability.

[0127] Unknown nodes refer to nodes in the sample preparation bias Bayesian model that are not directly covered by real-time discrete state evidence.

[0128] A probability distribution refers to the set of probability values ​​that a node can take in each of its preset discrete states.

[0129] In this embodiment of the invention, based on the global probability propagation results, the probability distribution of unknown nodes in the compiled sample preparation deviation Bayesian model that are not directly represented by discrete state evidence is updated. The prior probabilities of each unknown node, which were originally based on the conditional probability table, are corrected to posterior probabilities that incorporate the influence of real-time discrete state evidence. This makes the updated probability distribution more consistent with the real-time running state of the sample preparation process and accurately reflects the actual state probability of each unknown node.

[0130] Step S2046: Extract the posterior probability distribution of all nodes in the updated sample preparation deviation Bayesian model, and select the posterior probability distribution of the preset target ash deviation nodes.

[0131] It should be noted that the target ash content deviation node refers to the core node in the sample preparation deviation Bayesian model that is predefined and used to characterize the deviation of the ash content index of the final product (or key intermediate product) of coal sample preparation from the preset standard state. It belongs to the category of sub-nodes of the model.

[0132] The posterior probability distribution refers to the set of probability values ​​obtained by a node after global probability propagation and distribution update.

[0133] In this embodiment of the invention, the posterior probability distribution of all nodes that have been corrected is extracted from the updated sample preparation deviation Bayesian model. Then, the posterior probability distribution corresponding to the preset target ash deviation node is accurately located and screened according to the preset screening rules. This screening process can be directly associated with the predefined target node identifier in the model to ensure the accuracy of the screening results.

[0134] Step S2047: Extract the posterior probability value corresponding to the state of each node from the posterior probability distribution of the target ash deviation node and select the maximum posterior probability value.

[0135] It should be noted that the posterior probability value refers to a specific numerical value in the posterior probability distribution. Each value uniquely corresponds to a discrete state of the target ash deviation node, representing the probability that the node is in the corresponding state under the current sample preparation conditions.

[0136] The maximum posterior probability value refers to the largest value selected from the posterior probability values ​​corresponding to all states of the target gray deviation node.

[0137] In this embodiment of the invention, from the posterior probability distribution of the target ash deviation node, the posterior probability values ​​corresponding to each preset discrete state of the node (such as "ash deviation qualified", "ash slightly exceeding the standard", "ash seriously exceeding the standard") are extracted, and then all the extracted posterior probability values ​​are compared and sorted. The maximum posterior probability value with the largest value is selected, and the node state corresponding to this value can be used as the most likely state of ash deviation in the current sample preparation process.

[0138] Step S2048: Based on the maximum a posteriori probability value, determine the optimal state of the target ash deviation node and the deviation evaluation information.

[0139] It should be noted that the optimal state refers to the discrete state of the target gray deviation node corresponding to the maximum a posteriori probability value.

[0140] Deviation assessment information refers to a comprehensive assessment conclusion generated based on the optimal state of the target ash deviation node. The core content includes the judgment result of whether the ash deviation is qualified, the severity level of the deviation, and the key influencing factors (such as tool wear, abnormal feeding speed, etc.) extracted from the model that lead to the deviation state.

[0141] In this embodiment of the invention, based on the maximum posterior probability value, the discrete state of the target ash deviation node is located and the state is determined as the optimal state. At the same time, combined with the specific attributes of the optimal state (such as whether it exceeds the standard and the degree of exceeding the standard) and the posterior probability changes of each deviation-affected node in the model, deviation assessment information including the ash deviation state determination result, the deviation severity level and key influencing factor prompts are generated.

[0142] Step S2049: Compare the prior probability value and posterior probability value of each node, and determine the node corresponding to the key deviation influencing factor based on the comparison result.

[0143] It should be noted that the nodes of key deviation influencing factors refer to the deviation influencing factor nodes (parent nodes) whose probability difference exceeds the set threshold. The corresponding deviation influencing factors are the core causes that lead to the current target ash deviation state.

[0144] In this embodiment of the invention, the prior probability value and posterior probability value of each node (focusing on the parent node of the deviation influencing factor) are compared one by one, the probability difference between the two is calculated and a reasonable probability change threshold is set. If the probability difference of a certain node exceeds the threshold, it indicates that the deviation influencing factor corresponding to the node has a significant impact on the state of the target ash content deviation node under the current sample preparation conditions. Based on this, the node is determined to be the node corresponding to the key deviation influencing factor.

[0145] Step S20410: Calculate the difference between the prior probability value and the posterior probability value of each node to obtain the contribution of each node.

[0146] It should be noted that contribution refers to the absolute difference between the prior probability and the posterior probability.

[0147] In this embodiment of the invention, the absolute difference between the prior probability value and the posterior probability value of each node (focusing on the parent node of the deviation influencing factor) under the same state is calculated, and this absolute difference is taken as the contribution of each node. The magnitude of the contribution value directly reflects the degree of influence of the deviation influencing factor represented by the corresponding node on the optimal state of the target ash deviation node.

[0148] Step S20411: Using the deviation assessment information of the target ash deviation node, the nodes corresponding to the key deviation influencing factors, and the contribution of each node, the assessment result is generated.

[0149] In this embodiment of the invention, the deviation assessment information of the target ash deviation node, the nodes corresponding to the key deviation influencing factors, and the contribution of each node are used to link and integrate the three to generate an assessment result that includes the core judgment conclusion of ash deviation (such as whether it is qualified or not, the degree of exceeding the standard), a list of key deviation influencing factors, and the ranking of the contribution of each factor.

[0150] Specifically, the processed real-time coal sampling process data is input into a pre-constructed Bayesian model of sampling deviation, and a Bayesian inference algorithm (such as the connection tree algorithm) is run. The algorithm updates the probability distribution of all unknown nodes in the network. In particular, it can calculate the posterior probability of the target ash deviation node being in various deviation magnitude states given all observed process evidence. Simultaneously, it can calculate the posterior probability of each deviation factor node (such as "tool wear" or "reducer offset"), thereby identifying which factors are "more likely" to be the main causes of the overall deviation. Specifically, the core of probabilistic inference is Bayes' theorem, whose goal is to calculate the posterior probability distribution of other nodes (query variable Q, such as "target ash deviation" or "tool wear") given the observations of certain nodes (evidence variable E, such as real-time acquired data).

[0151] (1) The core formula is as follows: Posterior probability: P(Q|E) = [P(E|Q) × P(Q)] / P(E) In the formula, P(Q) represents the prior probability of the query variable (the probability in the absence of any evidence, usually derived from historical statistics); P(E|Q) represents the likelihood probability, which is the probability of observing the current evidence when the query variable is in a certain state, and is determined by the CPT in the network; P(E) represents the marginal probability of the evidence, which is a normalized constant; and P(Q|E) represents the posterior probability, which is the probability of the query variable being in each state after observing the current evidence.

[0152] (2) The specific inference algorithm is as follows: For complex Bayesian networks, an efficient algorithm, such as the connection tree algorithm, is required. The process is briefly described below: 1) Compile the network: Transform the Bayesian network into a data structure called a "connection tree" or "cluster tree", where each tree node contains a set of network variables and their joint probability table.

[0153] 2) Evidence injection: The observed evidence (such as "crusher current = too high" or "particle size after crushing = too coarse") is converted into probabilistic information and injected into the corresponding clustering nodes in the connection tree.

[0154] 3) Global Propagation: The algorithm performs a series of message passing operations between all nodes in the connection tree until the entire network reaches a new state with consistent probabilities. This process effectively propagates the influence of local evidence to the entire network.

[0155] 4) Result extraction: After inference is completed, the posterior probability distribution of any node can be directly queried from the connection tree.

[0156] (3) The specific tracing process is as follows: Maximum A posteriori (MAP) estimation: Used to assess the total deviation. For example, when querying the "target ash deviation" node, the algorithm will provide the posterior probability of each deviation state. The state with the highest probability (e.g., "+0.5%~+0.8%" has a probability of 90%) is taken as the most likely deviation assessment.

[0157] Posterior probability comparison: Used to identify key sources of deviation. Compare the prior probability of key deviation factor nodes (such as "tool wear" or "divider offset") before evidence input with the posterior probability after evidence input.

[0158] Contribution calculation: The contribution of a certain source of bias can be defined as contribution = posterior probability - prior probability, or it can be sorted directly based on the absolute value of the posterior probability.

[0159] Source identification: For example, if the posterior probability of the "secondary crusher tool wear" node increases sharply from the usual 10% (prior) to 95%, while other nodes remain relatively unchanged, it can be determined that this is the main contributor to the deviation. The system then generates a ranking list of deviation sources based on this.

[0160] (4) The evaluation results include: Total deviation estimate: The target ash / calorific value results may have a deviation of ±X% (based on the maximum a posteriori probability estimate).

[0161] Deviation source ranking: List the top 3 most likely deviation sources and their probabilities in descending order of posterior probability. For example: "1. Wear of secondary crusher blades (probability 75%); 2. Speed ​​fluctuation of primary divider (probability 60%); 3. Unstable drying temperature (probability 45%)".

[0162] Uncertainty assessment: Provide the confidence interval for the predicted deviation.

[0163] Step S205: Send the evaluation results to the operation interface or centralized control system to generate an inspection work order that matches the evaluation results.

[0164] It should be noted that the operating interface refers to the visual interactive terminal interface (such as an industrial control screen or on-site operating terminal) deployed at the sample preparation site.

[0165] A centralized control system refers to a remote control system used for centralized monitoring and management of multiple sample preparation stages or the entire sample preparation production line.

[0166] Inspection work orders refer to standardized work documents automatically generated by the system based on evaluation results.

[0167] In this embodiment of the invention, the evaluation results are pushed to the operating interface or centralized control system in real time. For example, if the system infers that the wear probability of the secondary crusher blades is extremely high, it can automatically generate an inspection work order that "suggests checking and replacing the secondary crusher blades".

[0168] Step S206: Based on the inspection work order, perform predictive maintenance operations.

[0169] It should be noted that predictive maintenance refers to preventative maintenance activities carried out in advance based on the equipment's operating status, historical maintenance data, and potential hazards indicated by inspection work orders.

[0170] In this embodiment of the invention, predictive maintenance is achieved by following the inspection work order that "suggests checking and replacing the secondary crusher cutters".

[0171] This embodiment also provides a device for tracing and quantifying deviations in coal sample preparation. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0172] This embodiment provides a device for tracing and quantifying deviations in the coal sample preparation process, such as... Figure 4 As shown, it includes: Input module 301 is used to input coal sample preparation process data of preset coal field operation procedures into preset sample preparation process model, and output multiple sample preparation process nodes and deviation influencing factors of each sample preparation process node. Module 302 is used to construct a directed acyclic graph by setting the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node. The allocation module 303 is used to assign conditional probability tables to each node of the directed acyclic graph and to construct a sample preparation bias Bayesian model based on all nodes and their corresponding conditional probability tables. The evaluation module 304 is used to input real-time coal sample preparation process data into the sample preparation deviation Bayesian model, evaluate the real-time coal sample preparation process through the sample preparation deviation Bayesian model, and obtain the evaluation results.

[0173] In some alternative implementations, the input module 301 includes: Acquisition unit, used to acquire coal field operation procedures; The input unit is used to input coal sample preparation process data from the coal field operation procedure into the preset sample preparation process model. The analysis unit is used to analyze the entire coal sample preparation process data through the sample preparation process model, and obtain a series of continuous sample preparation process nodes. The deviation unit is used to analyze the factors affecting deviations from the ideal state at each sample preparation process node.

[0174] In some alternative implementations, the construction module 302 includes: Set up node units to set all deviation influencing factors of each sample preparation process node as parent nodes, and set each sample preparation process node as child nodes; A connection unit is used to connect a parent node and its corresponding child nodes using directed arrows to obtain a directed acyclic graph.

[0175] In some alternative implementations, the allocation module 303 includes: Define the unit, used to define the node state of each node in a directed acyclic graph; The building unit is used to construct a conditional probability table for each node based on the combination of node states of each parent node and the probability distribution of node states of the corresponding child nodes. The parameter determination unit is used to determine the conditional probability parameters in the conditional probability table for each node based on historical operating data, equipment design tolerances, experimental test data, and expert experience. Construct model units to build a sample preparation bias Bayesian model using each node and its corresponding conditional probability table.

[0176] In some alternative implementations, the evaluation module 304 includes: The data acquisition unit is used to acquire real-time current values, equipment status parameters, operating parameters, and intermediate product quality parameters during the real-time coal sampling process. The input model unit is used to process real-time current values, equipment status parameters, operating parameters, and intermediate product quality parameters and input them into the sample preparation deviation Bayesian model. The processing unit is used to load each node and its corresponding conditional probability table through the sample preparation deviation Bayesian model, and to convert the real-time current value, equipment status parameters, operating parameters and intermediate product quality parameters after data processing into discrete state evidence that can be identified by the sample preparation deviation Bayesian model. The propagation unit is used to compile the sample preparation deviation Bayesian model using a preset Bayesian algorithm, and input all discrete state evidence into the compiled sample preparation deviation Bayesian model for global probability propagation. The update unit is used to update the probability distribution of unknown nodes in the compiled sample preparation bias Bayesian model based on the propagation results. The filtering unit is used to extract the posterior probability distribution of all nodes in the updated sample preparation deviation Bayesian model and filter out the posterior probability distribution of the preset target ash deviation nodes. The selection unit is used to extract the posterior probability value corresponding to each node state from the posterior probability distribution of the target ash deviation node and select the maximum posterior probability value. The optimal state unit is used to determine the optimal state of the target ash deviation node and the deviation evaluation information based on the maximum a posteriori probability value. The comparison unit is used to compare the prior probability value and the posterior probability value of each node, and determine the node corresponding to the key deviation influencing factor based on the comparison result; The calculation unit is used to calculate the difference between the prior probability value and the posterior probability value of each node to obtain the contribution of each node. The evaluation unit is used to generate evaluation results by using the deviation evaluation information of the target ash deviation node, the nodes corresponding to the key deviation influencing factors, and the contribution of each node.

[0177] In some alternative embodiments, the device includes: The sending unit is used to send the evaluation results to the operation interface or centralized control system and generate inspection work orders that match the evaluation results; The execution unit is used to perform predictive maintenance operations based on inspection work orders.

[0178] The coal sampling process deviation tracing and quantitative evaluation device provided in this embodiment of the invention can execute the coal sampling process deviation tracing and quantitative evaluation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0179] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0180] The following is a detailed reference. Figure 5 This diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 401, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 402 or a program loaded from memory 408 into random access memory (RAM) 403. RAM 403 also stores various programs and data required for the operation of the electronic device. The processor 401, ROM 402, and RAM 403 are interconnected via bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0181] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0182] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a memory 408, or installed from a ROM 402. When the computer program is executed by the processor 401, it performs the functions defined in the coal sampling process deviation tracing and quantitative evaluation method of the embodiments of the present invention.

[0183] Figure 5 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0184] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the method for tracing and quantifying deviations in the coal sampling process shown in the above embodiments is implemented.

[0185] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0186] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for tracing and quantitatively evaluating deviations in coal sample preparation, characterized in that, The method includes: Input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node. A directed acyclic graph is constructed by setting the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node. A conditional probability table is assigned to each node of the directed acyclic graph, and a sample preparation bias Bayesian model is constructed based on all the nodes and their corresponding conditional probability tables. The real-time coal sample preparation process data is input into the sample preparation deviation Bayesian model, and the real-time coal sample preparation process is evaluated through the sample preparation deviation Bayesian model to obtain the evaluation results.

2. The method according to claim 1, characterized in that, The process involves inputting coal sampling process data from a pre-defined on-site coal operation procedure into a pre-defined sampling process model, and outputting multiple sampling process nodes and the deviation influencing factors for each sampling process node, including: Obtain on-site coal operation procedures; Input the coal sample preparation process data from the coal field operation procedure into the preset sample preparation process model. By analyzing the entire coal sample preparation process data using the aforementioned sample preparation process model, a series of continuous sample preparation process nodes are obtained. The factors influencing deviations from the ideal state at each of the sample preparation process steps were analyzed.

3. The method according to claim 1, characterized in that, The construction of a directed acyclic graph, with the deviation influencing factors of each sample preparation process node as the parent node and each coal sample preparation process node as the child node, includes: Set all deviation influencing factors of each sample preparation process node as parent nodes, and set each sample preparation process node as child nodes; A directed acyclic graph is obtained by connecting the parent node and its corresponding child nodes with directed arrows.

4. The method according to claim 1, characterized in that, The step of assigning conditional probability tables to each node of the directed acyclic graph and constructing a sample preparation bias Bayesian model based on all the nodes and their corresponding conditional probability tables includes: Define the node state of each node in the directed acyclic graph; Based on the node state combinations of each parent node and the node state probability distribution of the corresponding child nodes, a conditional probability table for each node is constructed. Based on historical operating data, equipment design tolerances, experimental test data, and expert experience, the conditional probability parameters in the conditional probability table for each node are determined. A sample preparation bias Bayesian model is constructed using each node and its corresponding conditional probability table.

5. The method according to claim 1, characterized in that, The process involves inputting real-time coal sampling process data into the sampling deviation Bayesian model, evaluating the real-time coal sampling process using the sampling deviation Bayesian model, and obtaining evaluation results, including: Acquire real-time current values, equipment status parameters, operating parameters, and intermediate product quality parameters during the real-time coal sampling process; The real-time current value, the equipment status parameters, the operating parameters, and the intermediate product quality parameters are processed and input into the sample preparation deviation Bayesian model. The sample preparation deviation Bayesian model loads each node and its corresponding conditional probability table, and transforms the real-time current value, equipment status parameters, operating parameters and intermediate product quality parameters after data processing into discrete state evidence that can be identified by the sample preparation deviation Bayesian model. The sample preparation deviation Bayesian model is compiled using a preset Bayesian algorithm, and all the discrete state evidence is input into the compiled sample preparation deviation Bayesian model for global probability propagation. The probability distribution of unknown nodes in the compiled sample preparation bias Bayesian model is updated based on the propagation results. Extract the posterior probability distribution of all nodes in the updated sample preparation deviation Bayesian model, and then select the posterior probability distribution of the preset target ash content deviation nodes. Extract the posterior probability value corresponding to the state of each node from the posterior probability distribution of the target ash deviation node and select the maximum posterior probability value; Based on the maximum posterior probability value, the optimal state of the target ash deviation node and the deviation evaluation information are determined; The prior probability value and posterior probability value of each node are compared, and the node corresponding to the key deviation influencing factor is determined based on the comparison result. The difference between the prior probability value and the posterior probability value of each node is calculated to obtain the contribution of each node. An evaluation result is generated by using the deviation assessment information of the target ash deviation node, the nodes corresponding to the key deviation influencing factors, and the contribution of each node.

6. The method according to claim 1, characterized in that, The method further includes: The evaluation results are sent to the operation interface or centralized control system to generate an inspection work order that matches the evaluation results; Based on the inspection work order, predictive maintenance operations are performed.

7. A device for tracing and quantifying deviations in coal sample preparation, characterized in that, The device includes: The input module is used to input the coal sample preparation process data of the preset coal field operation procedure into the preset sample preparation process model, and output multiple sample preparation process nodes and the deviation influencing factors of each sample preparation process node. The construction module is used to construct a directed acyclic graph by setting the deviation influencing factors of each of the sample preparation process nodes as parent nodes and each of the coal sample preparation process nodes as child nodes. The allocation module is used to assign conditional probability tables to each node of the directed acyclic graph, and to construct a sample preparation bias Bayesian model based on all the nodes and the corresponding conditional probability tables. The evaluation module is used to input real-time coal sampling process data into the sampling deviation Bayesian model, evaluate the real-time coal sampling process through the sampling deviation Bayesian model, and obtain the evaluation result.

8. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the method for tracing and quantifying deviations in the coal sample preparation process as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for tracing and quantifying deviations in the coal sample preparation process as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes computer instructions for causing a computer to execute the method for tracing and quantifying deviations in the coal sample preparation process as described in any one of claims 1 to 6.

Images