Method and system for realizing process digital twin based on dynamic quantity analysis

Abstract

The application provides a process factory digital twin implementation method and system based on dynamic quantity analysis, relates to the technical field of process factory digital management, and first acquires process factory process link characteristic data and initial material input data, constructs a process flow model reflecting process link connection and material transmission relationship; theoretical quantity data is obtained through simulation calculation based on the process flow model and the initial data, and a material attribute transmission attenuation coefficient is introduced; real-time quantity data of key process links is collected, compared and analyzed with the theoretical data after multi-device data fusion processing, and quantity deviation analysis results containing deviation propagation path information are obtained; process parameter adjustment instructions containing adjustment effect prediction information are generated according to the quantity deviation analysis results and sent to a control terminal, so that dynamic monitoring, accurate analysis and intelligent control of the process factory are realized.

Description

Technical Field

[0001] This invention relates to the field of digital management technology for process plants, and more specifically, to a method and system for implementing digital twins in process plants based on dynamic data quality analysis. Background Technology

[0002] In the production and operation of process plants, achieving efficient, stable, and high-quality production is the core objective. Traditional process plant management and control methods primarily rely on manual experience and periodic testing and analysis. On the one hand, the understanding and optimization of the process flow are often based on the intuitive understanding of experienced engineers regarding each process step, lacking a systematic and precise model to describe the complex relationships between each process step and the changing patterns of materials. This makes it difficult to make quick and accurate adjustment decisions when facing dynamic changes in the production process. On the other hand, in terms of data collection and analysis, traditional methods typically use single devices for data acquisition, resulting in limited data volume and susceptibility to equipment inaccuracies and malfunctions. Furthermore, the lack of effective data fusion mechanisms means that the collected data cannot comprehensively and accurately reflect the actual production situation. Moreover, the collected data is often only subjected to simple comparative analysis, failing to delve into the underlying deviation propagation paths and fundamentally identify the key factors affecting production quantity and quality. Consequently, precise adjustment of process parameters is impossible, limiting the improvement of process plant production efficiency and product quality. Summary of the Invention

[0003] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for implementing a digital twin of a process plant based on dynamic numerical quality analysis, the method comprising:

[0004] Acquire process characteristic data and initial material input data of the process plant. The process characteristic data includes the regularity information of each process to realize the material handling function. The initial material input data includes the basic attribute information of the material.

[0005] A process flow model is constructed based on the process characteristic data and initial material input data. The process flow model is used to reflect the connection relationship between each process step and the material transfer relationship between each process step. The process flow model can receive material input data and output the material processing result data of the corresponding process step.

[0006] Based on the process flow model and the initial material input data, process flow simulation calculation is performed to obtain theoretical quantity and quality data for each process step. The theoretical quantity and quality data includes the quantity and quality information of the output materials in each process step. The process flow simulation calculation introduces a material attribute transmission attenuation coefficient.

[0007] Real-time numerical and quality data of each key process step in the process plant are collected. After processing the real-time numerical and quality data using a multi-device data fusion mechanism, the processed real-time numerical and quality data is compared and analyzed with the theoretical numerical and quality data of the corresponding process step to obtain numerical and quality deviation analysis results. The numerical and quality deviation analysis results include deviation propagation path information.

[0008] Based on the numerical deviation analysis results, process parameter adjustment instructions are generated and sent to the control terminal of the process plant to adjust the operating parameters of the corresponding process steps. The process parameter adjustment instructions include adjustment effect prediction information.

[0009] In another aspect, embodiments of the present invention also provide a process plant digital twin implementation system based on dynamic data quality analysis, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0010] Based on the above, this invention constructs a process flow model by acquiring characteristic data of process links and initial material input data. This model reflects the connection and material transfer relationships between each process link. Based on this process flow model, process flow simulation calculations are performed, and a material attribute transfer attenuation coefficient is introduced to obtain theoretical quantitative and qualitative data for each process link, making the simulation results closer to actual production conditions. Real-time quantitative and qualitative data of key process links are collected and processed using a multi-device data fusion mechanism. The processed real-time data is compared and analyzed with theoretical data to obtain quantitative and qualitative deviation analysis results containing deviation propagation path information. This allows for in-depth analysis of the root causes of problems in the production process. Based on the analysis results, process parameter adjustment instructions containing prediction information of adjustment effects are generated and sent to the control terminal. This achieves precise adjustment of process link operating parameters, enabling dynamic monitoring, precise analysis, and intelligent control of the process plant production process, effectively improving production efficiency, product quality, and production stability. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the execution flow of the process factory digital twin implementation method based on dynamic numerical quality analysis provided in the embodiments of the present invention.

[0012] Figure 2 This is a schematic diagram of exemplary hardware and software components of a process plant digital twin implementation system based on dynamic numerical quality analysis provided in an embodiment of the present invention. Detailed Implementation

[0013] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1This is a flowchart illustrating a method for implementing a digital twin of a process factory based on dynamic number quality analysis, according to an embodiment of the present invention. The following is a detailed description of this method for implementing a digital twin of a process factory based on dynamic number quality analysis.

[0014] Step S110: Obtain the process characteristics data and initial material input data of the process plant. The process characteristics data includes the regularity information of each process to realize the material handling function, and the initial material input data includes the basic attribute information of the material.

[0015] In coal washing and processing plants, acquiring process characteristic data needs to cover all key equipment combinations involved in material handling. For the raw coal crushing stage, it is necessary to collect the variation patterns of material particle size distribution under different crushing tooth shapes and rotational speeds, including the correlation data between the proportion of each particle size range in the crushed material and crushing time and feed rate.

[0016] For the heavy media separation process, it is necessary to obtain the corresponding relationship between separation density, media circulation volume and the yield and ash content of clean coal, middlings, gangue, etc., covering the separation efficiency variation law under different raw coal properties, as well as the impact data of the magnetic content of the media on the separation effect.

[0017] The characteristic data of the flotation process include the relationship between the types and dosages of flotation reagents, pulp concentration, aeration rate and flotation clean coal yield and ash content, as well as the influence of frothing speed on flotation effect. The above data were compiled by simulating the process operation under different working conditions in the laboratory and combining it with long-term on-site production records to form a structured process characteristic dataset.

[0018] The initial material input data focuses on raw coal and includes industrial analysis data such as ash content, volatile matter, and fixed carbon content, as well as elemental analysis data on carbon, hydrogen, oxygen, nitrogen, and sulfur content. It also includes physical property data such as true density, apparent density, hardness, grindability index, and particle size distribution data, i.e., the mass percentage of each particle size fraction. This data is obtained by sampling raw coal and then measuring it using standard coal testing methods. For example, ash content is determined using the slow ashing method, volatile matter is determined using a muffle furnace with air isolation heating, and particle size distribution is obtained through sieve analysis, forming a complete initial material input data archive.

[0019] Step S120: Construct a process flow model based on the process characteristic data and initial material input data. The process flow model is used to reflect the connection relationship between each process step and the transfer relationship of materials between each process step. The process flow model can receive material input data and output the material processing result data of the corresponding process step.

[0020] Step S121: Classify and organize the characteristic data of the process links, and extract the material processing rule information of each process link. The material processing rule information includes the change rules of the material input being processed by the process links and transformed into the output material.

[0021] Step S1211: Classify the process characteristic data according to the functional type of the process, including material mixing function type, material separation function type, and material transportation function type.

[0022] In coal washing and processing plants, when classifying the characteristic data of process steps by functional type, the material mixing function type covers process steps such as flotation machines and agitated tanks. Flotation machines use impellers to thoroughly mix slurry, reagents, and air; their characteristic data includes the relationship between stirring intensity, stirring time, and mixing uniformity. Agitated tanks are used for pre-mixing reagents and slurry; their characteristic data includes information on the correlation between stirring speed and reagent dispersion effect.

[0023] Material separation functions include heavy media separators, jigs, screening equipment, and filters. Heavy media separators separate materials of different compositions based on density differences; their characteristic data relates to the relationship between separation density and separation efficiency. Jigs achieve stratified separation of materials by density through water flow pulsation; their characteristic data includes the relationship between pulsation frequency, amplitude, and separation effect. Screening equipment separates materials based on particle size differences; its characteristic data covers the correlation between screen aperture, vibration frequency, and screening efficiency. Filters achieve solid-liquid separation through filter media; their characteristic data includes the relationship between filtration pressure, filter cloth type, and filter cake moisture content.

[0024] Material handling functions include process steps such as belt conveyors, scraper conveyors, and bucket elevators. The characteristic data for belt conveyors includes the relationship between conveyor belt speed, width, and conveying capacity; the characteristic data for scraper conveyors involves the correlation between scraper chain speed and conveying capacity; and the characteristic data for bucket elevators includes the correspondence between lifting speed and lifting capacity. This classification method makes the structure of the characteristic data for each process step clearer, facilitating the subsequent extraction of material handling patterns.

[0025] Step S1212: For the process characteristic data under each functional type, extract the information describing the correspondence between material input attributes, processing parameters and output attributes. The input attributes include the composition and state information of the material, the processing parameters include the operating parameters during the process, and the output attributes include the composition and state information of the processed material.

[0026] For flotation machines in the material mixing function type, the correspondence between input attributes, processing parameters, and output attributes is extracted from their characteristic data. Input attributes include the solid content in the slurry, the proportion of each mineral component, and the slurry temperature; processing parameters include the flotation machine impeller speed, reagent dosage, and aeration rate; output attributes include the uniformity of the mixed slurry and the dispersion concentration of reagents in the slurry.

[0027] For the heavy media separator in the material separation function type, the input attributes are the ash content, content of each density grade of the feed material, and moisture content; the processing parameters include the separation density, media circulation volume, and drum speed; the output attributes are the ash content, yield, and moisture content of clean coal, middlings, and gangue.

[0028] For belt conveyors in the material handling function type, the input attributes are the particle size distribution and bulk density of the conveyed material; the processing parameters include the conveyor belt speed and the conveyor belt inclination angle; the output attributes are the conveying volume per unit time and the uniformity of material distribution on the conveyor belt. This information is obtained by extracting these corresponding relationships.

[0029] Step S1213: Optimize the correspondence information between the extracted input attributes, processing parameters and output attributes, remove duplicate and contradictory information, and form a preliminary material processing relationship dataset.

[0030] In the processing of characteristic data for process steps in coal washing and beneficiation plants, the extracted correspondence information needs to be optimized. For duplicate information, such as in the characteristic data of flotation machines where the same set of input attributes and processing parameters correspond to multiple sets of identical output attribute records, only one record should be retained to reduce data redundancy.

[0031] For contradictory information, such as in the characteristic data of a heavy media separator where the same input attributes and processing parameters correspond to two sets of significantly different output attributes, verification is required. This involves reviewing the original experimental records or production logs to confirm the accuracy of the data. If the error is due to a record error, the data should be corrected. If the error is caused by fluctuations in actual operating conditions, both sets of data should be retained, and the cause of the fluctuation should be noted.

[0032] For cases of missing data, if the amount of missing data is small, reasonable interpolation can be used to supplement it based on the changing trends of adjacent data. If the amount of missing data is large, the missing data should be marked as missing and considered in subsequent analysis. After the above optimization process, a preliminary material handling relationship dataset is formed, ensuring the accuracy and consistency of the data.

[0033] Step S1214: Based on the preliminary material processing relationship dataset, through data fitting and rule induction, construct the change rules of material input after being processed by the process steps to transform into output material. The change rules are used to clarify the output attributes corresponding to the combination of input attributes and processing parameters.

[0034] Based on a preliminary data set of material handling relationships, taking the heavy media separation process as an example, a data fitting method is used to process the data between input attributes (such as feed ash content and particle size distribution), process parameters (such as separation density and media circulation volume), and output attributes (such as clean coal yield and ash content). By drawing scatter plots to observe the data distribution trend, a suitable fitting curve type is selected to establish the mathematical relationship between input and output.

[0035] Simultaneously, rules were summarized and analyzed to determine the variation patterns of clean coal yield and ash content when the sorting density changes within a certain range. For example, when the sorting density increases while other parameters remain constant, the clean coal yield shows an upward trend, and the clean coal ash content also increases accordingly; when the media circulation volume increases to a certain extent, the improvement in sorting efficiency tends to level off.

[0036] The above patterns are fixed in the form of textual descriptions to form change rules, which clarify the specific changes in output attributes under specific combinations of input attributes and processing parameters, such as "when the ash content of the feed is within a certain range, the sorting density is set to a certain value, and the media circulation volume is at a certain level, the clean coal yield is within a certain range, and the ash content is within a certain range."

[0037] Step S1215: Set a dynamic correction factor for the change rule. The dynamic correction factor can be adjusted according to the running time of the process and maintenance records.

[0038] A dynamic correction factor is set for the changing rules of the heavy media sorting process. This dynamic correction factor is initially set to a baseline value, and gradually increases as the heavy media separator's operating time increases. For example, for every certain increase in operating time, the correction factor increases proportionally from the baseline value to reflect the decrease in sorting efficiency caused by equipment wear. For instance, after the roller liner wears down, the deviation between the actual and set values ​​of the sorting density increases, and the calculated results of the output attributes are adjusted through the dynamic correction factor.

[0039] When the heavy media separator undergoes maintenance, such as replacing the drum liner or adjusting the transmission device, the dynamic correction factor is reset to the baseline value according to the maintenance record to reflect the recovery of equipment performance. For the changing rules of the flotation machine, its dynamic correction factor is adjusted according to the impeller's running time. After the impeller wears down, the stirring intensity decreases, and the correction factor is adjusted accordingly to correct the calculated mixing effect. When the impeller is replaced, the correction factor returns to its initial value.

[0040] Step S1216: Associate the change rules with the identification information of the corresponding process steps to form material handling rule information for each process step, and at the same time, perform format standardization processing on the material handling rule information.

[0041] The rules governing changes in the heavy media sorting process are linked to their identification information, which includes the equipment number, equipment name, and location within the workshop. For example, the "rules governing changes for heavy media sorter M-01" are associated with the equipment number "M-01", the name "heavy media drum sorter", and the location "sorting workshop area A", ensuring that each rule accurately corresponds to a specific process step.

[0042] Material handling pattern information is standardized in format, adopting a unified structured data format. For example, each pattern information entry includes input attribute fields, processing parameter fields, output attribute fields, dynamic correction factor fields, and process step identifier fields. Each field has a consistent naming convention and clearly defined data type. For instance, ash content in the input attribute field is expressed as a percentage, sorting density in the processing parameter field is expressed as grams per cubic centimeter, and yield in the output attribute field is expressed as a percentage. This ensures consistent formatting of material handling pattern information across different process steps, facilitating unified access and processing by the process flow model.

[0043] Step S122: Based on the actual production process of the process plant, determine the material transfer sequence between each process step and form a process step transfer relationship list. The process step transfer relationship list records the correspondence between the input ports and output ports of adjacent process steps.

[0044] In a coal washing and processing plant, the material transfer sequence of each process step is analyzed based on the actual production flow. Raw coal first enters the raw coal bunker and is then conveyed to the crushing stage by a coal feeder. The crushed material is then conveyed to the screening stage by a belt conveyor. The oversize material from the screening stage is returned to the crushing stage via another belt conveyor, while the undersize material is conveyed to the heavy media separation stage by a belt conveyor.

[0045] The clean coal product from the heavy media separation stage is conveyed to the dewatering stage via a scraper conveyor, while middlings and gangue are conveyed to their respective subsequent processing stages via belt conveyors. The clean coal processed in the dewatering stage is conveyed to the clean coal silo via a belt conveyor, and the filtrate produced during dewatering is transported to the circulating water tank via pipeline.

[0046] Following the above transmission sequence, a list of process link transmission relationships is created. The list details the correspondence between the input and output ports of adjacent process links. For example, the output port "Crushed Material Outlet" of the crushing link corresponds to the input port "Screwing Inlet" of the screening link; the output port "Oversize Outlet" of the screening link corresponds to the input port "Return Inlet" of the crushing link; the output port "Undersize Outlet" of the screening link corresponds to the input port "Sorting Inlet" of the heavy media separation link, and so on. Each port is labeled with its type, such as solid material outlet, liquid material outlet, solid material inlet, etc., to ensure the accuracy of material transmission relationships.

[0047] Step S123: Call the process flow modeling tool, drag and drop the graphic symbols corresponding to each process step according to the process step transfer relationship list, establish the connecting lines between the graphic symbols of each process step to represent the material transfer path, and initially construct the process flow model framework.

[0048] A specialized process flow modeling tool is invoked. The tool's graphic library contains graphic symbols corresponding to each process step, such as the crushing step corresponding to the crusher graphic symbol, the screening step corresponding to the screening machine graphic symbol, and the heavy media separation step corresponding to the separator graphic symbol, etc.

[0049] Based on the list of process flow relationships, drag and drop the graphic symbols of each process step onto the canvas of the modeling tool, arranging them according to the spatial layout and flow sequence in actual production. For example, place the graphic symbol of the raw coal bunker on the left, and place the graphic symbols of the coal feeder and crushing process on its right, then place the graphic symbol of the screening process to the right of the crushing process graphic symbol, and finally place the graphic symbol of the heavy media separation process to the right of the screening process graphic symbol, and so on.

[0050] Next, using the line drawing function in the modeling tool, connecting lines are created between the graphic symbols of adjacent process stages. The starting point of the connecting line is the graphic symbol corresponding to the output port of the previous process stage, and the ending point is the graphic symbol corresponding to the input port of the next process stage. For example, a line is drawn from the "Crushed Material Outlet" position of the crushing stage graphic symbol to the "Screening Inlet" position of the screening stage graphic symbol, thus representing the path of material transfer from the crushing stage to the screening stage. Through the above operations, a preliminary process flow model framework is constructed, intuitively displaying the connection relationships of each process stage and the material transfer path.

[0051] Step S124: Encapsulate the material handling rules information of each process step into the corresponding process step graphic symbols, and set a parameter configuration interface for the process flow model framework. The parameter configuration interface can receive externally input process parameter adjustment data.

[0052] In the process flow modeling tool, select the graphic symbol for each process step. Using the tool's attribute configuration function, encapsulate the previously compiled material handling rules information for each process step into the graphic symbol. During the encapsulation process, establish a mapping relationship between the graphic symbol and the material handling rules information, so that when the graphic symbol is triggered, the corresponding handling rules information can be invoked for calculation.

[0053] A parameter configuration interface is set up for the process flow model framework. This interface includes multiple parameter input items, corresponding to the processing parameters of each process step. For example, input items such as separation density and media circulation volume are set for the graphic symbols of the heavy media separation step, and input items such as reagent dosage and pulp concentration are set for the graphic symbols of the flotation step.

[0054] The parameter configuration interface supports the input of external data. Operators or other systems can use this interface to input process parameters to adjust data, such as adjusting the sorting density value of the heavy medium sorting stage. The adjusted parameters will be transmitted to the corresponding process stage graphic symbol to update the processing parameters in the material handling law information, thereby affecting the model's calculation results.

[0055] Step S125: Establish a dynamic update interface for process characteristics. The dynamic update interface for process characteristics can receive real-time collected process operation status data and update the material handling rules information of the corresponding process based on the process operation status data.

[0056] The dynamic update interface for process characteristics is connected to the factory's real-time data acquisition system, enabling it to receive real-time operational status data for each process stage. This operational status data includes equipment operating current, voltage, temperature, vibration frequency, motor speed, etc.

[0057] For example, real-time operating status data for heavy media separators includes the actual rotational speed of the drum, bearing temperature, and motor current; real-time operating status data for flotation machines includes impeller speed, aeration rate, and slurry level. When this operating status data is transmitted to the dynamic update interface, the interface determines its impact on material processing patterns based on preset association rules.

[0058] If the actual rotational speed of the drum in the heavy media separator deviates from the set rotational speed, the dynamic update interface will adjust the correction coefficient corresponding to the rotational speed parameter in the material handling pattern information of this process step according to the deviation value, and then update the output attribute calculation logic related to rotational speed in the change rules. If the bearing temperature of the flotation machine exceeds the normal range, the interface will trigger the adjustment of the dynamic correction factor in the material handling pattern information to reflect the impact of equipment overheating on the mixing effect, and ensure that the material handling pattern information can reflect the actual operating status of the process step in real time.

[0059] Step S126: Input the initial material input data into the parameter configuration interface to trigger the process flow model framework to perform material processing simulation according to the material processing rules information of each process step, generate simulation output data of each process step, and complete the construction of the process flow model.

[0060] Initial material input data, such as the ash content, particle size distribution, and moisture content of raw coal, is entered into the process flow model framework through the parameter configuration interface. The interface validates the format of the input data to ensure it matches the model requirements; for example, ash content is entered as a percentage, and particle size distribution is entered as the mass percentage of each particle size range. If the data format does not meet the requirements, the interface will return an error message, and the data must be re-entered after adjustment.

[0061] After the initial material input data passes verification, the process flow model framework is triggered to start running. The model framework, following the order in the process link transmission relationship list, starts from the process link corresponding to the raw coal bunker and sequentially calls the material handling rule information encapsulated in the graphic symbols of each process link.

[0062] Taking raw coal entering the crushing stage as an example, the model uses the raw coal particle size composition, moisture content, and other data in the initial material input data as input attributes for the crushing stage. Combined with the processing parameters of the crushing stage (such as crusher speed, parameters corresponding to crusher tooth shape, etc.), the model calculates the particle size composition, output, and other simulated output data of the crushed material based on the material processing law information of the crushing stage.

[0063] The simulated output data from the crushing stage is transmitted as input data to the screening stage via the material transfer path in the model. The screening stage utilizes its own material handling characteristics and, combined with processing parameters such as screen aperture and vibration frequency, simulates the input crushed material to generate simulated output data including particle size distribution and yield of the oversize and undersize materials. The simulated output data for the oversize material is returned to the crushing stage via the corresponding transfer path, while the simulated output data for the undersize material is transmitted to the heavy media separation stage.

[0064] After receiving the simulated output data of the undersize material in the heavy medium separation stage, the heavy medium separation stage uses this data as an input attribute and combines it with processing parameters such as separation density and medium circulation volume. Based on its own material processing law information, it performs simulation calculations to obtain simulated output data such as the yield, ash content, and moisture content of clean coal, middlings, and gangue. The above data are transmitted to the subsequent dewatering stage and middlings and gangue processing stages through the corresponding transmission paths.

[0065] Each subsequent process step is simulated in the same manner as described above, until the final simulated output data such as the quality of clean coal in the clean coal bin, ash content, and the amount of filtrate in the circulating water pool are generated.

[0066] Throughout the simulation, the model framework records the simulation output data of each process step in real time, including data generation time and corresponding process step identifiers. After the simulation output data of all process steps is generated, it is summarized and organized to form a complete simulation result dataset.

[0067] Step S130: Based on the process flow model and the initial material input data, perform process flow simulation calculation to obtain the theoretical quantity and quality data of each process step. The theoretical quantity and quality data includes the quantity and quality information of the output materials of each process step. The process flow simulation calculation introduces a material attribute transmission attenuation coefficient.

[0068] Once the completed process flow model is started, the initial material input data is re-entered into the model's parameter configuration interface. The model then begins process flow simulation calculations according to the preset calculation logic. During the calculation process, a material attribute transfer attenuation coefficient is introduced to correct for changes in quantity and quality attributes of materials caused by factors such as equipment wear and tear and material residue during material transfer.

[0069] The simulation calculations begin with the initial process step and proceed progressively. After each process step is calculated, the corresponding theoretical numerical and qualitative data are output, including the quantity (e.g., yield, volume) and quality (e.g., ash content, moisture, purity) of the output materials at that step. After all process steps are calculated, the theoretical numerical and qualitative data of each step are summarized to form a complete theoretical numerical and qualitative dataset.

[0070] Step S131: Input the initial material input data into the process flow model through the parameter configuration interface of the process flow model to determine the starting process step corresponding to the initial material input data.

[0071] The initial material input data is organized into a structured data format that the model can recognize, and then input into the model through the parameter configuration interface of the process flow model. The parameter configuration interface parses the input data and extracts material source information, such as the initial feed point of raw coal.

[0072] Based on the material source information, the corresponding process step is located in the process flow model. In the coal washing process, the initial process step corresponding to the initial material input data is usually the raw coal bunker. The model determines the process step corresponding to the raw coal bunker as the starting process step for this simulation calculation by matching the process step identification information.

[0073] Step S132: Using the process flow model, following the order in the process link transfer relationship list, starting from the initial process link, call the material handling rule information encapsulated in the process link to process the initial material input data and generate the intermediate material output data of the initial process link.

[0074] The process flow model reads the list of process link transmission relationships, clarifies that the first process link after the starting process link (raw coal bunker) is the coal feeder, and defines the material transmission path between the two.

[0075] The model starts from the initial process stage (raw coal bunker) and calls upon the material handling rules information encapsulated in that process stage. The material handling rules information of the raw coal bunker includes the relationship between the amount of raw coal stored and the discharge rate, as well as rules on particle size changes during discharge.

[0076] The data related to the raw coal bunker (such as raw coal storage volume, initial particle size, etc.) in the initial material input data are used as input attributes and substituted into the material processing law information of the raw coal bunker. Combined with the processing parameters such as the feeding speed of the coal feeder, the intermediate material output data such as the amount and particle size of the material output from the raw coal bunker to the coal feeder are calculated to obtain the intermediate material output data such as the amount and particle size of the material output from the raw coal bunker to the coal feeder.

[0077] The intermediate material output data is marked with the starting process stage identifier (raw coal bunker number) and processing timestamp, and stored in the model's temporary data storage area.

[0078] Step S1321: The process flow model reads the process link transmission relationship list, determines the identifier of the starting process link and the storage path of the material handling rule information corresponding to the starting process link.

[0079] The process flow model opens the list of process link transmission relationships and extracts the identification information of the starting process link, such as "raw coal bunker-Y01". Based on this identification information, the model searches for the corresponding material handling rule information storage path in the preset storage directory, which is usually in the form of a folder named after the process link identification, such as " / process_data / raw coal bunker-Y01 / handling_rules / ".

[0080] The model verifies the validity of the storage path, ensuring that the path exists and contains complete material handling rule information files, such as rule description files and parameter configuration files. If the path is invalid, the model will return an error message, prompting a check of the storage status of the process link transfer relationship list and material handling rule information.

[0081] Step S1322: Call the material handling rule information encapsulated in the initial process step through the storage path, and load the material handling change rules and dynamic correction factors therein.

[0082] The model accesses the corresponding folder based on the determined storage path and reads the material handling pattern information file. It parses the rule description file and extracts the material handling variation rules, such as the relationship between the discharge and storage volume of raw coal bunker, the opening of the discharge valve, and the variation rules of raw coal particle size under different storage volumes.

[0083] At the same time, the dynamic correction factor in the parameter configuration file is read. This dynamic correction factor is used to reflect the impact of the aging of the raw coal bunker equipment on the discharge accuracy, such as the correction value corresponding to the change in material residue caused by the decrease in the sealing performance of the discharge valve as the usage time increases.

[0084] The model stores the loaded material handling change rules and dynamic correction factors in memory for quick retrieval during subsequent calculations.

[0085] Step S1323: Convert the initial material input data according to the format required by the material processing rule information, extract the input attribute information from the initial material input data, and match it with the input attribute items in the material processing change rule.

[0086] Initial material input data may include data in various formats, such as text-based inspection reports and tabular analysis data. Based on the requirements for input data format in the material processing rules information, the model converts the initial material input data into a unified structured data format, such as JSON or XML.

[0087] After the transformation is complete, input attribute information, such as the storage volume of raw coal, average particle size, and moisture content, is extracted from the structured data. This input attribute information is then matched one by one with the input attribute items in the material handling change rules to check for any missing attribute items.

[0088] If an unmatched input attribute exists, the model will check if the attribute is optional. If it is optional, the model will use the default value to fill in the missing information; if it is required, the model will indicate that the corresponding input attribute information is missing and needs to be filled in before reprocessing.

[0089] Step S1324: Based on the matched input attribute information and the material processing change rules, determine the default value of the processing parameters in the current operating state of the initial process step. If the initial material input data contains processing parameter adjustment information, replace the default value of the processing parameters with the parameter value in the adjustment information.

[0090] Based on the matched input attribute information (such as the raw coal storage volume being within a certain range), the model searches for the corresponding default values ​​of the processing parameters in the material handling change rules, such as the initial opening degree of the raw coal bunker outlet valve and the initial feeding speed of the coal feeder.

[0091] Check if the initial material input data includes information on adjustments to processing parameters, such as the feeder speed adjusted by operators according to the production plan. If such adjustment information exists, the model will replace the default values ​​of the processing parameters with the values ​​in the adjustment information to ensure that the parameter settings meet actual production requirements.

[0092] After the replacement is completed, the model performs a rationality check on the processing parameters to ensure that the parameter values ​​are within the allowable range (such as valve opening between 0-100%). If the parameter values ​​are outside the range, the model will return a prompt message and continue processing after the parameters are adjusted correctly.

[0093] Step S1325: Based on the matched input attribute information, the determined processing parameter values ​​and the dynamic correction factor, substitute them into the material processing change rules to calculate and obtain the material output attribute information after the initial process step.

[0094] The matched input attribute information (such as raw coal storage quantity, average particle size, etc.), the determined processing parameter values ​​(such as valve opening degree, coal feeder feeding speed, etc.), and the dynamic correction factor (such as the correction value reflecting equipment aging) are substituted into the calculation logic of the material processing change rule.

[0095] Taking the material output calculation of raw coal bunker as an example, the basic output volume is calculated based on the storage capacity and valve opening, and then the basic output volume is corrected by a dynamic correction factor to obtain the actual output volume; at the same time, based on the average particle size of raw coal and the processing parameters, the particle size distribution change of the output is calculated to obtain the particle size attributes of the processed material.

[0096] The above calculations yield the material output attributes after the initial process, including output quantity, particle size distribution, and moisture content.

[0097] Step S1326: Associate the material output attribute information with the starting process step identifier and processing time information to generate intermediate material output data for the starting process step, and store the intermediate material output data in the temporary data storage module of the process flow model.

[0098] The material output attribute information after the initial process step (such as the material quantity, particle size distribution, moisture content, etc. of the raw coal bunker) is associated with the identification information of the initial process step (such as "raw coal bunker-Y01"), the processing start time, the processing end time, and other processing time information to form a complete set of intermediate material output data.

[0099] Intermediate material output data is stored in a structured format, including fields such as "process step identifier", "process start time", "process end time", "output amount", "particle size distribution" and "moisture content".

[0100] The generated intermediate material output data is transmitted to a temporary data storage module in the process flow model. This temporary data storage module uses a partitioned storage method, classifying and storing data according to process step identifiers to facilitate quick retrieval and retrieval in subsequent process steps. Simultaneously, the temporary data storage module backs up the data to prevent data loss.

[0101] Step S133: Determine the material attribute transfer attenuation coefficient based on the material transfer distance and transfer method between process links. The material attribute transfer attenuation coefficient includes the material quantity transfer attenuation coefficient and the material quality attribute attenuation coefficient, which are used to correct the attribute loss of materials during the transfer process.

[0102] The material transfer distance between adjacent process steps is measured. For example, the material transfer distance from the coal feeder to the crusher is obtained by measuring the length of the conveyor belt between the two, and the transfer distance from the crusher to the screening machine is obtained by measuring the length of the corresponding pipe or conveyor belt.

[0103] Determine the material transfer methods between each process stage, such as using a belt conveyor to transfer materials from the coal feeder to the crusher, using a chute to transfer materials from the crusher to the screening machine, and using a pipeline pump to transfer materials from the heavy medium separator to the dewatering machine.

[0104] Based on the transmission distance and transmission method, query the preset attenuation coefficient calculation rules. For belt conveyor transmission, the longer the transmission distance, the more material spills and remains during the transmission process, and the greater the material quantity transmission attenuation coefficient. For chute transmission, considering the friction loss of material in the chute, an increase in transmission distance will lead to an increase in the quantity attenuation coefficient.

[0105] For the material quality attribute decay coefficient, if the transmission method is pipeline pumping and the material is mineral slurry, long-term transportation may cause solid particles in the mineral slurry to settle, affecting the particle size distribution of the material. In this case, the quality attribute decay coefficient is determined based on the transportation time (calculated from the transportation distance and transportation speed). If there is a temperature change during the transportation process, it may affect the moisture content of the material. The quality attribute decay coefficient needs to be adjusted in combination with the temperature change range.

[0106] Taking into account the above factors, the material quantity transfer attenuation coefficient and the material quality attribute attenuation coefficient between each process step are calculated, forming an attenuation coefficient list for subsequent material transfer correction calculations.

[0107] Step S134: Use the material quantity attenuation coefficient to attenuate and correct the quantity information in the intermediate material output data of the initial process step, and use the material quality attribute attenuation coefficient to attenuate and correct the quality information in the intermediate material output data of the initial process step to obtain corrected intermediate material output data. Use the corrected intermediate material output data as input data for the next process step, call the material processing rule information encapsulated in the next process step for processing, and generate intermediate material output data for the next process step.

[0108] Quantity information, such as the discharge volume, is extracted from the intermediate material output data of the initial process stage (raw coal bunker). Based on the material quantity transfer attenuation coefficient between the raw coal bunker and the next process stage (coal feeder), the discharge volume is corrected and calculated. The corrected discharge volume is equal to the original discharge volume multiplied by (1 minus the quantity transfer attenuation coefficient) to reflect the quantity loss of material during the transfer process.

[0109] Extract quality information, such as moisture content and particle size distribution, from the intermediate material output data of the initial process. Based on the corresponding material quality attribute attenuation coefficient, correct the moisture content (e.g., considering evaporation during transport) and the particle size distribution (e.g., considering minor particle size changes caused by collisions during transport) to obtain the corrected quality information.

[0110] The corrected quantity and quality information are combined to form the corrected intermediate material output data, which is then used as input data for the coal feeder and transmitted to the corresponding process stage of the coal feeder.

[0111] The coal feeder process calls upon its own encapsulated material handling rules information, combines it with the input corrected intermediate material output data and its own processing parameters (such as coal feeding speed adjustment parameters) to generate intermediate material output data for the coal feeder, including information such as coal feeding rate and coal feeding uniformity.

[0112] Step S135: Repeat the above steps of determining the material property transmission attenuation coefficient, attenuating and correcting the intermediate material output data of the previous process step and using it as input data for the next process step, until the material processing simulation of all process steps is completed, and the final material output data of the final process step is obtained.

[0113] Starting with the intermediate material output data of the coal feeder, determine the material transfer distance and transfer method (such as belt conveyor) between the coal feeder and the crusher, and query the corresponding material attribute transfer attenuation coefficient.

[0114] The quantity information (such as coal feed rate) and quality information (such as coal particle size) in the intermediate material output data of the coal feeder are attenuated and corrected, and the corrected input data is then transmitted to the crusher process.

[0115] The crusher process calls upon its own material handling pattern information, combines it with the corrected input data and the crusher's processing parameters (such as crushing speed, crushing gap, etc.) to generate the crusher's intermediate material output data (such as the particle size distribution and output of the crushed coal).

[0116] Following the above process, the screening machine, heavy media separator, and dewatering machine are processed sequentially. The steps of determining the attenuation coefficient, correcting previous data, and processing the current step are repeated for each process step.

[0117] Until the material has undergone all process steps, the final material output data includes the amount of clean coal, ash content, and moisture in the clean coal bin, as well as the amount and quality of materials in the middlings bin and gangue bin, and the amount of filtrate in the circulating water pool.

[0118] Step S136: Organize the intermediate material output data generated in each process step and the final material output data in the final process step, extract the quantity and quality information, and form the theoretical quantity and quality data for each process step.

[0119] Collect intermediate and final material output data for all process steps and sort them according to the order of the process steps.

[0120] Extract quantitative information from the output data of each process step, such as the output and conveying volume of each step; extract quality information, such as ash content, moisture, particle size distribution, and purity.

[0121] The extracted quantity and quality information are associated with the corresponding process step identifiers and data generation time to form theoretical quantity and quality data records for each process step.

[0122] The theoretical numerical and quality data of all process steps are recorded and summarized to form a structured theoretical numerical and quality dataset, which can be stored in the form of a database table or an Excel spreadsheet for easy subsequent querying, comparative analysis and other operations.

[0123] Step S140: Collect real-time numerical and quality data of each key process link in the process plant, process the real-time numerical and quality data using a multi-device data fusion mechanism, compare and analyze the processed real-time numerical and quality data with the theoretical numerical and quality data of the corresponding process link to obtain numerical and quality deviation analysis results, which include deviation propagation path information.

[0124] Based on the process flow and product characteristics of the process plant, heavy media separators, dewatering machines, and flotation machines were identified as key process steps. Numerical and qualitative data acquisition devices were installed at the output ports of these key process steps. For example, belt scales and online ash analyzers were installed at the clean coal output port of the heavy media separator, and moisture analyzers and yield meters were installed at the discharge port of the dewatering machine.

[0125] The data acquisition equipment collects real-time quantity data (such as output) and real-time quality data (such as ash content and moisture content) of key process links according to a preset acquisition cycle (such as once per hour).

[0126] The collected real-time data quality data is processed through a multi-device data fusion mechanism to eliminate outliers and redundant information in the data, resulting in fused real-time data quality data.

[0127] The fused real-time numerical and quality data are compared with the theoretical numerical and quality data for the corresponding process steps calculated by the process flow model to calculate quantity deviation and quality deviation. Based on the connection relationship of each key process step, the propagation path of deviation between process steps is analyzed, and finally, numerical and quality deviation analysis results containing information such as deviation value and deviation propagation path are generated.

[0128] Step S141: Based on the process flow and product composition of the process plant, identify the key process links that have a direct impact on the quantity and quality of the final product, and form a list of key process links.

[0129] Draw a complete process flow diagram of the process plant, labeling the name of each process step, input / output ports, and their interconnections. Analyze the composition of the final product (such as clean coal), clarifying that clean coal is formed by processing the clean coal product from the heavy media separator through a dewatering machine. Therefore, the heavy media separator and dewatering machine have a direct impact on the quantity and quality of the final clean coal.

[0130] Meanwhile, considering the recovery effect of the flotation process on fine coal, the processing effect of the flotation machine will also affect the final output and quality of clean coal. Therefore, the flotation machine has also been identified as a key process link.

[0131] These key process steps are identified, and information such as equipment number, location, and main function is recorded to form a list of key process steps.

[0132] Step S1411: Based on the complete process flow of the process plant, draw a process flow diagram, and mark the input port, output port and material transfer path between each process link in the process flow diagram.

[0133] Collect process design data and equipment layout diagrams from the process plant, and outline all process steps from raw coal input to final product output, including raw coal bunkers, coal feeders, crushers, screening machines, heavy media separators, dewatering machines, flotation machines, filters, and product silos. Draw a process flow diagram according to the sequence and connection relationships of each process step.

[0134] In the block diagram, draw corresponding graphic symbols for each process step. For example, a rectangle represents the raw coal bunker, a graphic with a crushing tooth pattern represents the crusher, and a graphic with a screen pattern represents the screening machine. Label each graphic symbol with its input and output ports. Arrows point from the input ports to the graphic symbol, and arrows extend from the graphic symbol to the output ports. The material type is indicated next to the arrows, such as "raw coal," "crushed material," "undersize material," or "refined coal."

[0135] Solid lines are used to connect the output ports of each process stage to the input ports of the next process stage, forming material transfer paths. For example, the output port of the raw coal bunker is connected to the input port of the coal feeder via a solid line; the output port of the coal feeder is connected to the input port of the crusher; the output port of the crusher is connected to the input port of the screening machine; one output port of the screening machine is connected to the input port of the heavy media separator, and the other output port is connected to the input port of the crusher (indicating that the material oversize is returned for crushing), and so on. In this way, the input-output relationships and material transfer paths of each process stage are clearly shown in the process flow diagram.

[0136] Step S1412: Analyze the composition of the final product and determine the material branches that constitute the final product. The material branches are the sequence of process steps that the materials go through from the initial material input to the final product output.

[0137] The final product is refined coal. Analysis of its composition shows that refined coal consists of coarse refined coal produced by the heavy media separator and fine refined coal produced by the flotation machine. Based on this, the material branches constituting the final product are determined.

[0138] The first material branch is: raw coal - raw coal bunker - coal feeder - crusher - screening machine - heavy media separator - dewatering machine - clean coal bunker. This branch corresponds to the production path of coarse clean coal and covers all the process steps from raw coal input to coarse clean coal entering the clean coal bunker.

[0139] The second material branch is: raw coal—raw coal bunker—coal feeder—crusher—screening machine—flotation machine—filter—clean coal bunker. This branch corresponds to the production path of fine-grained clean coal, including all process steps from raw coal input to the entry of fine-grained clean coal into the clean coal bunker. By identifying these material branches, the critical paths affecting the final product are clarified.

[0140] Step S1413: For each material branch, from the initial process step to the final process step, analyze the degree of influence of the output material of each process step on the material processing results of subsequent process steps. Calculate the influence factor value of each process step using the influence factor evaluation method. The influence factor value represents the degree of influence of the process step on the material processing results of subsequent processes.

[0141] Taking the first material branch (raw coal—raw coal bunker—coal feeder—crusher—screening machine—heavy media separator—dewatering machine—clean coal bunker) as an example, the analysis begins sequentially from the initial process stage. The output material of the raw coal bunker is raw coal, and its quality and quantity directly affect the conveying effect of the coal feeder. If the particle size of the raw coal in the bunker is uneven, it can lead to fluctuations in the conveying capacity of the coal feeder, which in turn affects the subsequent crushing stage and has a certain impact on the subsequent processing results of this branch. The impact factor value is calculated using the impact factor evaluation method.

[0142] The stability of the output material flow rate of the coal feeder affects the load of the crusher. If the coal feed rate is unstable, it can lead to fluctuations in the crusher's processing effect and affect the subsequent screening process. The influence factor value is calculated, and this influence factor value is slightly higher than that of the raw coal bunker.

[0143] The output particle size of the crusher directly determines the screening efficiency of the screening machine. If the particle size of the crushed material does not meet the requirements, it can lead to the undersize material of the screening machine exceeding the standard, affecting the heavy media separation effect. Its influence factor value is higher than that of the coal feeder.

[0144] The quality and quantity of the material undersize from the screening machine directly enter the heavy media separator. If the material undersize contains too many large particles, it can increase the burden on the heavy media separator and affect the separation accuracy. Its influence factor value is higher than that of the crusher.

[0145] The quality and quantity of clean coal products from heavy media separators are the input basis for subsequent dewatering processes. The separation effect directly determines the quality of the final coarse clean coal, and its influence factor value is relatively high in this branch.

[0146] The dewatering effect affects the moisture content of clean coal. Insufficient dewatering can lead to excessive moisture content in the clean coal, affecting the quality of the final product. Its impact factor value is lower than that of the heavy media separator, but higher than that of some of the preceding stages.

[0147] The same method was used for the second material branch (raw coal - raw coal bunker - coal feeder - crusher - screening machine - flotation machine - filter - clean coal bunker) to analyze the influence factor values ​​of each process step in turn. The influence factor value of the flotation machine was relatively high in this branch because it directly determines the recovery rate and quality of fine clean coal.

[0148] Step S1414: Select the process steps in each material branch whose influence factor value is greater than the influence factor threshold as candidate key process steps for that material branch.

[0149] The impact factor threshold is determined based on the production experience and process requirements of the process plant, and is used to distinguish process steps that have a significant impact on the results of subsequent material processing.

[0150] For the first material branch, the influence factor values ​​of each process step are compared with the influence factor threshold. The influence factor values ​​of the heavy medium separator and the dewatering machine are greater than the threshold, and they are selected as candidate key process steps for this branch.

[0151] In the second material branch, by comparison, the influence factor values ​​of the flotation machine and the filter machine are greater than the threshold, and they become the candidate key process links of this branch. The above-mentioned candidate key process links are the links that have a more significant impact on the quantity and quality of the final product in each material branch.

[0152] Step S1415: Summarize the candidate critical process steps of all material branches, remove duplicate process steps, and label the material branch to which each candidate critical process step belongs and its influence factor value to form a list of critical process steps.

[0153] The candidate critical process steps (heavy media separator, dewatering machine) of the first material branch and the candidate critical process steps (flotation machine, filter machine) of the second material branch were summarized, and it was checked whether there were any duplicate process steps. After checking, there were no duplicates.

[0154] For each candidate critical process step, label its corresponding material branch and influencing factor value. For example, label the heavy media separator as "Branch: Coarse Coal Production Path, Influence Factor Value: XX"; the dewatering machine as "Branch: Coarse Coal Production Path, Influence Factor Value: XX"; the flotation machine as "Branch: Fine Coal Production Path, Influence Factor Value: XX"; and the filter as "Branch: Fine Coal Production Path, Influence Factor Value: XX". Organize this information into a table to form a list of critical process steps.

[0155] Step S1416: Introduce historical data on process failures, prioritize candidate key process steps in the list of key process steps, and rank candidate key process steps with higher failure frequency, and update the list of key process steps.

[0156] Collect historical failure data for each candidate key process step, including the number of failures, duration of failures, and product quality fluctuations caused by failures in the past period for heavy media separators, dewatering machines, flotation machines, and filters.

[0157] The failure frequency of each candidate key process step was statistically analyzed. The heavy media separator had a high failure frequency due to its complex structure and harsh operating conditions. The flotation machine had a second highest failure frequency due to its involvement in reagent addition and slurry state control. The dewatering machine and filter had relatively low failure frequencies.

[0158] Candidate critical process steps are prioritized based on their failure frequency, with the heavy media separator ranked first, followed by the flotation machine, then the dewatering machine, and finally the filter. The list of critical process steps is updated according to this ranking to reflect the failure risk level of each critical process step, facilitating subsequent focused monitoring.

[0159] Step S142: For each critical process step in the list of critical process steps, select at least two suitable quantity and quality acquisition devices. The quantity and quality acquisition devices are used to collect the real-time quantity information and real-time quality information of the output materials of the corresponding critical process step.

[0160] For the heavy media separator listed in the key process steps, suitable data acquisition equipment was selected. For quantity acquisition, electromagnetic flowmeters and vortex flowmeters were chosen, both suitable for the pipeline transport environment of the heavy media separator's output material, and capable of acquiring the volumetric flow rate to reflect quantity information. For quality acquisition, online ash analyzers and X-ray fluorescence analyzers were selected. The online ash analyzer can detect the ash content of clean coal in real time, while the X-ray fluorescence analyzer can analyze the elemental composition of clean coal, indirectly reflecting quality information.

[0161] For flotation machines, turbine flow meters and ultrasonic flow meters are selected as quantity acquisition devices, which are suitable for measuring the flow rate of the slurry output from the flotation machine; near-infrared spectrometers and laser particle size analyzers are selected as quality acquisition devices. Near-infrared spectrometers can detect the composition of flotation clean coal, and laser particle size analyzers can analyze the particle size distribution of clean coal, serving as a basis for quality assessment.

[0162] For the dewatering machine, the quantity acquisition equipment uses scraper flow meters and rotor flow meters to measure the conveying volume of clean coal after dewatering; the quality acquisition equipment uses microwave moisture meters and image analyzers. The microwave moisture meter can detect the moisture content of the clean coal in real time, and the image analyzer assesses its physical state by taking pictures of the appearance of the clean coal.

[0163] The quantity acquisition equipment for the filter press uses electromagnetic flowmeters and oval gear flowmeters, while the quality acquisition equipment uses infrared moisture meters and online viscometers, used to collect quantity and quality information of the filtered material, respectively. Each key process step is equipped with at least two quantity and quality acquisition devices to improve the reliability and accuracy of data collection.

[0164] Step S143: Start the data and quality acquisition device and collect the real-time quantity and quality information of the output materials of each key process step according to the preset acquisition frequency to obtain multiple sets of real-time data of each key process step.

[0165] Start the electromagnetic flowmeter, vortex flowmeter, online ash analyzer and X-ray fluorescence analyzer equipped for the heavy media separator, and simultaneously start the turbine flowmeter, ultrasonic flowmeter, near-infrared spectrometer and laser particle size analyzer equipped for the flotation machine, as well as various quantity and quality acquisition devices equipped for the dewatering machine and filter machine.

[0166] The preset acquisition frequency is determined based on the dynamic characteristics of the process and production monitoring needs to ensure timely capture of changes in material quantity and quality. Each quantity and quality acquisition device collects real-time quantity and quality information of the materials output from the corresponding key process steps according to this frequency.

[0167] The electromagnetic flowmeter and vortex flowmeter of the heavy medium separator continuously collect the volumetric flow rate data of the output material according to the acquisition frequency, forming multiple sets of real-time quantitative data; the online ash analyzer and X-ray fluorescence analyzer simultaneously collect the ash content and elemental composition data of the clean coal, forming multiple sets of real-time quality data.

[0168] The turbine flow meter and ultrasonic flow meter of the flotation machine collect and output slurry flow data at frequency to generate multiple sets of real-time quantity data; the near-infrared spectrometer and laser particle size analyzer collect the composition and particle size distribution data of the clean coal to form multiple sets of real-time quality data.

[0169] The data acquisition devices for the dewatering machine and the filter also operate at the same acquisition frequency, generating multiple sets of real-time quantity data and real-time quality data for each. Through continuous acquisition, real-time quantity and quality data for each key process stage are accumulated.

[0170] Step S144: The multi-device data fusion mechanism is used to process the multiple sets of real-time quantity and quality data. The multi-device data fusion mechanism calculates the weighted average of the real-time quantity data and the real-time quality data respectively, and combines the acquisition accuracy weights of each quantity and quality acquisition device to generate fused real-time quantity data and real-time quality data.

[0171] For multiple sets of real-time quantity data (from electromagnetic flowmeters and vortex flowmeters) from the heavy media separator, the accuracy weight of each acquisition device in quantity acquisition is first determined. Based on the calibration reports of the devices and comparison with historical data, the measurement accuracy of the electromagnetic flowmeter is higher than that of the vortex flowmeter under this operating condition. Therefore, the weight of the electromagnetic flowmeter is set to a larger value, and the weight of the vortex flowmeter is set to a smaller value.

[0172] The real-time quantitative data collected by the electromagnetic flowmeter is multiplied by its corresponding weight, and the real-time quantitative data collected by the vortex flowmeter is multiplied by its corresponding weight. The two products are then added together and finally divided by the sum of the two weights to obtain the real-time quantitative data after fusion by the heavy medium separator.

[0173] For multiple sets of real-time mass data from the heavy media separator (from the online ash analyzer and X-ray fluorescence analyzer), the accuracy weight of each acquisition device in mass acquisition is determined. The online ash analyzer has higher accuracy in detecting the ash content of clean coal, so its weight is set to a larger value. The X-ray fluorescence analyzer's weight is set according to its accuracy in elemental analysis. The multiple sets of real-time mass data from the online ash analyzer are multiplied by their respective weights, and the multiple sets of real-time mass data from the X-ray fluorescence analyzer are multiplied by their respective weights. The sum of these two products is then divided by the sum of the two weights to obtain the merged real-time mass data from the heavy media separator.

[0174] For multiple sets of real-time quantitative and qualitative data from flotation machines, dewatering machines, and filters, the same processing method as that used for heavy media separators is adopted. The weights of each device are determined according to their acquisition accuracy, and a weighted average calculation is performed to generate their respective merged real-time quantitative and qualitative data.

[0175] Step S145: Extract the theoretical quantity and quality data corresponding to the key process step from the theoretical quantity and quality data of each process step, compare the fused real-time quantity data with the corresponding theoretical quantity data, and calculate the quantity deviation value; compare the fused real-time quality data with the corresponding theoretical quality data, and calculate the quality deviation value.

[0176] From the theoretical quantitative and qualitative data of each process step generated by the process flow model, the theoretical quantitative and qualitative data corresponding to the key process steps (heavy media separator, flotation machine, dewatering machine, and filter) are selected. Specifically, the theoretical quantitative data for the heavy media separator is the theoretical value of the clean coal volumetric flow rate calculated by the model, and the theoretical qualitative data is the theoretical value of the clean coal ash content and elemental composition; the theoretical quantitative data for the flotation machine is the theoretical value of the flow rate of the flotation clean coal, and the theoretical qualitative data is the theoretical value of the clean coal composition and particle size distribution; the theoretical quantitative data for the dewatering machine is the theoretical value of the conveying capacity of the dewatered clean coal, and the theoretical qualitative data is a theoretical description of the clean coal moisture content and physical state; the theoretical quantitative data for the filter is the theoretical value of the flow rate of the filtered material, and the theoretical qualitative data is the theoretical value of the material moisture content and viscosity.

[0177] For heavy media separators, the fused real-time quantity data (fused volumetric flow rate data) is compared with the corresponding theoretical quantity data (theoretical value of volumetric flow rate), and the difference between the two is calculated to obtain the quantity deviation value of the heavy media separator. Simultaneously, the fused real-time quality data (fused value of ash content and elemental composition) is compared with the theoretical quality data (theoretical value of ash content and elemental composition), and the difference is calculated to obtain the quality deviation value of the heavy media separator.

[0178] Using the same method, the quantity deviation and quality deviation values ​​of the flotation machine, dewatering machine, and filter machine were calculated respectively. That is, by comparing the fused real-time quantity and quality data with the corresponding theoretical quantity and quality data, the deviation values ​​of each key process link were obtained.

[0179] Step S146: Based on the connection relationship of each key process link and the material transfer sequence, track the transmission path of deviation between each process link, generate deviation propagation path information, and generate quantity and quality deviation analysis results by combining the quantity deviation value, quality deviation value and deviation propagation path information.

[0180] Step S1461: Compare the real-time quantity data and theoretical quantity data of the same key process step, and calculate the quantity deviation value using the absolute deviation calculation method. The quantity deviation value is the absolute value of the difference between the real-time quantity data and the theoretical quantity data.

[0181] Taking a heavy media separator as an example, its real-time quantity data (such as the volumetric flow rate at a certain moment) and corresponding theoretical quantity data (the theoretical value of the volumetric flow rate at the same moment) are obtained after fusion. The absolute deviation calculation method is used to subtract the theoretical quantity data from the real-time quantity data, and then take the absolute value of the difference to obtain the quantity deviation value of the heavy media separator at that moment.

[0182] The same absolute deviation calculation method is used for flotation machines, dewatering machines, and filters. The corresponding theoretical quantity data is subtracted from the merged real-time quantity data, and the absolute value of the difference is taken to obtain the quantity deviation value for each key process step. The quantity deviation value calculated in this way directly reflects the degree of deviation between the real-time quantity data and the theoretical quantity data.

[0183] Step S1462: Compare the real-time quality data and theoretical quality data of the same key process step, and select the corresponding deviation calculation method according to the type of quality data. If the quality data is component content data, the relative deviation calculation method is used to calculate the quality deviation value, which is the absolute value of the ratio to the theoretical quality data. If the quality data is state data, the quality deviation value is calculated through state matching degree evaluation. The state matching degree evaluation outputs the deviation value by comparing the degree of conformity between the real-time state and the theoretical state.

[0184] For the quality data (ash content and elemental composition of clean coal) of the heavy media separator, which belongs to the component content category, the relative deviation calculation method is adopted. The theoretical quality data is subtracted from the merged real-time quality data to obtain the difference. Then, the difference is divided by the theoretical quality data, and finally, the absolute value of the result is taken to obtain the quality deviation value of the heavy media separator.

[0185] In the quality data of the flotation machine, the clean coal composition belongs to the component content data, and the deviation value is calculated using the above relative deviation calculation method. Although the particle size distribution is numerical data, it focuses more on the distribution state. The deviation value is calculated by evaluating the state matching degree. The real-time particle size distribution is compared with the theoretical particle size distribution, and the degree of conformity between the two is analyzed. The lower the degree of conformity, the larger the deviation value.

[0186] In the quality data of the dewatering machine, the moisture content of the clean coal belongs to the component content category and is calculated using the relative deviation method; the physical state is obtained through image analysis and belongs to the state category, and the deviation value is calculated by comparing the matching degree between the real-time physical state and the theoretical physical state.

[0187] The quality data processing method for filters is similar. Moisture is calculated using relative deviation, and viscosity is calculated using an appropriate deviation method based on its characteristics, ensuring that the quality deviation value can accurately reflect the difference between real-time quality and theoretical quality.

[0188] The transmission between nodes. For example, after the quantity deviation value of the heavy medium separator is substituted into the model, the model will calculate the value of the deviation transmitted to the dewatering machine based on the material transfer ratio between it and the dewatering machine, and then determine whether the dewatering machine will generate a new quantity deviation or aggravate the original deviation, thereby tracing the transmission path of the quantity deviation from the initial process stage to the subsequent process stages.

[0189] The quality deviation value is also substituted into the quality deviation propagation path model. After the quality deviation value of the heavy medium separator is entered into the model, the model analyzes the impact of the quality deviation on the processing effect of the dewatering machine based on the quality influence relationship between the heavy medium separator and the dewatering machine, calculates the quality deviation value transmitted to the dewatering machine, and traces the propagation path of the quality deviation.

[0190] The quantity and quality deviations of the flotation machine are also substituted into the corresponding deviation propagation path model in the same way as described above to track their transmission to the filter. By combining the transmission paths of all quantity and quality deviations, the starting point, transmission node, and final impact range of each deviation are identified, generating the final deviation propagation path information.

[0191] Step S1464: Organize the identification information, real-time quantity data, theoretical quantity data, quantity deviation value, real-time quality data, theoretical quality data, quality deviation value and deviation propagation path information of each key process step into a structured data format.

[0192] Collect identification information such as equipment number and name for each key process step (heavy media separator, flotation machine, dewatering machine, filter). Simultaneously, summarize real-time quantity data, theoretical quantity data, quantity deviation values, real-time quality data, theoretical quality data, and quality deviation values ​​for each key process step.

[0193] The above information, along with the deviation propagation path information, is organized and stored in a structured data format. This includes clearly defining the name and content of each field, such as "Equipment Number: XXX, Equipment Name: Heavy Medium Separator, Real-time Quantity Data: XXX, Theoretical Quantity Data: XXX, Quantity Deviation Value: XXX, Real-time Quality Data: XXX, Theoretical Quality Data: XXX, Quality Deviation Value: XXX, Deviation Propagation Path: XXX". This structuring process makes the data clear and organized, facilitating subsequent analysis and use.

[0194] Step S1465: Set quantity deviation threshold and quality deviation threshold, compare the quantity deviation value of each key process step with the quantity deviation threshold, compare the quality deviation value with the quality deviation threshold, and determine the deviation of each key process step.

[0195] Based on the production standards and process requirements of the process plant, set quantity deviation thresholds and quality deviation thresholds. The quantity deviation threshold is used to determine whether the quantity deviation exceeds the acceptable range, while the quality deviation threshold is used to determine whether the quality deviation exceeds the tolerance.

[0196] The quantity deviation value of the heavy medium separator is compared with the quantity deviation threshold. If the quantity deviation value is greater than the quantity deviation threshold, the quantity deviation of this process step is determined to be out of tolerance; otherwise, it is not out of tolerance. Similarly, its quality deviation value is compared with the quality deviation threshold to determine whether the quality deviation is out of tolerance.

[0197] Using the same method, the quantity and quality deviations of the flotation machine, dewatering machine, and filter machine were compared with the corresponding thresholds to determine the deviations of each key process step, such as which steps had quantity deviations, which steps had quality deviations, and which steps had both quantity and quality deviations.

[0198] Step S1466: Integrate the structured data, the deviation and excess information of each key process step, and the deviation threshold information to generate the quantity and quality deviation analysis results. The quantity and quality deviation analysis results also include the method used for deviation calculation and the data acquisition time information.

[0199] The structured data, deviations in each key process step (such as the quantity of heavy media separators and the quality of flotation machines), and the set quantity and quality deviation thresholds are integrated.

[0200] During the integration process, the method used for deviation calculation is explained, such as using the absolute deviation calculation method for quantity deviation and the relative deviation calculation method for component content quality deviation, etc. At the same time, the time information of data collection is recorded, including year, month, day, hour, minute and second.

[0201] Through the above integration, a complete numerical quality deviation analysis result is formed, which presents information such as the numerical quality deviation, out-of-tolerance status, calculation method, and data source time of each key process link.

[0202] Step S150: Generate process parameter adjustment instructions based on the quantity and quality deviation analysis results, and send the process parameter adjustment instructions to the control terminal of the process plant to adjust the operating parameters of the corresponding process link. The process parameter adjustment instructions include adjustment effect prediction information.

[0203] A detailed interpretation of the logarithmic quality deviation analysis results is conducted to clarify the type (quantitative or qualitative) of deviations, their magnitudes, and propagation paths in each key process step. Based on this information, the process steps requiring adjustment and their corresponding operating parameters are determined, such as the separation density of the heavy media separator and the reagent dosage of the flotation machine.

[0204] By analyzing the correlation between deviations and operating parameters, specific adjustment amounts are calculated, such as how much to adjust the separation density of the heavy media separator or how much to increase or decrease the reagent dosage of the flotation machine. Simultaneously, the changes in deviations after adjustment are predicted, i.e., the predicted adjustment effect information, such as how much the quantity deviation is expected to decrease and to what extent the quality deviation is expected to be reduced.

[0205] The process step identification, operating parameter type, specific adjustment amount, and adjustment effect prediction information are integrated to form a process parameter adjustment instruction. This instruction is then sent to the control terminal of the process plant to adjust the operating parameters of the corresponding process step.

[0206] Step S151: Analyze the quantity deviation value, quality deviation value and deviation propagation path information in the quantity and quality deviation analysis results to determine the key process links and deviation types corresponding to the deviations. The deviation types include quantity deviation types and quality deviation types.

[0207] The logarithmic quality deviation analysis results are broken down to extract the quantity deviation value and the quality deviation value. At the same time, the deviation propagation path information is analyzed to trace the source and scope of the deviation.

[0208] Based on the process step identifiers corresponding to the deviation values, the key process steps corresponding to each deviation are determined. For example, a certain quantity deviation value corresponds to a heavy media separator, and a certain mass deviation value corresponds to a flotation machine, etc.

[0209] Based on the content reflected by the deviation value, the deviation type is distinguished. If the deviation value is related to the quantity indicators of the material (such as flow rate, conveying capacity), it is determined to be a quantity deviation type; if the deviation value is related to the quality indicators of the material (such as ash content, moisture, composition), it is determined to be a quality deviation type.

[0210] Step S152: Based on the deviation type and deviation magnitude, call the preset parameter adjustment rule library to determine the type and direction of the operating parameters that need to be adjusted for the corresponding key process steps. The type of operating parameters includes the types of process parameters that affect the quantity and quality of material processing.

[0211] The preset parameter adjustment rule library stores the corresponding rules between different deviation types, deviation magnitudes, and operating parameter adjustments. For example, when the heavy media separator experiences a quantity deviation and the deviation is large, the rule library specifies that its feed speed needs to be adjusted; when a quality deviation (ash content exceeding the standard) occurs and the deviation is small, the sorting density needs to be adjusted.

[0212] Based on the deviation type (quantity deviation or quality deviation) and deviation magnitude obtained from the analysis, the corresponding rules are called from the parameter adjustment rule library to determine the type of operating parameters that need to be adjusted for the corresponding key process links, such as the feed rate and separation density of the heavy medium separator, and the reagent dosage and pulp concentration of the flotation machine.

[0213] At the same time, the adjustment direction is determined according to the direction of the deviation (such as quantity being too much or too little, quality exceeding or falling short of standards), such as the feeding speed needing to be reduced or increased, the sorting density needing to be increased or decreased, etc.

[0214] Step S153: Based on the adjustment direction and the magnitude of the deviation, the specific adjustment amount of the operating parameter that needs to be adjusted is calculated by the parameter optimization algorithm. The parameter optimization algorithm can output the parameter adjustment amount that reduces the deviation according to the deviation situation.

[0215] For example, in step S1531: the adjustment direction, deviation magnitude and the current operating parameter value of the corresponding key process link are used as input data for the parameter optimization algorithm, and the constraint range of the operating parameter of the key process link is determined. The constraint range includes the maximum and minimum allowable values ​​of the operating parameter.

[0216] Taking a heavy media separator as an example, the algorithm is optimized by inputting the adjustment direction (such as reducing the sorting density), the magnitude of the quality deviation (such as the value of high ash content), and the current sorting density value. At the same time, based on the equipment performance and process requirements of the heavy media separator, the constraint range of the sorting density is determined, that is, the maximum and minimum allowable values, such as the sorting density cannot be lower than a certain value or higher than a certain value.

[0217] For flotation machines, if the adjustment direction is to increase the amount of reagent, the adjustment direction, the corresponding deviation, and the current reagent dosage value are input into the algorithm, and the maximum and minimum allowable values ​​of reagent dosage are determined as the constraint range to ensure that the adjusted reagent dosage is within the range allowed by the equipment and process.

[0218] Step S1532: Initialize the iteration parameters of the parameter optimization algorithm. The iteration parameters include an upper limit for the number of iterations, a convergence threshold, and an initial search step size. The convergence threshold is the standard for reducing the deviation to an acceptable range.

[0219] Set an upper limit on the number of iterations for the parameter optimization algorithm. This upper limit is determined based on the process complexity and computational accuracy requirements. For example, setting it to several tens of iterations can prevent the algorithm from getting stuck in infinite iteration.

[0220] A convergence threshold is determined; when the deviation decreases to below this threshold, the deviation is considered to be within an acceptable range, and the algorithm can stop iterating. For example, the convergence threshold can be set as a certain percentage of the initial deviation; when the adjusted deviation is less than the value corresponding to this percentage, the convergence condition is met.

[0221] Set the initial search step size. The step size is determined based on the sensitivity of the operating parameters. For example, the initial search step size for sorting density is a fixed value, and the initial search step size for reagent dosage is another fixed value. If the step size is too large, it may lead to excessive adjustment, and if the step size is too small, it may prolong the iteration time.

[0222] Step S1533: Within the constraints of the operating parameters, generate multiple candidate adjustment values ​​for the operating parameters according to the adjustment direction and the initial search step size.

[0223] Within the constraints of the heavy media separator's sorting density, multiple candidate adjustment values ​​are generated according to the adjustment direction of reducing the sorting density and the initial search step size. For example, if the current sorting density is a certain value and the initial search step size is a certain value, then the candidate adjustment value can be the current value minus one step size, minus two step sizes, etc., and all these candidate adjustment values ​​are between the maximum and minimum allowable values.

[0224] For adjusting the reagent dosage of the flotation machine, within its constraints, multiple candidate adjustment values ​​are generated according to the direction of increasing the reagent dosage and the initial search step size, such as adding one step size to the current dosage, adding two step sizes, etc., to ensure that the candidate adjustment values ​​are within the constraints.

[0225] Step S1534: Substitute each candidate adjustment value of the operating parameter into the process flow model, simulate and calculate the changes in the theoretical mass data of the corresponding key process links, and obtain the simulated deviation value corresponding to each candidate adjustment value.

[0226] Substitute each candidate adjustment value of the heavy medium separator into the process flow model. The model will simulate the operating state of the heavy medium separator at that separation density, calculate the corresponding theoretical mass data (such as the ash content of clean coal), and then compare the data with the theoretical target value to obtain the simulated deviation value corresponding to each candidate adjustment value.

[0227] Similarly, by substituting each candidate adjustment value of reagent dosage for the flotation machine into the model, the corresponding theoretical mass data change is simulated and calculated to obtain the respective simulation deviation value.

[0228] Step S1535: Introduce a stability assessment of the simulated deviation value, calculate the fluctuation range of the simulated deviation value corresponding to each candidate adjustment value in multiple consecutive simulations, and screen out candidate adjustment values ​​whose fluctuation range is less than a preset fluctuation threshold.

[0229] For each candidate adjustment value of the heavy medium separator, the process flow model is continuously substituted into the simulation calculation multiple times to obtain multiple simulated deviation values. The difference between these simulated deviation values, such as the difference between the maximum and minimum values, is calculated, and this difference is the fluctuation range.

[0230] A preset fluctuation threshold is set. If the fluctuation range of a candidate adjustment is less than the threshold, it indicates that the simulated deviation value corresponding to the adjustment is relatively stable, and the candidate adjustment is retained; otherwise, the candidate adjustment is removed.

[0231] The same stability assessment and screening were also performed on the candidate adjustment quantities for the flotation machine to ensure that the retained candidate adjustment quantities have a stable adjustment effect.

[0232] Step S1536: Compare the simulated deviation value corresponding to the selected candidate adjustment amount with the current deviation value, and select the candidate adjustment amount that makes the simulated deviation value less than the current deviation value as the effective candidate adjustment amount.

[0233] The simulated deviation value corresponding to each candidate adjustment amount of the sorting density after screening by the heavy medium separator is compared with the current mass deviation value. If the simulated deviation value of a certain candidate adjustment amount is less than the current deviation value, it indicates that the adjustment amount can reduce the deviation, and it is determined as an effective candidate adjustment amount.

[0234] The candidate adjustment quantities after screening by the flotation machine are compared with the current deviation value in the same way to select the effective candidate adjustment quantities.

[0235] Step S1537: Select the effective candidate adjustment amount that minimizes the simulation deviation value as the current optimal adjustment amount. If the simulation deviation value corresponding to the current optimal adjustment amount is less than the convergence threshold, then the current optimal adjustment amount is taken as the final adjustment amount; if it is greater, adjust the search step size and repeat the steps of generating candidate adjustment amounts, calculating simulation deviation values, performing stability evaluation and screening effective candidate adjustment amounts until the number of iterations reaches the upper limit or the simulation deviation value is less than the convergence threshold.

[0236] From the effective candidate adjustment values ​​of the heavy media separator, the one with the smallest simulated deviation value is selected as the current optimal sorting density adjustment value. If the simulated deviation value corresponding to this optimal adjustment value is less than the convergence threshold, then it is determined as the final sorting density adjustment value.

[0237] If the simulation deviation value is greater than the convergence threshold, the search step size is adjusted, such as reducing the step size to half of the original size. Then, within the constraints, candidate adjustment quantities are regenerated according to the new step size and adjustment direction, and simulation calculations, stability assessments, and effective candidate adjustment quantity screening are performed again. This process is repeated until the number of iterations reaches the set upper limit, or the obtained simulation deviation value is less than the convergence threshold.

[0238] The process of determining the optimal adjustment amount for the flotation machine is the same as that for the heavy media separator. The final effective adjustment amount is obtained through iterative calculation and step size adjustment.

[0239] Step S1538: If there is no valid candidate adjustment amount, adjust the fine adjustment range of the adjustment direction, regenerate the candidate adjustment amount of the operating parameter, and perform the simulation calculation, stability evaluation and screening steps again until the final adjustment amount that meets the requirements is obtained. The final adjustment amount is used as the specific adjustment amount of the operating parameter that needs to be adjusted.

[0240] If there are no valid candidate adjustment values ​​among the candidate adjustment values ​​of the heavy medium separator, that is, the simulated deviation value corresponding to all candidate adjustment values ​​is not less than the current deviation value, then the adjustment direction is fine-tuned. For example, if the original adjustment direction is to significantly reduce the sorting density, the fine-tuning is to slightly reduce the sorting density, and then candidate adjustment values ​​are regenerated according to the fine-tuned direction.

[0241] Next, simulation calculations, stability assessments, and screening of effective candidate adjustment values ​​are performed again. If no effective candidate adjustment value is found, the adjustment direction is further fine-tuned, such as slightly increasing the sorting density. Candidate adjustment values ​​are regenerated and related steps are performed until the final adjustment value that meets the requirements is obtained. This final adjustment value is then used as the specific adjustment value for the operating parameters that the heavy medium separator needs to be adjusted.

[0242] If a similar situation occurs in the flotation machine, adjust the fine-tuning range of the adjustment direction in the same way as described above, regenerate candidate adjustment amounts and process them until the final adjustment amount is determined.

[0243] Step S154: Substitute the key process step identifier, the type of operating parameter to be adjusted, and the specific adjustment amount into the process flow model, simulate the operating state of the adjusted process step, obtain the adjusted simulated numerical quality data, calculate the reduction of the deviation between the adjusted simulated numerical quality data and the theoretical numerical quality data, and use the reduction of the deviation as the prediction information of the adjustment effect.

[0244] Substitute the heavy media separator's identifier, the type of operating parameter to be adjusted (such as sorting density), and the specific adjustment amount into the process flow model. The model will simulate the heavy media separator's operating status under the new sorting density and output the adjusted simulated quantity data and simulated quality data.

[0245] The deviations between these adjusted simulated quality data and the corresponding theoretical quality data are calculated and compared with the deviations before adjustment to obtain the magnitude of the deviation reduction. For example, if the quality deviation before adjustment is a certain value and after adjustment it is another value, the ratio of the difference between the two to the deviation before adjustment is the magnitude of the deviation reduction. This magnitude of the deviation reduction is used as information to predict the adjustment effect.

[0246] For key process steps requiring adjustment, such as flotation machines, dewatering machines, and filters, the above method was used to obtain prediction information on their respective adjustment effects.

[0247] Step S155: Integrate the key process step identifier, the type of operating parameter to be adjusted, the specific adjustment amount, and the prediction information of the adjustment effect to generate a process parameter adjustment instruction. The process parameter adjustment instruction contains format information for communication with the control terminal.

[0248] The identification of each key process step that needs adjustment (such as equipment number), the type of operating parameter that needs adjustment (such as sorting density, reagent dosage), the specific adjustment amount (such as how much to reduce sorting density, how much to increase reagent dosage), and the corresponding prediction information of the adjustment effect (such as the magnitude of deviation reduction) are summarized.

[0249] This information is encapsulated according to the communication format agreed upon with the control terminal. The communication format information includes the data transmission protocol type, data packet structure, verification method, etc., to ensure that the control terminal can correctly receive and parse the instruction content and generate process parameter adjustment instructions.

[0250] Step S156: Send the process parameter adjustment command to the control terminal of the process plant.

[0251] The generated process parameter adjustment instructions are sent to the corresponding control terminal via industrial Ethernet or dedicated communication lines within the process plant. After receiving the instructions, the control terminal parses them, extracts information such as key process step identifiers, operating parameter types, and specific adjustment amounts, and then adjusts the operating parameters of the corresponding process steps according to the instructions to achieve optimized control of the process.

[0252] Figure 2 The illustration shows exemplary hardware and software components of a process plant digital twin implementation system 100 based on dynamic number quality analysis, which can implement the ideas of this application, according to some embodiments of this application. For example, processor 120 can be used on the process plant digital twin implementation system 100 based on dynamic number quality analysis and to perform the functions in this application.

[0253] For example, a process plant digital twin implementation system 100 based on dynamic data quality analysis may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the process plant digital twin implementation system 100 based on dynamic data quality analysis may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The process plant digital twin implementation system 100 based on dynamic data quality analysis also includes an I / O interface 150 between the computer and other input / output devices.

[0254] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned method for realizing a process plant digital twin based on dynamic data quality analysis is implemented.

[0255] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for implementing a digital twin of a process plant based on dynamic numerical quality analysis, characterized in that, The method includes: Acquire process characteristic data and initial material input data of the process plant. The process characteristic data includes the regularity information of each process to realize the material handling function. The initial material input data includes the basic attribute information of the material. A process flow model is constructed based on the process characteristic data and initial material input data. The process flow model is used to reflect the connection relationship between each process step and the material transfer relationship between each process step. The process flow model can receive material input data and output the material processing result data of the corresponding process step. Based on the process flow model and the initial material input data, process flow simulation calculation is performed to obtain theoretical quantity and quality data for each process step. The theoretical quantity and quality data includes the quantity and quality information of the output materials in each process step. The process flow simulation calculation introduces a material attribute transmission attenuation coefficient. Real-time numerical and quality data of each key process step in the process plant are collected. After processing the real-time numerical and quality data using a multi-device data fusion mechanism, the processed real-time numerical and quality data is compared and analyzed with the theoretical numerical and quality data of the corresponding process step to obtain numerical and quality deviation analysis results. The numerical and quality deviation analysis results include deviation propagation path information. Based on the numerical deviation analysis results, process parameter adjustment instructions are generated and sent to the control terminal of the process plant to adjust the operating parameters of the corresponding process link. The process parameter adjustment instructions include adjustment effect prediction information. The construction of the process flow model based on the process characteristic data and initial material input data includes: The characteristic data of the process steps are classified and organized, and the material handling rules information of each process step is extracted. The material handling rules information includes the rules of change of the material input into the output material after being processed by the process steps. Based on the actual production process of the process plant, the material transfer sequence between each process step is determined, and a process step transfer relationship list is formed. The process step transfer relationship list records the correspondence between the input ports and output ports of adjacent process steps. Using the process flow modeling tool, according to the process link transfer relationship list, drag and drop the graphic symbols corresponding to each process link to establish the connecting lines between the graphic symbols of each process link to represent the material transfer path, and initially construct the process flow model framework. The material handling rules information of each process step is encapsulated into the corresponding process step graphic symbols, and a parameter configuration interface is set for the process flow model framework. The parameter configuration interface can receive externally input process parameter adjustment data. A dynamic update interface for process characteristics is established. This interface can receive real-time collected process operation status data and update the material handling rules information of the corresponding process based on the process operation status data. Input the initial material input data into the parameter configuration interface to trigger the process flow model framework to simulate material handling according to the material handling rules of each process step, generate simulation output data of each process step, and complete the construction of the process flow model.

2. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 1, characterized in that, The process of classifying and organizing the characteristic data of the process steps and extracting the material handling pattern information of each process step includes: The process characteristic data are classified according to the functional type of the process, which includes material mixing function type, material separation function type, and material transportation function type. For the process characteristic data under each functional type, extract the information describing the correspondence between material input attributes, processing parameters and output attributes. The input attributes include the composition and state information of the material, the processing parameters include the operating parameters during the process, and the output attributes include the composition and state information of the processed material. The information on the correspondence between the extracted input attributes, processing parameters and output attributes is optimized, and duplicate and contradictory information is removed to form a preliminary material processing relationship dataset. Based on the preliminary material handling relationship dataset, by means of data fitting and rule induction, a change rule is constructed to transform the input material into the output material after being processed by the process. The change rule is used to clarify the output attribute corresponding to the combination of input attributes and processing parameters. A dynamic correction factor is set for the change rule, and the dynamic correction factor can be adjusted according to the running time of the process and maintenance records; The change rules are associated with the identification information of the corresponding process steps to form material handling rule information for each process step, and the material handling rule information is then standardized in format.

3. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 1, characterized in that, The process flow simulation calculation based on the process flow model and the initial material input data, to obtain the theoretical quality data of each process step, includes: The initial material input data is input into the process flow model through the parameter configuration interface of the process flow model to determine the starting process step corresponding to the initial material input data; Using the process flow model, in the order of the process link transmission relationship list, starting from the initial process link, the material handling rule information encapsulated in the process link is called to process the initial material input data and generate the intermediate material output data of the initial process link. Based on the material transfer distance and transfer method between process steps, a material attribute transfer attenuation coefficient is determined. The material attribute transfer attenuation coefficient includes a material quantity transfer attenuation coefficient and a material quality attribute attenuation coefficient, which are used to correct for the attribute loss of materials during the transfer process. The quantity information in the intermediate material output data of the initial process stage is attenuated and corrected by the material quantity transmission attenuation coefficient, and the quality information in the intermediate material output data of the initial process stage is attenuated and corrected by the material quality attribute attenuation coefficient, so as to obtain the corrected intermediate material output data. The corrected intermediate material output data is used as the input data of the next process stage, and the material processing rule information encapsulated in the next process stage is called for processing to generate the intermediate material output data of the next process stage. Repeat the above steps of determining the material property transmission attenuation coefficient, attenuating and correcting the intermediate material output data of the previous process step and using it as input data for the next process step, until the material processing simulation of all process steps is completed, and the final material output data of the final process step is obtained. The intermediate material output data generated in each process step and the final material output data in the final process step are organized, and the quantity and quality information are extracted to form the theoretical quantity and quality data of each process step.

4. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 3, characterized in that, The process flow model, following the order in the process step transmission relationship list, starts from the initial process step, calls the material handling rule information encapsulated in that process step to process the initial material input data, and generates the intermediate material output data of that initial process step, including: The process flow model reads the process link transmission relationship list to determine the identifier of the starting process link and the storage path of the material handling rule information corresponding to the starting process link; The material handling pattern information encapsulated in the initial process step is called through the storage path, and the material handling change rules and dynamic correction factors are loaded therein; The initial material input data is converted according to the format required by the material processing rule information, and the input attribute information in the initial material input data is extracted and matched with the input attribute items in the material processing change rule. Based on the matched input attribute information and the material handling change rules, the default values ​​of the processing parameters in the current operating state of the initial process are determined. If the initial material input data contains processing parameter adjustment information, the default values ​​of the processing parameters are replaced with the parameter values ​​in the adjustment information. Based on the matched input attribute information, the determined processing parameter values ​​and the dynamic correction factor, the material processing change rules are substituted into the calculation to obtain the material output attribute information after the initial process step. The material output attribute information is associated with the starting process step identifier and processing time information to generate intermediate material output data for the starting process step. At the same time, the intermediate material output data is stored in the temporary data storage module of the process flow model.

5. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 1, characterized in that, The real-time numerical and quality data of each key process step in the collection process plant are processed using a multi-device data fusion mechanism. The processed real-time numerical and quality data are then compared and analyzed with the theoretical numerical and quality data of the corresponding process step to obtain numerical and quality deviation analysis results, including: Based on the process flow and product composition of the process plant, identify the key process steps that have a direct impact on the quantity and quality of the final product, and form a list of key process steps. For each critical process step in the list of critical process steps, at least two suitable quantity and quality acquisition devices are selected. The quantity and quality acquisition devices are used to collect the real-time quantity information and real-time quality information of the output materials of the corresponding critical process step. The data acquisition device is activated, and real-time quantity and quality information of the output materials of each key process step are collected according to the preset acquisition frequency to obtain multiple sets of real-time data of each key process step. A multi-device data fusion mechanism is used to process the multiple sets of real-time quantity and quality data. The multi-device data fusion mechanism calculates the weighted average of the real-time quantity data and real-time quality data respectively, and combines the acquisition accuracy weights of each quantity and quality acquisition device to generate fused real-time quantity data and real-time quality data. From the theoretical quantity and quality data of each process step, extract the theoretical quantity and quality data corresponding to the key process step, compare the fused real-time quantity data with the corresponding theoretical quantity data, and calculate the quantity deviation value; compare the fused real-time quality data with the corresponding theoretical quality data, and calculate the quality deviation value. Based on the connection relationship of each key process link and the material transfer sequence, the transmission path of deviation between each process link is tracked to generate deviation propagation path information. Combined with the quantity deviation value, quality deviation value and deviation propagation path information, the quantity and quality deviation analysis results are generated.

6. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 5, characterized in that, Based on the process flow and product composition of the process plant, the key process steps that directly affect the quantity and quality of the final product are identified, and a list of key process steps is formed, including: Based on the complete process flow of the process plant, draw a process flow diagram, and mark the input port, output port and material transfer path between each process step in the process flow diagram. Analyze the composition of the final product and determine the material branches that constitute the final product. The material branches are the sequence of process steps that the materials go through from the initial material input to the final product output. For each material branch, from the initial process step to the final process step, the influence of the output material of each process step on the material processing results of subsequent process steps is analyzed in turn. The influence factor evaluation method is used to calculate the influence factor value of each process step. The influence factor value represents the degree of influence of the process step on the material processing results of subsequent processes. In each material branch, process steps with influence factor values ​​greater than the influence factor threshold are selected as candidate key process steps for that material branch. Summarize the candidate critical process steps of all material branches, remove duplicate process steps, and label the material branch to which each candidate critical process step belongs and its influencing factor value to form a list of critical process steps. Historical data on process failures are introduced, and candidate critical process steps in the list of critical process steps are prioritized. Candidate critical process steps with higher failure frequencies are ranked higher, and the list of critical process steps is updated.

7. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 5, characterized in that, The process involves comparing the fused real-time quantitative data with the corresponding theoretical quantitative data to calculate the quantitative deviation value; comparing the fused real-time quality data with the corresponding theoretical quality data to calculate the quality deviation value; and generating a quantitative-quality deviation analysis result based on the quantitative deviation value, the quality deviation value, and the deviation propagation path information, including: The real-time quantity data and theoretical quantity data of the same key process are compared, and the quantity deviation value is calculated using the absolute deviation calculation method. The quantity deviation value is the absolute value of the difference between the real-time quantity data and the theoretical quantity data. Real-time quality data and theoretical quality data for the same key process step are compared. The corresponding deviation calculation method is selected according to the type of quality data. If the quality data is component content data, the relative deviation calculation method is used to calculate the quality deviation value, which is the absolute value of the ratio to the theoretical quality data. If the quality data is state data, the quality deviation value is calculated through state matching degree assessment. The state matching degree assessment outputs the deviation value by comparing the degree of conformity between the real-time state and the theoretical state. Based on the connection sequence and material transfer relationship of each key process step in the process flow, a deviation propagation path model is constructed. The quantity deviation value of each key process step is substituted into the quantity deviation propagation path model to track the transmission path of the quantity deviation from the initial process step to the subsequent process steps. The quality deviation value of each key process step is substituted into the quality deviation propagation path model to track the transmission path of the quality deviation from the initial process step to the subsequent process steps. By combining the propagation path information of quantity deviation and quality deviation, the final deviation propagation path information is generated. The identification information, real-time quantity data, theoretical quantity data, quantity deviation value, real-time quality data, theoretical quality data, quality deviation value and deviation propagation path information of each key process step are organized into a structured data format. Set quantity deviation thresholds and quality deviation thresholds, compare the quantity deviation values ​​of each key process step with the quantity deviation thresholds, and compare the quality deviation values ​​with the quality deviation thresholds to determine the deviation exceeding the tolerance of each key process step. The structured data, deviations and threshold information of each key process step are integrated to generate quantity and quality deviation analysis results. The quantity and quality deviation analysis results also include a description of the method used for deviation calculation and data acquisition time information.

8. The method for implementing a digital twin of a process plant based on dynamic numerical quality analysis according to claim 1, characterized in that, The step of generating process parameter adjustment instructions based on the quantitative deviation analysis results and sending the process parameter adjustment instructions to the control terminal of the process plant to adjust the operating parameters of the corresponding process step includes: The quantity deviation value, quality deviation value, and deviation propagation path information in the quantity and quality deviation analysis results are analyzed to determine the key process links and deviation types corresponding to the deviations. The deviation types include quantity deviation types and quality deviation types. Based on the deviation type and deviation magnitude, a preset parameter adjustment rule library is invoked to determine the type and direction of the operating parameters that need to be adjusted for the corresponding key process steps. The type of operating parameters includes the types of process parameters that affect the quantity and quality of material processing. Based on the adjustment direction and the magnitude of the deviation, the specific adjustment amount of the operating parameter that needs to be adjusted is calculated by the parameter optimization algorithm. The parameter optimization algorithm can output the parameter adjustment amount that reduces the deviation according to the deviation situation. Substitute the key process step identifiers, the types of operating parameters to be adjusted, and the specific adjustment amounts into the process flow model, simulate the operating state of the adjusted process steps, obtain the adjusted simulated numerical quality data, calculate the reduction in the deviation between the adjusted simulated numerical quality data and the theoretical numerical quality data, and use the reduction in the deviation as the information for predicting the adjustment effect. The key process step identifiers, the types of operating parameters that need to be adjusted, the specific adjustment amounts, and the prediction information of the adjustment effects are integrated to generate process parameter adjustment instructions. The process parameter adjustment instructions include format information for communication with the control terminal. The process parameter adjustment command is sent to the control terminal of the process plant.

9. A digital twin implementation system for process plants based on dynamic numerical quality analysis, characterized in that, The device includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the process plant digital twin implementation method based on dynamic numerical quality analysis as described in any one of claims 1-8.

Images