An intelligent simulation system for Lyocell fiber production line based on digital twin

Abstract

The present invention relates to the field of digital twin simulation technology, and discloses an intelligent simulation system for a Lyocell fiber production line based on digital twins. The system includes: a 3D modeling and digital twin module, which constructs a three-dimensional model of the Lyocell fiber production line and associates real-time production data, and finally forms a three-dimensional digital twin model of the production line; a key equipment status prediction module, which predicts the remaining life of a thin-film evaporator based on its vibration data; an energy consumption prediction module, which obtains the key parameters of the glue-making process according to the process sheet of the Lyocell fiber glue-making process and predicts the energy consumption value; an early warning module, which gives an early warning in the equipment model corresponding to the target thin-film evaporator in the three-dimensional digital twin model according to the remaining life prediction result; and a process management module responsible for the compilation, management, and distribution of process documents. The system of the present invention combines the above-mentioned functional modules to form a complete system, which is of great significance for guiding Lyocell production and can help enterprises achieve better cost control.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twin simulation, and more specifically to an intelligent simulation system for a Lyocell fiber production line based on digital twin. Background Technique

[0002] As an important part of new fiber materials, Lyocell fiber has natural raw materials, products that can be naturally degraded, and a green and environmentally friendly production process, with broad application prospects, belonging to the national strategic emerging industries. In terms of economic construction, promoting the production of Lyocell fiber helps to optimize resource allocation, improve the overall efficiency of the industrial chain, relieve the pressure on resources and the environment, promote employment and local economic development, meeting multiple needs of social development.

[0003] The entire Lyocell production system is complex, and the production process belongs to process industry, with many production processes and a long process. Among them, the glue-making process is a key link in the production process of Lyocell fiber, directly affecting the quality and performance of the fiber, and the steam and electricity energy consumption account for a relatively large proportion in the cost of the Lyocell glue-making process. Accurate prediction of it helps enterprises optimize the production process, reduce costs and improve market competitiveness. At present, there is less research on energy consumption prediction in the production process of Lyocell fiber, and there is no perfect technical method for the energy consumption prediction direction. For the current energy consumption prediction direction, it mostly focuses on the building and equipment fields, with the prediction object mainly being electric energy, and these energy consumption prediction methods often regard energy consumption data as time-series data, combined with environmental parameters, through time prediction algorithms for real-time dynamic prediction, lacking attention to the static energy consumption prediction under a given process sheet. For the production of Lyocell fiber, the lack of energy consumption prediction obtained under a given process sheet will cause the power supply and gas supply departments of the enterprise to be unable to provide future electricity and gas usage plan reports, making it difficult to cope with sudden power outages and gas outages during peak electricity and gas usage times in the future, and the need for more energy consumption. Objectively, it affects the reasonable allocation of enterprise resources and long-term energy efficiency planning.

[0004] Furthermore, the status of key equipment on the Lyocell production line, including remaining life and health degree, has a very significant impact on the entire production. The maintenance and overhaul of key equipment on the production line require the cooperation of line shutdown, which leads to large fluctuations in production costs, increases the complexity and cost of production management, affects production plans and resource allocation; and the status of key equipment also affects product quality.

[0005] Therefore, it is necessary to simulate the key equipment and key processes in Lyocell production, and establish a series of models to achieve the prediction, analysis, and monitoring of various production factors in the Lyocell fiber production process. Among them, the most important difficulty lies in prediction, including energy consumption prediction, key equipment status prediction, etc. These predictions all revolve around the common goal of cost control, which is of great significance for guiding production and can significantly improve the economic benefits of Lyocell production. Summary of the Invention

[0006] The present invention proposes an intelligent simulation system for a Lyocell fiber production line based on digital twin. By predicting, analyzing, and monitoring various production factors under a given parameter combination, including energy consumption prediction, key equipment status prediction, etc., the present invention provides a basis and support for the production guidance of enterprises; moreover, it can help enterprises achieve better cost control and significantly improve the economic benefits of Lyocell production.

[0007] In order to achieve the above invention objectives, the technical solution of the present invention is as follows:

[0008] An intelligent simulation system for a Lyocell fiber production line based on digital twin, the system includes: a 3D modeling and digital twin module, a key equipment status prediction module, an energy consumption prediction module, an early warning module, and a process management module; wherein,

[0009] The 3D modeling and digital twin module is used to construct a three-dimensional model of the Lyocell fiber production line, associate the real-time production data of the production line with the three-dimensional model, and generate a three-dimensional digital twin model of the Lyocell fiber production line based on digital twin technology;

[0010] The key equipment status prediction module includes a remaining life prediction sub-module for a thin film evaporator. The remaining life prediction sub-module for the thin film evaporator acquires the real-time vibration data collected by a vibration sensor on the target thin film evaporator and the real-time main shaft speed transmitted by the central control platform of the target thin film evaporator; segments the real-time vibration data at a set time interval to obtain multiple data segments at the set time interval, then performs time-domain and frequency-domain analysis on each data segment to obtain the root mean square, kurtosis, and frequency centroid of each data segment; obtains the main shaft fault characteristic frequency based on the acquired real-time main shaft speed, and then obtains the main shaft bearing fault characteristic frequency according to the real-time main shaft speed and the size parameters of the main shaft bearing; finally, inputs the obtained root mean square, kurtosis, frequency centroid, main shaft fault characteristic frequency, and main shaft bearing fault characteristic frequency into the remaining life prediction model, and finally outputs the remaining life prediction result of the target thin film evaporator to the early warning module;

[0011] The energy consumption prediction module extracts the slurry weight, process parameter group 1, process parameter group 2, and process parameter group 3 corresponding to the current process sheet according to the process sheet of the glue-making process provided by the process management module; then comprehensively obtains the glue solution weight based on the slurry weight and process parameter group 1, obtains the glue-making consumption time based on the slurry weight and process parameter group 3, and obtains the refractive index of the glue solution based on process parameter group 2; finally, inputs the slurry weight, glue solution weight, refractive index of the glue solution, and glue-making consumption time into the trained energy consumption prediction model to output the energy consumption prediction result corresponding to the glue-making process; wherein, the process parameter group 1 includes the motor speed of the thin-film evaporator, the temperature of the fourth zone of the thin-film evaporator, the pressure of the vacuum system, and the pump speed of the vacuum pump; the process parameter group 2 includes the motor speed of the thin-film evaporator, the temperature of the fourth zone of the thin-film evaporator, the pump speed of the vacuum pump, and the outlet temperature of the heat preservation water circulation pump; the process parameter group 3 includes the motor speed of the thin-film evaporator and the pump speed of the vacuum pump;

[0012] The warning module is used to give a warning in the equipment model corresponding to the target thin-film evaporator in the three-dimensional digital twin model according to the remaining life prediction result of the target thin-film evaporator;

[0013] The process management module is used for the compilation, management, and distribution of process documents.

[0014] Preferably, the real-time vibration data includes acceleration signals in three directions respectively collected by vibration sensors at the upper and lower ends and the middle region of the barrel of the target thin-film evaporator; wherein, the vibration sensor at the upper end of the barrel is used to monitor the fault condition of the upper bearing of the main shaft, the vibration sensor at the lower end of the barrel is used to monitor the fault condition of the lower bearing of the main shaft, and the vibration sensor in the middle region of the barrel is used to monitor the fault condition of the main shaft.

[0015] Preferably, the construction process of the remaining life prediction model of the thin-film evaporator is as follows:

[0016] Obtain the historical vibration data and historical main shaft speed data of the whole process from the normal operation of the thin-film evaporator to the failure of the main shaft and / or the main shaft bearing resulting in the abnormal operation of the thin-film evaporator;

[0017] The historical vibration data includes acceleration signals in three directions respectively collected by vibration sensors at the upper and lower ends and the middle region of the barrel of the thin-film evaporator; wherein, the vibration sensor at the upper end of the barrel is used to monitor the fault condition of the upper bearing of the main shaft, the vibration sensor at the lower end of the barrel is used to monitor the fault condition of the lower bearing of the main shaft, and the vibration sensor in the middle region of the barrel is used to monitor the fault condition of the main shaft;

[0018] The acceleration signal is segmented into multiple data segments at set time intervals, and then time-domain analysis and frequency-domain analysis are performed on each data segment to extract the root mean square, kurtosis, and frequency centroid of each data segment; the fault characteristic frequency of the main shaft is obtained based on the historical main shaft speed; the fault characteristic frequency of the main shaft bearing is obtained based on the historical main shaft speed and the main shaft bearing size parameters.

[0019] Vector addition is performed on the acceleration signals in three directions of each vibration sensor to obtain multiple vibration vector data, and a one-dimensional convolutional neural network model including convolution, pooling, activation function, and fully connected layers is built for each vibration vector data; the response trend of the vibration response of each vibration vector data over time is extracted through the one-dimensional convolutional neural network model of each vibration vector data to obtain the vibration vector data characteristic curves of the upper and lower bearings of the main shaft and the main shaft.

[0020] Multiple high-order exponential functions are used to fit the vibration vector data characteristic curves of the upper and lower bearings of the main shaft and the main shaft respectively, and the remaining life characteristic curves of the upper bearing of the main shaft, the remaining life characteristic curve of the lower bearing of the main shaft, and the remaining life characteristic curve of the main shaft are formed and expressed in percentage form.

[0021] Taking the set time interval as the time resolution, a data set for the remaining life prediction model is jointly constituted according to the root mean square, kurtosis, frequency centroid, the fault characteristic frequency of the main shaft bearing, the fault characteristic frequency of the main shaft, and the data on the remaining life characteristic curves of the upper bearing of the main shaft, the remaining life characteristic curve of the lower bearing of the main shaft, and the remaining life characteristic curve of the main shaft; among them, the data on the remaining life characteristic curve is the label data.

[0022] The obtained data set is divided into a training set, a validation set, and a test set; the constructed remaining life prediction deep learning model is trained, validated, and tested respectively using the divided data set, and finally a trained remaining life prediction model is obtained; the remaining life prediction model is a one-dimensional convolutional neural network model.

[0023] Preferably, the fault characteristic frequencies of the main shaft bearing include:

[0024] Inner ring fault characteristic frequency:

[0025] ;

[0026] Outer ring fault characteristic frequency:

[0027] ;

[0028] Rolling element fault characteristic frequency:

[0029] ;

[0030] Cage fault characteristic frequency:

[0031] ;

[0032] Where: Q is the number of rolling elements, d the diameter of the rolling element, D the pitch diameter of the bearing, β is the contact angle, f r is the rotational frequency of the main shaft.

[0033] Preferably, the main shaft fault characteristic frequency is:

[0034] ;

[0035] Where: n 1 is the multiple frequency number, f r is the rotational frequency of the main shaft; is the main shaft disturbance frequency.

[0036] Preferably, after forming the remaining life characteristic curves of the upper bearing of the main shaft, the remaining life characteristic curve of the lower bearing of the main shaft, and the remaining life characteristic curve of the main shaft, the calculation formula expressed in percentage is as follows:

[0037] ;

[0038] Where: , , , , are adjustable parameters; is the abscissa of the vibration data, that is, the system time of the vibration data; is the defined time limit for the equipment to change from completely normal to abnormal; is the remaining life percentage; e is the natural logarithm base.

[0039] Preferably, the construction process of the energy consumption prediction model is as follows:

[0040] Collect the evaporation system parameter data, glue liquid and seal system parameter data, heat preservation water and vacuum system parameter data, pulp porridge weight, glue liquid weight, glue liquid refractive index, glue making consumption time, and energy consumption data of the historical original data in the Lyocell fiber glue making process, and perform data preprocessing to obtain historical data;

[0041] Calculate the process similarity of the corresponding historical data according to the process sheet of the Lyocell fiber glue making process, and rearrange the historical data according to the process similarity to obtain historical sorted data;

[0042] Select the reference group sequences from the historical sorting data. By calculating the process similarity of each reference group sequence, obtain the historical sorting data sequence, and find the threshold interval of the process similarity. Then, use the historical sorting data sequence within its threshold interval to calculate and obtain the energy consumption prediction value sequence;

[0043] Compare the energy consumption prediction value sequence with the historical energy consumption data of the corresponding reference group sequence respectively, and record their absolute error and X;

[0044] Increase the percentage P of the threshold interval of the process similarity to obtain the current threshold interval of the process similarity until the current threshold interval of the process similarity reaches the set maximum value, and obtain the percentage P of the threshold interval of the process similarity corresponding to the minimum absolute error sum X. Mark P as P s ;

[0045] Construct an energy consumption prediction model, and pass the historical sorting data through P s Screen to obtain the historical screened sorting data, and use the porridge weight, glue solution weight, glue solution refractive index, and glue-making consumption time in the historical screened sorting data as the data set of the energy consumption prediction model. After training the energy consumption prediction model with the data set, finally obtain the trained energy consumption prediction model.

[0046] Preferably, the historical original data includes historical original process parameter data, historical original energy consumption data, and historical original intermediate parameter data; the historical data includes historical process parameter data, historical energy consumption data, and historical intermediate parameter data; the historical sorting data includes historical sorted process parameter data, historical sorted intermediate parameter data, and historical sorted energy consumption data; the historical screened sorting data includes historical screened sorted process parameter data, historical screened sorted intermediate parameter data, and historical screened sorted energy consumption data.

[0047] Preferably, the process sheet includes process parameter data; the process parameter data includes evaporation system parameter data, glue solution and sealing system parameter data, insulation water and vacuum system parameter data, and porridge weight; the evaporation system parameter data includes evaporation heating system parameter data, thin film evaporator parameter data, and evaporation condensate water system parameter data; the glue solution and sealing system parameter data includes sealing liquid system parameter data and glue solution conveying parameter data; the insulation water and vacuum system parameter data includes insulation water system parameter data and vacuum system parameter data.

[0048] Preferably, the parameter data of the evaporation heating system includes the temperature of the heating system, the liquid level of the heating system, and the pressure of the heating system; the parameter data of the thin-film evaporator includes the temperatures of the four zones of the thin-film evaporator, the motor speed of the thin-film evaporator, the circulating temperature of the bearing lubricating oil of the thin-film evaporator, the speed of the reducer bearing, the speed of the cooling fan, the liquid level of the thin-film evaporator, the outlet pressure of the thin-film evaporator, and the bottom outlet temperature of the thin-film evaporator; the temperatures of the four zones of the thin-film evaporator include the jacket temperatures of the four evaporation zones and the glue solution temperatures of the four evaporation zones; the parameter data of the glue solution transportation includes the temperature of the in-pump heat exchanger for heat preservation water, the flow rate of the in-pump heat exchanger for heat preservation water, the temperature of the out-pump heat exchanger for heat preservation water, the flow rate of the out-pump heat exchanger for heat preservation water, the speed of the glue solution transfer pump, the outlet pressure of the glue solution transfer pump, and the outlet temperature of the glue solution transfer pump; the parameter data of the sealant system includes the temperature of the sealant tank, the sealant flow rate, and the liquid level of the sealant tank; the parameter data of the heat preservation water system includes the outlet temperature of the heat preservation water circulation pump and the hot water pressure of the heat preservation system; the parameter data of the evaporation condensate water system includes the temperature of the evaporation condenser, the liquid level of the evaporation condenser, the pressure of the evaporation condenser, the flow rate of the cooling water to the evaporation condenser, and the outlet flow rate of the evaporation condensate pump; the parameter data of the vacuum system includes the liquid level of the vacuum system, the pressure of the vacuum system, and the speed of the vacuum pump.

[0049] Preferably, the intermediate parameter data includes the weight of the glue solution, the refractive index of the glue solution, and the time consumed for glue production.

[0050] Preferably, the data preprocessing is specifically as follows: based on the 3 σ principle, the outliers are removed; after the outliers are removed, the missing value part is filled by weighted averaging of K adjacent data values near the missing value.

[0051] Preferably, the specific method for selecting the reference group from the historical sorting data is as follows: the first Z rows of the historical sorting data form the reference group sequence ; through each sequence in the reference group sequence , the historical sorting data corresponding to each reference group sequence is obtained , thereby forming the historical sorting data sequence of the reference group sequence , and the maximum process similarity value in the process is recorded ; among them, the threshold interval of the process similarity is , is the minimum process similarity value.

[0052] Preferably, using the historical sorting data sequence within the threshold interval of the process similarity, the energy consumption prediction value sequence is calculated, and the specific method is as follows:

[0053] Taking the intermediate parameter data of the historical sorting and the gruel weight of the process parameter data of the historical sorting in the historical sorting data sequence of the reference group sequence as the training set of the neural network model, so as to obtain the energy consumption prediction value sequence of the reference group sequence; the process similarity corresponding to the intermediate parameter data of the historical sorting belongs to the threshold range of the process similarity.

[0054] Preferably, the energy consumption prediction value sequence of the reference group sequence is respectively compared with the historical energy consumption data of the corresponding reference group sequence, and their absolute error and X are recorded. The specific method is as follows:

[0055] The energy consumption prediction value sequence of the reference group sequence is respectively compared with the historical energy consumption data of the respective corresponding reference group sequence, and their absolute error values are recorded. Traverse the Z historical sorting data rows to obtain the energy consumption prediction absolute error value sequence , and then accumulate the values of the energy consumption prediction absolute error value sequence to obtain the absolute error sum X and record it.

[0056] Preferably, the gruel weight, glue solution weight, glue solution refractive index and consumption time in the historical screening and sorting data are used as the data set of the energy consumption prediction model. After training the energy consumption prediction model with the data set, the finally trained energy consumption prediction model is obtained, including:

[0057] The glue solution weight, glue solution refractive index and consumption time in the intermediate parameter data of the historical screening and sorting and the gruel weight in the process parameters of the historical screening and sorting are used as the input features of the training set of the energy consumption prediction model, and the energy consumption data of the historical screening and sorting is used as the target output of the energy consumption prediction model. The above data is used to train the model to obtain the trained energy consumption prediction model.

[0058] Preferably, the energy consumption data is the sum of the electricity energy consumption and the steam energy consumption.

[0059] Preferably, the key equipment status prediction module further includes a health prediction sub-module for thin film evaporators; the health prediction sub-module for thin film evaporators obtains the current production parameters of the target thin film evaporator and the current production parameters of the upstream pulper, and then preprocesses the collected data, including steps of removing outliers and normalization; then, based on the expert scoring method, weights are assigned to each collected parameter from three dimensions: the degree of parameter fluctuation, the degree of influence of the parameter on the equipment health status, and the severity of the parameter deviation from the normal range, and the sum of the products of each dimension of each parameter and the corresponding weight is calculated to obtain the influence weight of each parameter in health prediction; finally, the weighted data of each parameter is input into the trained health prediction model to predict the health degree of the target thin film evaporator, and the health index of the target thin film evaporator is output; among them, the real-time production parameters of the target thin film evaporator include evaporation chamber pressure, stirring motor current, feed flow rate, feed pipe pressure, dissolving liquid temperature, solution viscosity, and solvent recovery rate; the real-time production parameters of the upstream pulper include discharge flow rate, pulp concentration, and pulp temperature.

[0060] Preferably, when collecting the real-time production parameters of the thin film evaporator and the real-time production parameters of the pulper, time synchronization processing is performed on each parameter to ensure the alignment of the time series of different data sources; among them, the parameters of the pulper are compensated in advance according to the material conveying time to match the operating state of the thin film evaporator.

[0061] Preferably, the step of removing outliers is performed by combining a physical constraint method and a statistical method. The physical constraint method refers to setting a parameter threshold based on the equipment operating range, and values outside the range are regarded as abnormal; the statistical method refers to using the 3σ principle to remove abnormal points based on the distribution of historical data.

[0062] Preferably, the degree of parameter fluctuation is measured by the normalized standard deviation, and the calculation method is as follows:

[0063] ;

[0064] In the formula, represents the normalized standard deviation of the i th production parameter; is the standard deviation of the i th production parameter, is the sum of the standard deviations of all production parameters.

[0065] Preferably, the degree of influence of the production parameter on the equipment health status is measured by the normalized expert score, and the calculation method is as follows:

[0066] ;

[0067] In the formula, represents thei Normalized expert scores for production parameters For the i expert score of the production parameter, is the total score.

[0068] Preferably, the severity of the production parameter deviating from the normal range is measured by the physical constraint weight, and the calculation method is as follows:

[0069] ;

[0070] In the formula, represents the severity of the i th production parameter deviating from the normal range; is the production parameter i the number of times exceeding the normal range, is the total sample number of the production parameter.

[0071] Preferably, the calculation method of the health index is as follows:

[0072] ;

[0073] In the formula, is the health index of the thin-film evaporator; is the predicted failure probability of the thin-film evaporator.

[0074] Preferably, the health prediction model is a long short-term memory artificial neural network. The model includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The input layer is used to receive time series feature data, and its shape is determined by the step size and the total number of features at each step size; the LSTM layer is a single layer with 64 hidden units, which is used to capture long-term dependencies in the time series; the fully connected layer is used to generate prediction results; the output layer is used to output the failure probability.

[0075] Preferably, the system further includes a cost prediction module. The cost prediction module includes a raw material cost prediction sub-module. The raw material cost prediction sub-module obtains the raw material data in the production and manufacturing process of Lyocell and performs preprocessing. Then, the preprocessed raw material data is input into the total raw material consumption prediction model, and the predicted values of raw material consumption corresponding to different machine learning algorithm models are output. Then, the predicted values of each machine learning algorithm model are weighted and fused to obtain the final predicted value of raw material consumption; finally, the raw material price is combined to obtain the predicted cost of raw materials; among them, the raw material data includes wood pulp parameter data, NMMO solution parameter data, and chemical additive parameter data; the wood pulp parameter data includes wood pulp viscosity, wood pulp polymerization degree, and wood pulp methyl cellulose content; the NMMO solution parameter data includes NMMO solution concentration; the chemical additive parameter data includes propylene glycol content and hydroxylamine content.

[0076] Preferably, the predicted raw material consumption values include the predicted wood pulp consumption value, the predicted NMMO solution consumption value, the predicted propylene glycol consumption value, and the predicted hydroxylamine consumption value; the predicted raw material costs include the predicted wood pulp cost, the predicted NMMO solution cost, the predicted propylene glycol cost, and the predicted hydroxylamine cost.

[0077] Preferably, the total predicted raw material consumption model includes a first predicted raw material consumption sub-model constructed based on a random forest, a second predicted raw material consumption sub-model constructed based on a support vector machine, and a third predicted raw material consumption sub-model constructed based on a long short-term memory network.

[0078] Preferably, the raw material data is preprocessed as follows:

[0079] The 3 σ principle is used to detect outliers in the numerical data of the raw material data, and the interquartile range method is used to detect outliers in the non-normal distribution data of the raw material data, and the outliers are removed; then missing value processing and duplicate data deduplication are performed;

[0080] The raw material data is normalized using Min-Max.

[0081] Pearson correlation analysis and principal component analysis are used to extract key features from the raw material data respectively.

[0082] K-means clustering analysis and information entropy screening method are used to perform final screening on the raw material data respectively.

[0083] Preferably, the specific method for missing value processing is as follows: for time series data, linear interpolation is used to fill; for numerical data, if the missing ratio is less than 20%, k-nearest neighbor interpolation is used to fill, and if the missing exceeds 20%, the sample is removed; for non-normal distribution data, if the missing ratio is less than 20%, k-nearest neighbor interpolation is used to fill, and if the missing exceeds 20%, the sample is removed.

[0084] Preferably, the predicted values of each machine learning algorithm model are weighted and fused to obtain the final predicted raw material consumption value, specifically: set the initial weight distribution of the total predicted raw material consumption model, and calculate the mean square error of each machine learning algorithm model; then update the Beta distribution parameters according to the mean square error, and at the same time introduce Gaussian process regression to calculate the uncertainty of each machine learning algorithm model inside the total model, and then update the Beta distribution parameters again through the uncertainty; finally, calculate according to the Beta distribution parameters updated again combined with variational Bayesian inference to obtain the fusion weights of each machine learning algorithm model respectively.

[0085] Preferably, the system further includes a production management module, which is used to formulate, adjust, approve, query, and issue annual production plans, monthly production plans, and daily production plans.

[0086] Preferably, the system further includes an equipment management module, which is used to create and manage the information of production line equipment, configure and issue inspection plans, and configure and issue equipment maintenance strategies; the equipment management module includes an equipment information maintenance sub-module, an equipment inspection sub-module, and an equipment repair and maintenance sub-module; among them,

[0087] The equipment information maintenance sub-module is used to create and manage the information of production line equipment, including basic data for configuring the name, type, and number of equipment.

[0088] The equipment inspection sub-module is used to configure and issue inspection plans to employee terminals, and collect inspection results to generate corresponding inspection logs.

[0089] The equipment repair and maintenance sub-module is used to configure and issue maintenance strategies to employee terminals, and collect equipment repair and maintenance to generate corresponding repair and maintenance logs.

[0090] Advantages of the present invention:

[0091] 1. The key equipment status prediction module of the system of the present invention can realize the remaining life prediction and / or health prediction of the Lyocell thin film evaporator, can detect potential problems of the equipment in advance, avoid sudden failure shutdown of the equipment during operation, resulting in interruption of the production process, improve the reliability of the equipment, and can reasonably arrange equipment maintenance, repair and replacement as well as production tasks according to the prediction results, reduce unnecessary equipment maintenance and replacement, reduce costs and improve production efficiency. Further, the key equipment status prediction module can also cooperate with the early warning module, the 3D modeling and digital twin module. When there are problems with the equipment status, it can quickly locate the position of the problem equipment through the three-dimensional digital twin model of the production line. Furthermore, the energy consumption prediction module of the present invention can realize the energy consumption prediction during the production process of the Lyocell fiber sizing process under the given process sheet conditions. The obtained energy consumption prediction value helps the enterprise's power supply and gas supply related departments to provide future electricity and gas usage plan reports to cope with sudden power outages and gas outages during peak electricity and gas usage times in the future, and the need for more energy consumption. Still further, the cost prediction module of the present invention considers from two aspects of raw material consumption and equipment maintenance, which helps to achieve refined cost management during the production and manufacturing process of Lyocell. In summary, the system of the present invention combines functional modules such as the 3D modeling and digital twin module, the key equipment status prediction module, the energy consumption prediction module, the cost prediction module, and the early warning module together to form a complete system. This system is of great significance for guiding the production of Lyocell fibers, can help enterprises to conduct better cost control and production task scheduling, and significantly improve the economic benefits of Lyocell fiber production.

[0092] 2. The system of the present invention integrates data from systems such as MES and DCS to establish an intelligent simulation system for the Lyocell fiber production line based on digital twin, which can break the data barriers between systems, achieve data consistency and accuracy, and improve information circulation efficiency. The system of the present invention can provide a relatively comprehensive production data view. During the production process, each department can work together based on unified data, reducing errors and delays caused by data inconsistency.

[0093] 3. The prediction of the thin film evaporator by the system of the present invention considers from two aspects of remaining life and health. Among them, the remaining life prediction focuses on the physical state changes of the key components of the equipment, such as the vibration data at the upper and lower ends and the middle area of the thin film evaporator cylinder. By analyzing and processing these data, a remaining life prediction model is constructed to judge the duration for which the equipment can still operate normally. The health prediction comprehensively considers the process production parameters of the equipment itself and related equipment, such as the temperature of the dissolving solution, the pressure in the evaporation chamber, the slurry concentration, etc., and evaluates the health status of the equipment from two aspects of the equipment operation state and the process environment. The combination of the two captures fault information from multiple dimensions at the physical level and the operation state, avoiding the limitations of a single prediction method.

[0094] 4. For the remaining life prediction of the thin film evaporator, the system of the present invention forms a data matrix by using multi-component characteristic frequency data and big data characteristic extraction data, combines the vibration data collected from the bearings at the upper and lower ends of the thin film evaporator cylinder and the main shaft in the middle region with the characteristic frequencies of the key components of the main shaft and bearings, and captures the equipment degradation information more accurately in practical applications, making the correlation between the characteristic data and the life label higher, and more effectively constructing the relationship model between the fault characteristics and the life label, and then forming a more accurate remaining life prediction model.

[0095] 5. For the remaining life prediction of the thin film evaporator, the present invention collects the vibration data within a single life cycle from normal to abnormal in four regions at the upper and lower ends and the middle region of the thin film evaporator cylinder, and designs the remaining life characteristic curve by using the coupling drive method of a multi-term high-order exponential function empirical mathematical model and big data characteristic extraction, reducing the complexity of the characteristic extraction network model, improving the real-time operation of the model, meeting the real-time life prediction requirements, and providing more accurate data labels at the same time, which can effectively improve the accuracy of the remaining life prediction.

[0096] 6. For the remaining life prediction of the thin film evaporator, the present invention obtains the remaining life characteristic curve of the main shaft by taking the average value of 4 characteristic curves of the main shaft in the middle region of the thin film evaporator, and respectively constructs the characteristic curves of the upper bearing, lower bearing and main shaft of the thin film evaporator that independently reflect the health state, simplifying the calculation process.

[0097] 7. For the remaining life prediction of the thin film evaporator, the present invention forms a three-dimensional training data set by integrating the multi-component characteristic frequency data, big data characteristic extraction data and remaining life characteristic curve data in the historical data, trains the remaining life prediction model designed based on the deep learning model, trains the remaining life prediction model based on the data set, establishes the correlation mapping between the characteristic data and the life label, improves the prediction accuracy, and thus obtains the remaining life prediction model with the accuracy meeting the requirements.

[0098] 8. For the health prediction of the thin film evaporator, the present invention improves the prediction accuracy by fusing the equipment-level parameters and process-level parameters of the thin film evaporator. Based on time series analysis and the LSTM network, by collecting the equipment parameters and in-process product parameters of the thin film evaporator, and combining with the equipment parameters and in-process product parameters of the upstream pulper, a cross-device and cross-process data coupling relationship is established. Through this multi-level data fusion method, the production conditions can be more comprehensively reflected, enabling the prediction model to more accurately identify faults, thereby reducing the false alarm and missed alarm rates and improving the prediction reliability.

[0099] 9. For the health prediction of the thin-film evaporator, the present invention adopts a method combining standard deviation, expert experience scoring, and physical constraints to assign weights to various parameters, ensuring that the importance of key parameters in the prediction model is reasonably reflected.

[0100] 10. The health prediction model of the present invention uses an LSTM time series model, improving the accuracy of equipment health prediction. The LSTM (Long Short-Term Memory network) is used to learn time series data, enhancing the long-term dependence ability of the prediction model. LSTM can capture long-term trends and avoid misjudgments caused by short-term fluctuations.

[0101] 11. The present invention improves the intelligence of equipment management through a maintenance recommendation mechanism. By predicting and calculating the health index ( HI ), maintenance recommendations for different health states are provided: when HI > 0.8, it indicates that the equipment is operating normally and only requires daily monitoring; when HI is in the range of 0.6 - 0.8, it indicates that the equipment has minor abnormalities and the process parameters can be adjusted; when HI is in the range of 0.4 - 0.6, planned maintenance needs to be arranged; when HI is in the range of 0.2 - 0.4, it indicates that the equipment has serious abnormalities and it is recommended to check immediately; when HI < 0.2, it indicates that the equipment has failed and it is recommended to stop the machine for maintenance or replace parts.

[0102] 12. Since steam energy consumption and electricity energy consumption account for a relatively large proportion in the cost of the Lyocell fiber sizing process, the present invention particularly constructs an energy consumption prediction module in the system to achieve the energy consumption prediction of the Lyocell fiber sizing process, which helps to optimize the production process of Lyocell fiber by the enterprise and reduce the cost of the Lyocell fiber sizing process.

[0103] 13. For the energy consumption prediction module, the method adopted by the present invention controls the scale of the training set by calculating the percentage of the most suitable threshold interval of process similarity, improving the credibility and accuracy of model prediction.

[0104] 14. For the energy consumption prediction of the Lyocell sizing process, the present invention screens out the historical screening and sorting data, and uses the intermediate parameter data and energy consumption parameter data of the historical screening and sorting in the historical screening and sorting data as the training set of the neural network. After training by the neural network, an energy consumption prediction model is obtained. Compared with the prior art that uses a large amount of data as the training set of the neural network model, this method only uses the screened historical screening and sorting data, effectively reducing the complexity of the model and improving the fitting ability and generalization ability of the model.

[0105] 15. The present invention extracts and processes the glue liquid weight, the refractive index of the glue liquid, and the consumption time from the current process sheet by constructing a glue liquid weight calculation model, a glue liquid refractive index calculation model, and a glue making consumption time calculation model, as the input results of the energy consumption prediction model, facilitating the energy consumption prediction model to achieve the technical effect of energy consumption prediction.

[0106] 16. For the energy consumption prediction of the Lyocell glue making process, the present invention adjusts each model parameter by solving the process similarity, reducing the risk of model failure caused by process changes and enhancing the reliability of the model.

[0107] 17. For the energy consumption prediction of the Lyocell glue making process, the present invention selects the training feature set by analyzing the mechanism affecting the output of each model. On the one hand, the model has strong interpretability, and on the other hand, it reduces the input features, facilitating the implementation of the energy consumption prediction model.

[0108] 18. The present invention constructs a glue liquid weight calculation model, a glue liquid refractive index calculation model, and a glue making consumption time calculation model based on the BP neural network algorithm and the nonlinear parameter fitting algorithm based on the least squares method, ensuring the accuracy of the finally obtained glue liquid weight, glue liquid refractive index, and consumption time, and making the final energy consumption prediction result more accurate.

[0109] 19. For the energy consumption prediction value of the glue making process output by the energy consumption prediction module of the present invention, it helps the power supply and gas supply departments of the enterprise to provide future electricity and gas usage plan reports to cope with sudden power outages and gas outages during peak electricity and gas usage times in the future, and the need for more energy consumption.

[0110] 20. For the energy consumption prediction value of the glue making process output by the energy consumption prediction module of the present invention, relevant staff judge whether the process and process equipment are unstable according to the obtained energy consumption prediction value. If instability occurs, relevant staff can check and judge the specific problems according to the instability situation, achieving the technical effect of warning relevant staff to pay attention to the process and process equipment, and ensuring the normal operation of the process and process equipment.

[0111] 21. The cost prediction module of the present invention optimizes the weight coefficient, first accurately predicts the raw material consumption and equipment maintenance, and then accurately calculates and predicts the total cost of Lyocell production and manufacturing through the obtained accurate prediction values, providing a scientific basis for Lyocell process optimization and resource scheduling, and assisting the enterprise to achieve refined cost management in the process of Lyocell production and manufacturing.

[0112] 22. The present invention constructs a total raw material consumption prediction model, a total energy consumption prediction model, and a total equipment maintenance prediction model based on three different machine learning methods: random forest, support vector machine, and long short-term memory network. Then, the three obtained raw material consumption prediction values, the three energy consumption prediction values, and the three equipment maintenance prediction values are respectively and precisely obtained through weighted voting to obtain the final raw material consumption prediction value, the final energy consumption prediction value, and the final equipment maintenance prediction value, improving the accuracy of the total production manufacturing cost of the final prediction.

[0113] 23. For the prediction of the manufacturing cost of lyocell fiber, the present invention combines variational Bayesian inference, Gaussian process regression, and Beta distribution update to optimize the calculation of the model fusion weights, making it adapt to different data distributions and improving the stability and accuracy of the system cost prediction.

[0114] 24. For the prediction of the manufacturing cost of lyocell fiber, the present invention introduces Gaussian process regression to calculate the uncertainty of the three sub-models within each total model and incorporates it into the weight update rule, achieving the technical effect of reducing the weights of high-uncertainty models and improving the stability of the system for cost prediction.

[0115] 25. For the prediction of the manufacturing cost of lyocell fiber, the Beta distribution update adopted in the present invention, compared with the traditional Beta distribution update, uses variational inference to achieve the dynamic correlation of the parameter update of the Beta distribution with data error and uncertainty, achieving the technical effect of improving the speed and adaptability of weight adjustment.

[0116] 26. For the prediction of the manufacturing cost of lyocell fiber, the present invention calculates the fusion weights of the three sub-models within each total model through variational Bayesian inference, achieving the dynamic optimization of the model weights, ensuring that the weight calculation is not only based on the error distribution but also can adapt to different data characteristics, and improving the technical effect of the prediction accuracy.

[0117] 27. The cost prediction module of the present invention predicts the obtained raw material cost and equipment maintenance cost, which is beneficial to guiding the enterprise's future production and sales, effectively reducing the enterprise's cost in production and manufacturing, and increasing the profit of the enterprise in future sales.

[0118] 28. The cost prediction value obtained by the cost prediction module of the present invention can assist in judging the process consumption cost involved in the lyocell manufacturing process. Relevant staff can specifically judge the recovery cost involved in the process consumption cost, so as to conduct refined management of the costs involved in the entire process production and manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0119] The foregoing and following specific descriptions of the present invention will become clearer when read in conjunction with the following drawings, in which:

[0120] Figure 1 is the system architecture diagram of the present invention;

[0121] Figure 2 is the layout diagram of vibration sensor measuring points on the thin film evaporator cylinder of the present invention;

[0122] Figure 3 is the vibration vector data diagram of the present invention;

[0123] Figure 4 is the vibration vector data characteristic curve of the present invention;

[0124] Figure 5 is the data processing architecture diagram of the energy consumption prediction module of the present invention;

[0125] Figure 6 is the data processing architecture diagram of the cost prediction module of the present invention. Detailed implementation manners

[0126] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions for achieving the objectives of the present invention will be further described below through specific embodiments. It should be noted that the technical solutions claimed by the present invention include, but are not limited to, the following embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the protection scope of the present invention.

[0127] Embodiment 1

[0128] The embodiment of the present invention proposes an intelligent simulation system for a Lyocell fiber production line based on digital twin. Figure 1 is the system architecture diagram of the present invention. Referring to the attached drawings of the specification Figure 1 , the system mainly includes the following functional modules connected to the central command and dispatch system: a 3D modeling and digital twin module, a key equipment status prediction module, an energy consumption prediction module, an early warning module, and a process management module.

[0129] The following will explain and illustrate each functional model in detail.

[0130] First, for the 3D modeling and digital twin module, the 3D modeling and digital twin module uses SolidWorks software to establish a three-dimensional model of the Lyocell production line according to the actual site information (spatial layout information) of the Lyocell production line scanned by the data acquisition terminal. The three-dimensional model includes equipment models and fixed building models (this model details the plane geometric structure and assembly structure information of the production line and equipment, including the precise modeling of key production equipment such as pulpers, thin film evaporators, and spinning equipment); then uses 3Ds Max 3D rendering software to perform format conversion and lightweight processing on the constructed three-dimensional model and convert it into an FBX format file; then imports the model into Unity 3D software and establishes a virtual simulation model according to the actual environment of the factory production line; finally, by connecting with sensors, MES (Manufacturing Execution System), ERP (Enterprise Resource Planning System), DCS (Distributed Control System), and OPC servers, real-time data of physical entity equipment, work-in-progress, and environment are obtained, such as the operating temperature, rotation speed, and material flow rate of the equipment. Bind these real-time data to the equipment in the corresponding three-dimensional scene, and under the drive of the Unity 3D virtual engine, the virtual model and the actual production process are synchronized in real time, and finally the three-dimensional digital twin model of the Lyocell production line is constructed. Operators can intuitively observe the real-time operating status of the production line, such as the operation of equipment and the flow trajectory of materials, through the constructed three-dimensional digital twin model of the Lyocell production line, which provides convenience for production monitoring and management.

[0131] For the key equipment status prediction module, it includes a remaining life prediction sub-module for the thin film evaporator; the remaining life prediction sub-module for the thin film evaporator processes the real-time vibration data collected and transmitted by the vibration sensor on the target thin film evaporator and the real-time spindle speed transmitted by the central control platform of the target thin film evaporator, and then inputs the processed real-time vibration data and real-time spindle speed data into the trained remaining life prediction model to obtain the remaining life prediction result of the target thin film evaporator. Specifically, the remaining life prediction sub-module for the thin film evaporator segments the vibration data at a set time interval to obtain multiple vibration data segments of the set time interval, and then performs time-domain and frequency-domain analysis on each vibration data segment to obtain the root mean square, kurtosis, and frequency centroid of each vibration data segment; obtains the spindle fault characteristic frequency according to the acquired real-time spindle speed, and then obtains the spindle bearing fault characteristic frequency according to the real-time spindle speed and the size parameters of the spindle bearing; finally, inputs the obtained root mean square, kurtosis, frequency centroid, spindle fault characteristic frequency, and spindle bearing fault characteristic frequency into the trained remaining life prediction model to output the remaining life prediction result of the target thin film evaporator; the remaining life prediction of the target thin film evaporator includes the remaining life of the spindle and / or the remaining life of the spindle bearing.

[0132] The energy consumption prediction module extracts the slurry weight, process parameter group 1, process parameter group 2, and process parameter group 3 corresponding to the current process sheet according to the process sheet provided by the process management module; then, through the glue solution weight calculation model constructed based on the BP neural network algorithm, the slurry weight and process parameter group 1 are comprehensively used to obtain the glue solution weight; through the glue-making consumption time calculation model constructed based on the least squares method for fitting non-linear parameter algorithms, the slurry weight and process parameter group 3 are comprehensively used to obtain the glue-making consumption time; through the glue solution refractive index calculation model constructed based on the BP neural network algorithm, the process parameter group 2 is used to obtain the glue solution refractive index; finally, the slurry weight, glue solution weight, glue solution refractive index, and glue-making consumption time are input into the trained energy consumption prediction model to output the energy consumption prediction result corresponding to the glue-making process.

[0133] In this embodiment, for the remaining life prediction sub-module of the thin film evaporator, the equipment structure characteristics of the thin film evaporator and the impact of vibration sensitivity on production are mainly considered. By collecting the vibration data during the operation of the thin film evaporator and analyzing it in combination with the physical characteristic frequencies, the remaining life of the equipment is finally predicted through a deep learning model. The main shaft of the large thin film evaporator has a large axial dimension and is sensitive to vibration. A complex structure is connected to the main shaft, the system structure mode is complex, and the vibration of the main shaft is likely to excite the system structure mode, affecting the normal operation of the equipment and even damaging the equipment. At the same time, the locking / sticking of the main shaft is a common fault during the operation of the large thin film evaporator, and the excessive vibration displacement of the main shaft is an important cause of this fault. The reasons for the abnormal vibration of the main shaft include imperfect equipment design, easy resonance during the operation process, as well as faults such as damage to the main shaft bearing, imbalance of main shaft wear, and bending of the main shaft. These faults will lead to a decrease in the evaporation efficiency of the thin film evaporator, unstable product quality, and even cause equipment shutdown, seriously affecting the continuity and economy of the spinning production.

[0134] Specifically, the construction process of the remaining life prediction sub-module of the thin film evaporator is as follows:

[0135] (1) Construct a thin film evaporator vibration data acquisition unit for obtaining the vibration data and main shaft speed of the thin film evaporator;

[0136] The vibration data are the acceleration signals in three directions respectively collected by a total of 6 vibration sensors arranged at the upper and lower ends and the middle area of the thin film evaporator cylinder, with a total of 18 acceleration signals; among them, 1 vibration sensor at the upper end of the cylinder is used to monitor the fault condition of the upper bearing of the main shaft, 1 vibration sensor at the lower end of the cylinder is used to monitor the fault condition of the lower bearing of the main shaft, and 4 vibration sensors in the middle area of the cylinder are used to monitor the fault condition of the main shaft;

[0137] (2)Write the vibration data preprocessing calculation template into the vibration data preprocessing unit to construct the vibration data preprocessing unit; the vibration data preprocessing calculation template is used to process the acquired vibration data and spindle speed data of the thin-film evaporator, and the data processing process involved in the preprocessing calculation template is as follows:

[0138] Segment the acquired vibration data at set time intervals respectively to obtain multiple data segments at set time intervals; perform time-domain analysis and frequency-domain analysis on each data segment respectively to obtain the root mean square, kurtosis, and frequency centroid of each vibration data segment; obtain the fault characteristic frequency of the spindle according to the spindle speed data, and then obtain the fault characteristic frequency of the spindle bearing according to the spindle speed data and the size parameters of the spindle bearing;

[0139] For the root mean square, kurtosis, and frequency centroid of each acceleration signal, as well as the fault characteristic frequency of the bearing and the fault characteristic frequency of the spindle, these data can be integrated into a data matrix, and the root mean square, kurtosis, and frequency centroid of each acceleration signal, the fault characteristic frequency of the bearing, and the fault characteristic frequency of the spindle in the data matrix are normalized and then output;

[0140] The normalization process used for the vibration data of the present invention adopts the most conventional method, and the calculation method is: the difference between the current data size and the minimum value of this type of data is divided by the difference between the maximum value and the minimum value of this type of data to obtain a number between 0 and 1;

[0141] (3)Use the historical vibration data of the thin-film evaporator to train the constructed remaining life prediction model to obtain a trained remaining life prediction model, and finally construct the remaining life prediction output unit;

[0142] (4)Associate the vibration data acquisition unit, the vibration data preprocessing unit, and the remaining life prediction output unit according to the signal relationship, and combine them to form the remaining life prediction sub-module of the thin-film evaporator.

[0143] The remaining life prediction model is constructed and trained through the following methods:

[0144] Step S1, Obtain the historical vibration data and historical spindle speed data of the whole process from normal operation to the failure of the spindle and / or spindle bearing resulting in the abnormal operation of the thin-film evaporator.

[0145] The vibration data includes acceleration signals in three directions respectively collected by vibration sensors at the upper and lower ends and the middle region of the cylinder of the thin-film evaporator; among them, 1 vibration sensor is set at each of the upper and lower ends respectively, which are used to monitor the fault conditions of the upper and lower bearings of the spindle, and 4 vibration sensors are set in the middle region of the cylinder to monitor the fault conditions of the spindle.

[0146] Step S2: Segment the obtained historical vibration data at set time intervals to obtain multiple data segments of the set time intervals; perform time-domain analysis and frequency-domain analysis on each data segment respectively. Among them, time-domain analysis is used to extract the root mean square and kurtosis of each data segment, and frequency-domain analysis is used to perform the conversion from time domain to frequency domain on each data segment and extract the frequency centroid; then obtain the fault characteristic frequency of the main shaft based on the historical main shaft speed data obtained in Step S1, and finally, based on the main shaft speed data and the dimensional parameters of the main shaft bearing, obtain the fault characteristic frequency of the main shaft bearing;

[0147] It can be understood that the data processing process involved in Step S2 is the same as that involved in the vibration data preprocessing calculation template. And, in implementation, the set time interval is 0.5 s.

[0148] Step S3: Perform vector addition on the acceleration signals in three directions of each vibration sensor obtained in Step S1 to obtain a total of 6 vibration vector data. For each vibration vector data, build a one-dimensional convolutional neural network model including convolution, pooling, activation function, and fully connected layers; extract the response trend of the vibration response of each vibration vector data changing with time through the one-dimensional convolutional neural network model of each vibration vector data to obtain the vibration vector data characteristic curves of the upper and lower bearings of the main shaft and the main shaft; use this model to capture the changing trend presented by the vibration response of the vibration vector data over time, and this trend reflects the characteristic data of the equipment life degradation, presenting as a curve as shown in Figure 4 shown. In this figure, the abscissa is time, and the data on the curve reflects the relationship between vibration data and life at the time point.

[0149] In Step S3, the output of the one-dimensional convolutional neural network model changes with time to obtain Figure 4 the changing trend presented by the vibration response of the vibration vector data over time as shown.

[0150] Step S4: Use multiple high-order exponential functions to fit the vibration vector data characteristic curves of the upper and lower bearings of the main shaft and the main shaft obtained in Step S3 respectively to form the remaining life characteristic curves of the upper bearing of the main shaft, the remaining life characteristic curve of the lower bearing of the main shaft, and the remaining life characteristic curve of the main shaft, and then express them in percentage form.

[0151] In Step S4, the vibration vector data characteristic curve of the main shaft is obtained by taking the average value of the vibration vector data characteristic curves corresponding to the vibration sensors in the middle area of the cylinder.

[0152] In Step S4, the calculation formulas for expressing the remaining life characteristic curves of the upper bearing of the main shaft, the remaining life characteristic curve of the lower bearing of the main shaft, and the remaining life characteristic curve of the main shaft in percentage form are as follows:

[0153] ;

[0154] wherein: , , , , are adjustable parameters, and the specific debugging process is as follows: First, set the initial values to 1, further observe the fitting situation between the vibration vector data characteristic curve and the remaining life characteristic curve obtained after fitting, and then adjust , , , , until the vibration vector data characteristic curve fits well with the remaining life characteristic curve obtained after fitting; is the system time of the vibration data; is the time limit defined for the equipment from completely normal to abnormal, which can be obtained from the abscissa of the vibration data; is the remaining life percentage; e is the natural base.

[0155] Step S5: Using the time interval set in Step S2 as the time resolution, the root mean square, kurtosis, frequency centroid, fault characteristic frequencies of the main shaft bearings, and fault characteristic frequencies of the main shaft obtained in Step S2, as well as the data on the remaining life characteristic curves of the upper main shaft bearing, lower main shaft bearing, and main shaft obtained in Step S4, these data together constitute the data set of the remaining life prediction model; among them, the data on the remaining life characteristic curves are label data.

[0156] In practical applications, for the sake of easy understanding, the root mean square, kurtosis, frequency centroid, fault characteristic frequencies of the main shaft bearings, fault characteristic frequencies of the main shaft obtained in Step S2, as well as the data on the remaining life characteristic curves of the upper main shaft bearing, lower main shaft bearing, and main shaft obtained in Step S4 are plotted on the same two-dimensional coordinate graph. Each moment on the two-dimensional coordinate graph corresponds to a series of data, and these data together constitute the data set.

[0157] Step S6: Based on the one-dimensional convolutional neural network model, construct a remaining life prediction model. Divide the data set obtained in Step S5 into a training set, a validation set, and a test set according to the ratio of 8:1:1. Use the divided data sets to train, validate, and test the constructed remaining life prediction model respectively, and finally obtain a trained remaining life prediction model with the required accuracy.

[0158] Since the mechanical system vibrates during operation, the vibration will change when components fail, and different degrees of failure result in different vibration responses. Therefore, it is possible to predict the time of component failure or even failure, that is, to achieve life prediction, by analyzing the characteristic changes of the vibration response. The root mean square, kurtosis, frequency centroid, fault characteristic frequencies of the main shaft, and fault characteristic frequencies of the main shaft bearing reflect the characteristics of the vibration response, and these characteristics can be used to describe the vibration situation. The remaining life prediction model is established based on a deep learning model, and this establishment process is achieved by training the deep learning model with the characteristic data of the vibration response throughout the life cycle of the equipment components. The deep learning model learns the characteristics of the component degradation process through training with a large amount of data, and thus realizes the prediction of the component life.

[0159] In the embodiment depicted in the present invention, the design method of the remaining life label data in the remaining life prediction is as follows: First, design the label data set. First, in multiple tests, obtain 18 acceleration signals in three directions of six vibration sensors throughout the process from the equipment running without damage to the equipment being unable to operate normally due to the failure of each of the main shaft and the main shaft bearing through the data acquisition module. Then, vectorially add the vibration accelerations in the three directions of each sensor to obtain 6 vibration vector data. A schematic diagram of one of the vibration vector data is as Figure 3 shown. The amplitude represents the magnitude of the vibration acceleration collected by the vibration sensor, and the sample points represent the sampling points of the vibration acceleration collected by the vibration sensor.

[0160] For the vibration vector data collected by each vibration sensor, build a one-dimensional convolutional neural network deep learning mathematical model including convolutional, pooling, activation function, and fully connected layers. Extract the response trend of the vibration response over time through this one-dimensional convolutional neural network deep learning mathematical model. Finally, form the vibration vector data characteristic curves of the vibration response over time of the vibration vector data of the upper and lower bearings and the main shaft on the main shaft, Figure 4 which is a schematic diagram of one of the vibration vector data characteristic curves. Among them, the amplitude represents the magnitude of the vibration acceleration collected by the vibration sensor, and the sample points represent the sampling points of the vibration acceleration collected by the vibration sensor. The curve has obvious characteristics. As time goes by, the damage of the main shaft and the main shaft bearing gradually accumulates and quickly collapses when it accumulates to a certain degree. This phenomenon is the same as the damage phenomenon of mechanical components in reality. Therefore, this curve can be used as a reference for estimating the remaining life of the main shaft and the main shaft bearing.

[0161] Among these six vibration vector data characteristic curves, two vibration vector data characteristic curves are derived from the data of the upper and lower sensors of the thin-film evaporator, respectively reflecting the health status of the upper and lower main shaft bearings. There are also vibration vector data characteristic curves derived from the data of four vibration sensors on the cylinder body of the thin-film evaporator, reflecting the health status of the main shaft. For the convenience of calculation, the data of these four vibration vector data characteristic curves are averaged to obtain a vibration vector data characteristic curve of the main shaft. Finally, a total of 3 vibration vector data characteristic curves reflecting the health status of the main shaft and main shaft bearings of the thin-film evaporator will be obtained, namely 1 vibration vector data characteristic curve of the upper bearing of the main shaft, 1 vibration vector data characteristic curve of the lower bearing of the main shaft, and 1 vibration vector data characteristic curve of the main shaft.

[0162] However, the obtained vibration vector data characteristic curves still have defects. For example, the fluctuations are obvious, which does not match the damage accumulation caused by the actual damage of the components of the thin-film evaporator. This is because there are many interferences in the measured vibration signals. In view of this, in this embodiment, a method combining an empirical mathematical model and big data feature extraction is used to design the remaining life characteristic curve. In this embodiment, a multi-term high-order exponential function is used to fit the vibration vector data characteristic curve, and the corresponding remaining life characteristic curves of the upper bearing of the main shaft, the lower bearing of the main shaft, and the main shaft are expressed in percentage form.

[0163] So far, with a time resolution of 0.5 s, the response time of the vibration data can be used as the abscissa, and then the root mean square, kurtosis, frequency centroid, fault characteristic frequency of the bearing, fault characteristic frequency of the shaft, and remaining life characteristic curve data of the vibration data can be plotted on the same two-dimensional coordinate graph. Each moment in this two-dimensional coordinate graph corresponds to a series of data, and these data together constitute a data set. Among them, the data on the remaining life characteristic curve are label data.

[0164] In the embodiment depicted in the present invention, time-domain and frequency-domain analyses are performed on the vibration data segments to obtain the root mean square, kurtosis, and frequency centroid of the vibration data segments, including:

[0165] Perform time-domain analysis on the data segments, extract the root mean square and kurtosis of each data segment, and the calculation methods are as follows:

[0166] ;

[0167] ;

[0168] In the formula: is the root mean square of the data segment; is the number of samples in the vibration data segment; is the i th sampling value in the data segment: is the kurtosis of the data segment; u is the average value of the data segment;

[0169] Perform a time-domain to frequency-domain conversion on the data segment using frequency-domain analysis, and then extract the frequency centroid; where

[0170] The calculation method for performing a time-domain to frequency-domain conversion on the data segment is:

[0171] ;

[0172] In the formula: is the frequency-domain value after the transformation of the data segment; N 1 is the number of discrete sampling points of the data segment; is the sequence number of the time-domain discrete value of the data segment; is the n th data sample in the data segment; J is the imaginary unit;

[0173] Furthermore, the calculation method for extracting the frequency centroid is:

[0174] ;

[0175] In the formula: is the frequency centroid; is the k th spectral line frequency value; is the total number of spectral lines.

[0176] Furthermore, in step A2, the fault characteristic frequencies of the main shaft bearing include:

[0177] Inner ring fault characteristic frequency:

[0178] ;

[0179] Outer ring fault characteristic frequency:

[0180] ;

[0181] Rolling element fault characteristic frequency:

[0182] ;

[0183] Cage fault characteristic frequency:

[0184] ;

[0185] In the formula: Q is the number of rolling elements, d the diameter of the rolling element, D the pitch diameter of the bearing,β is the contact angle, f r is the rotational frequency of the main shaft, obtained according to the rotational speed of the main shaft;

[0186] For the fault characteristic frequency of the main shaft, the calculation method is as follows:

[0187] ;

[0188] In the formula: n 1 is the multiple frequency number, f r is the rotational frequency of the main shaft; is the disturbance frequency of the main shaft.

[0189] In the present invention, the fault characteristic frequency of the main shaft and the fault characteristic frequency of the main shaft bearing together constitute the characteristic frequency of the key components of the thin film evaporator.

[0190] For the construction process of the energy consumption prediction module, it is as follows:

[0191] (5) Construct a process sheet receiving unit for obtaining the process sheet of the viscose production stage of Lyocell fiber;

[0192] (6) Construct a process sheet data extraction unit internally storing a glue liquid weight calculation model, a glue production consumption time calculation model, and a glue liquid refractive index calculation model;

[0193] (7) Construct an energy consumption prediction output unit internally storing a trained energy consumption prediction model;

[0194] (8) Associate the above-mentioned process sheet receiving unit, process sheet data extraction unit, and energy consumption prediction output unit according to the signal relationship, and combine them to form the energy consumption prediction module.

[0195] For the construction and training process of the energy consumption prediction model, it is specifically as follows:

[0196] Step B1: Collect the evaporation system parameter data, glue liquid and seal system parameter data, heat preservation water and vacuum system parameter data, pulp porridge weight, glue liquid weight, glue liquid refractive index, consumption time, and energy consumption data in the historical raw data of the Lyocell fiber viscose production process, and perform data preprocessing to obtain historical data; The data preprocessing technology is a mature technical method in the art in the prior art, and the outliers in the historical raw data can be removed through these methods, and the missing values in the historical raw data can be filled.

[0197] In this embodiment, the data preprocessing in step B1 uses 3 σ principles to remove outliers; after removing outliers, the missing value part is filled by weighted averaging of K adjacent data values near the missing value.

[0198] Step B2: Calculate the process similarity of the historical process parameter data corresponding to the process sheet of the Lyocell fiber sizing process, and rearrange the historical data in descending order of process similarity to obtain historical sorting data.

[0199] In this embodiment, in Step B2, the process similarity is calculated using cosine similarity, and the calculation method is as follows:

[0200] ;

[0201] In the formula, θ i is the process similarity; X i is the vector composed of the historical process parameter data in the i th row; X now is the vector composed of the process sheet data of the Lyocell fiber sizing process.

[0202] Step B3: Select a reference group sequence from the historical sorting data. By calculating the process similarity of each reference group sequence, obtain the historical sorting data sequence of the reference group sequence, and obtain the threshold interval of the process similarity during the process of obtaining the historical sorting data sequence of the reference group sequence. Then, use the historical sorting data sequence of the reference group sequence within the threshold interval of the process similarity to calculate the energy consumption prediction value sequence of the reference group sequence; the method for solving the process similarity in Step B3 is the same as that in Step B2.

[0203] In this embodiment, the specific method for selecting a reference group from the historical sorting data is as follows: The first Z rows of data in the historical sorting data form the reference group sequence , Z The value of Z should not be too large, as it will increase the calculation amount. Usually, can take a random even number in [10, 20]; Each sequence in the reference group sequence is used as the current process sheet, and the corresponding historical sorting data of each reference group sequence is obtained respectively, thus forming the historical sorting data sequence of the reference group sequence, and record the maximum process similarity value during the process; The threshold interval of the process similarity is , and the calculation method of the minimum process similarity value

[0204] ;

[0205] In the formula, is the initial value of the threshold interval percentage of the set process similarity;

[0206] Further, in step B3, the specific method for calculating the energy consumption prediction value sequence of the reference group sequence using the historical sorting data sequence of the reference group sequence is as follows:

[0207] Take the intermediate parameter data of the historical sorting and the porridge weight of the process parameter data in the historical sorting data sequence of the reference group sequence as the training set of the neural network model, so as to obtain the energy consumption prediction value sequence of the reference group sequence; the process similarity corresponding to the intermediate parameter data of the historical sorting belongs to the threshold interval range of the process similarity;

[0208] Furthermore, in step B3, before using the intermediate parameter data of the historical sorting and the porridge weight of the process parameter data as the training set of the neural network model for training, it is necessary to eliminate the difference between the input data dimensions through normalization processing, and the calculation method is: ;

[0209] In the formula, is the result after normalization processing of the i th row j and th column of the training set; i is the original data of the j th row and j th column of the training set; is the minimum value of the process parameters in the j th column of the training set;

[0210]

[0211] Step B4. Compare the energy consumption prediction value sequence of the reference group sequence with the historical energy consumption data of the corresponding reference group sequence respectively, and record their absolute error and X.

[0212] In this embodiment, the specific method for comparing the energy consumption prediction value sequence of the reference group sequence with the historical energy consumption data of the corresponding reference group sequence respectively and recording their absolute error and X is as follows:

[0213] The energy consumption prediction value sequences of the reference group sequences are respectively compared with the historical energy consumption data of their corresponding reference group sequences, and their absolute error values are recorded. Traverse the Z historical sorting data rows to obtain an energy consumption prediction absolute error value sequence . Accumulate the values of the energy consumption prediction absolute error value sequence to obtain the absolute error sum X and record it

[0214] Step B5: Gradually increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat steps B3 to B4 until the current threshold interval of the process similarity reaches the set maximum value, and obtain the threshold interval percentage P of the process similarity corresponding to the minimum absolute error sum X during the process of repeating steps B3 to B4. Mark P as P s ; Among them, the set maximum value of the threshold interval is set artificially and should not be too large. If the set maximum value of the threshold interval is too large, it will lead to too much training data, causing overfitting and prolonging the calculation time. In practice, the set maximum value of the threshold interval takes half of the maximum similarity

[0215] In this embodiment, the threshold interval percentage for gradually increasing the process similarity has the following specific calculation formula:

[0216] ;

[0217] In the formula, P represents the threshold interval percentage of the current process similarity; represents the initial value of the threshold interval percentage of the set process similarity; S represents the step value of the threshold interval percentage of the set process similarity; represents the current step number;

[0218] The current threshold interval of the process similarity is :

[0219] ;

[0220] In the formula, represents the current minimum process similarity value; represents the set current maximum process similarity value

[0221] Step B6: Construct an energy consumption prediction model based on the BP neural network, and then pass the historical sorting data through the threshold interval percentage P of the process similarity sThe historical screening and sorting data is obtained through the set range screening, and the gruel weight, glue solution weight, refractive index of the glue solution, consumption time, and energy consumption data in the historical screening and sorting data are used as the data set of the energy consumption prediction model. The data set is divided into a training set, a test set, and a validation set. Then, the divided data set is used to train, test, and validate the constructed energy consumption prediction model. Finally, a trained energy consumption prediction model is obtained.

[0222] In this embodiment, step B6 is specifically as follows: The glue solution weight, refractive index of the glue solution, and consumption time in the intermediate parameter data of the historical screening and sorting, and the gruel weight in the process parameters of the historical screening and sorting are used as the input features of the training set of the above-mentioned constructed energy consumption prediction model, and the energy consumption data of the historical screening and sorting is used as the target output of the energy consumption prediction model. The above data is used to train the model to obtain a trained energy consumption prediction model; the energy consumption prediction model selects a three-layer BP neural network, the number of neurons in the input layer is 4, and the number of neurons in the output layer is 1; the hidden layer is 1 layer, and the number of neurons in this layer is 4.

[0223] In this embodiment, the training set of the energy consumption prediction model needs to be normalized to eliminate the differences between the data dimensions, and then the training set is used to train the model.

[0224] In the embodiment described in the present invention, the energy consumption data is the sum of the electricity energy consumption and the steam energy consumption; among them, the specific calculation formula for converting the steam energy consumption into electricity energy consumption is as follows:

[0225] ;

[0226] In the formula, is the electricity energy consumption of the steam; is the price of the steam, and the unit is yuan / m 3 ; is the price of the electric energy, and the unit is yuan / KW·h; is the consumption of the steam, and the unit is m 3 .

[0227] In the embodiment described in the present invention, the historical raw data includes historical raw process parameter data, historical raw energy consumption data, and historical raw intermediate parameter data; the historical data includes historical process parameter data, historical energy consumption data, and historical intermediate parameter data; the historical sorting data includes historical sorted process parameter data, historical sorted intermediate parameter data, and historical sorted energy consumption data; the historical screening and sorting data includes historical screened and sorted process parameter data, historical screened and sorted intermediate parameter data, and historical screened and sorted energy consumption data; the intermediate parameter data includes glue solution weight, refractive index of the glue solution, and consumption time; the energy consumption data includes electricity energy consumption and steam energy consumption.

[0228] In the embodiments described in the present invention, the process sheet includes process parameter data; the process parameter data includes evaporation system parameter data, glue solution and sealing system parameter data, heat preservation water and vacuum system parameter data, and gruel weight; the evaporation system parameter data includes evaporation heating system parameter data, thin film evaporator parameter data, and evaporation condensate water system parameter data; the glue solution and sealing system parameter data includes sealing liquid system parameter data and glue solution conveying parameter data; the heat preservation water and vacuum system parameter data includes heat preservation water system parameter data and vacuum system parameter data.

[0229] In the embodiments described in the present invention, the evaporation heating system parameter data includes heating system temperature, heating system liquid level, and heating system pressure; the thin film evaporator parameter data includes the four-zone temperature of the thin film evaporator, the motor speed of the thin film evaporator, the circulating temperature of the bearing lubricating oil of the thin film evaporator, the bearing speed of the reducer, the speed of the cooling fan, the liquid level of the thin film evaporator, the outlet pressure of the thin film evaporator, and the bottom outlet temperature of the thin film evaporator; the four-zone temperature of the thin film evaporator includes the jacket temperature of the four evaporation zones and the glue solution temperature of the four evaporation zones; the glue solution conveying parameter data includes the temperature of the in-pump heat preservation water heat exchanger, the flow rate of the in-pump heat preservation water heat exchanger, the temperature of the out-pump heat preservation water heat exchanger, the flow rate of the out-pump heat preservation water heat exchanger, the speed of the glue solution conveying pump, the outlet pressure of the glue solution conveying pump, and the outlet temperature of the glue solution conveying pump; the sealing liquid system parameter data includes the temperature of the sealing liquid tank, the flow rate of the sealing liquid, and the liquid level of the sealing liquid tank; the heat preservation water system parameter data includes the outlet temperature of the heat preservation water circulation pump and the hot water pressure of the heat preservation system; the evaporation condensate water system parameter data includes the temperature of the evaporation condenser, the liquid level of the evaporation condenser, the pressure of the evaporation condenser, the flow rate of the cooling water to the evaporation condenser, and the outlet flow rate of the evaporation condensate liquid pump; the vacuum system parameter data includes the liquid level of the vacuum system, the pressure of the vacuum system, and the speed of the vacuum pump.

[0230] In this embodiment, the glue solution weight calculation model is constructed based on the BP neural network algorithm. Specifically: First, based on the BP neural network algorithm, a glue solution weight calculation model is constructed. Then, the gruel weight, the motor speed of the thin film evaporator, the four-zone temperature of the thin film evaporator, the pressure of the vacuum system, and the speed of the vacuum pump in the process parameter data sorted by historical screening are used as the input features of the training set of the glue solution weight calculation model, and the glue solution weight in the intermediate parameter data sorted by historical screening is used as the target output of the training set of the glue solution weight calculation model. The above data is used to train, test, and verify the model, and finally a trained glue solution weight calculation model is obtained.

[0231] It is understandable that the number of neurons in the input layer of the glue liquid weight calculation model is set to 8, and the number of neurons in the output layer is 1; the hidden layer has two layers, with 5 neurons in the first layer and 4 neurons in the second layer; the activation function selects the tanh activation function; the loss function selects MSE; the initial weights are randomly generated using a Gaussian distribution with a mean of 0 and a standard deviation of 0.01; the initial bias value is zero; the learning rate is set to 0.001; the precision is set to three decimal places.

[0232] In this embodiment, the glue liquid refractive index calculation model is constructed based on the BP neural network algorithm. Specifically: First, based on the BP neural network algorithm, a glue liquid refractive index calculation model is constructed. Then, the motor speed of the thin film evaporator, the temperature of the four zones of the thin film evaporator, the vacuum pump speed, and the outlet temperature of the heat preservation water circulation pump in the historical screened and sorted process parameter data are used as the input features of the training set of the glue liquid refractive index calculation model, and the glue liquid refractive index in the historical screened and sorted intermediate parameter data is used as the target output of the training set of the glue liquid refractive index calculation model; the above data is used to train, test, and verify the model, and finally a trained glue liquid refractive index calculation model is obtained.

[0233] It is understandable that the number of neurons in the input layer of the glue liquid refractive index calculation model is set to 8, and the number of neurons in the output layer is set to 1; the hidden layer has two layers, with 6 neurons in the first layer and 4 neurons in the second layer; the activation function selects the tanh activation function; the loss function selects MSE; the initial weights are in the range of -1 to 1; the initial bias value is zero; the learning rate is set to 0.001; the precision is set to four decimal places.

[0234] In this embodiment, the glue making consumption time calculation model is constructed based on the least squares fitting of nonlinear parameters algorithm. Specifically: First, a glue making consumption time calculation model is constructed based on the least squares fitting of nonlinear parameters algorithm. Then, the porridge weight, the motor speed of the thin film evaporator, and the vacuum pump speed in the historical screened and sorted process parameter data are used as the input features of the training set of the glue making consumption time calculation model, and the consumption time in the historical screened and sorted intermediate parameter data is used as the target output of the training set of the glue making consumption time calculation model; finally, the above data is used to train, test, and verify the constructed model, and finally a trained glue making consumption time calculation model is obtained. The glue making consumption time calculation model is as follows:

[0235] ;

[0236] In the formula, is the glue making consumption time; and are both parameters to be fitted; is the motor speed of the thin-film evaporator; is the rotational speed of the vacuum pump; is the weight of the gruel.

[0237] It can be understood that each time a new process sheet is input, fitting needs to be performed once. The fitting method is a well-known technology in the field and will not be elaborated in detail here. When initializing for the first time, and are both initialized to 1, otherwise they are initialized to the parameters of the previous fitting.

[0238] In this embodiment, for the construction of the warning module: the warning module is constructed based on the key equipment status warning threshold information. That is, the warning module internally stores the key equipment status warning threshold information, and compares and warns according to the equipment status prediction value transmitted by the key equipment status prediction module and the internally stored threshold information. For example, when the warning module determines that the remaining life of the current target thin-film evaporator is lower than the threshold, in the equipment model corresponding to the target thin-film evaporator in the 3D digital twin model, it is displayed in red, and the predicted remaining life value is synchronously displayed to achieve the warning effect; if it is determined that the remaining life of the current target thin-film evaporator is higher than the threshold, in the equipment model corresponding to the target thin-film evaporator in the 3D digital twin model, it is displayed in green. The above warning method facilitates the operator to quickly locate the position of the faulty equipment and execute the equipment replacement and maintenance plan as soon as possible.

[0239] In this embodiment, a process management module is constructed based on the process database. The process management module can be used for the compilation, management, and distribution of process documents.

[0240] Based on the above-mentioned method of constructing each functional module, finally, the 3D modeling and digital twin module, the key equipment status prediction module, the energy consumption prediction module, the fault prediction module, and the process management module are associated according to the signal relationship, and the connection relationship and interaction logic between the functional modules are established, and finally the intelligent simulation system of the present invention is constructed.

[0241] After the construction of the intelligent simulation system is completed, through the cooperation of the above-mentioned functional modules, the intelligent control of the whole process of Lyocell fiber production can be realized, which can be used to guide the production of Lyocell fiber, help enterprises to control costs better, and significantly improve the economic benefits of Lyocell production.

[0242] Embodiment 2

[0243] On the basis of Embodiment 1, the remaining life prediction sub-module of the thin-film evaporator can be used to predict the remaining life of the target thin-film evaporator. The overall working principle and process are as follows.

[0244] For the prediction of the remaining life of the thin-film evaporator, the specific steps are as follows:

[0245] Step A1: The real-time vibration data collected by a total of 6 vibration sensors at the upper and lower ends and the middle area of the cylinder of the target thin-film evaporator are collected through the vibration data acquisition unit. At the same time, the real-time spindle speed is obtained from the central control platform of the target thin-film evaporator.

[0246] The real-time vibration data includes the acceleration signals in three directions collected by each of the 6 vibration sensors set at the upper and lower ends and the middle area of the cylinder of the thin-film evaporator, totaling 18 acceleration signals in all. Here, the three directions refer to the x-axis, y-axis, and z-axis directions. Among them, the vibration sensors set at the upper end of the cylinder are used to monitor the faults of the upper bearing of the spindle, the vibration sensors set at the lower end of the cylinder are used to monitor the faults of the lower bearing of the spindle, and the 4 vibration sensors set in the middle area of the cylinder are used to monitor the faults of the spindle.

[0247] Step A2: The vibration data acquisition unit inputs the above 18 acceleration signals obtained into the vibration data preprocessing unit, and processes the collected data based on the vibration data preprocessing calculation template stored internally to obtain the root mean square, kurtosis, and frequency centroid of each real-time vibration data segment, as well as the fault characteristic frequency of the spindle and the fault characteristic frequency of the spindle bearing (the specific data processing process can refer to Step S2).

[0248] Step A3: Then the vibration data preprocessing unit inputs the root mean square, kurtosis, frequency centroid, the fault characteristic frequency of the spindle, and the fault characteristic frequency of the spindle bearing obtained in Step A2 into the remaining life prediction output unit, predicts the remaining life of the target thin-film evaporator through the internally constructed and trained remaining life prediction model, and finally outputs the remaining life of the target thin-film evaporator and sends the life prediction result to the warning module of the system. The specifically predicted remaining life includes the remaining life of the spindle and / or the remaining life of the spindle bearing.

[0249] As Figure 2 shown, a plurality of vibration sensors distributed in the preset area on the surface of the cylinder of the thin-film evaporator collect the vibration data of the upper and lower bearings of the spindle and the vibration data of the spindle in real time.

[0250] Figure 2Among the six sensors, the first vibration sensor 1 is used to collect the vibration data of the upper bearing in the upper region of the thin-film evaporator, the sixth vibration sensor 6 is used to collect the vibration data of the lower bearing in the lower region, and the second vibration sensor 2, the third vibration sensor 3, the fourth vibration sensor 4, and the fifth vibration sensor 5 are used to collect the vibration data of the main shaft in the middle region. In implementation, preferably, the first vibration sensor 1 is installed on the cylinder body in the area corresponding to the upper bearing and the upper mounting plate, and the second vibration sensor 2 is installed on the cylinder body in the area corresponding to the lower bearing and the lower mounting plate. The vibration data includes 18 acceleration signals in three directions respectively collected by 6 vibration sensors at the upper and lower ends and the middle region of the thin-film evaporator cylinder body; among them, the vibration sensors at the upper end of the cylinder body are used to monitor the fault conditions of the upper bearing of the main shaft, the vibration sensors at the lower end of the cylinder body are used to monitor the fault conditions of the lower bearing of the main shaft, and the 4 vibration sensors in the middle region of the cylinder body are used to monitor the fault conditions of the main shaft.

[0251] The vibration data obtained via the vibration sensors above is preprocessed through the thin-film evaporator vibration data preprocessing calculation template to obtain the root mean square, kurtosis, and frequency centroid, and the fault characteristic frequencies of the bearing and the main shaft are calculated. After normalizing these data, they are input into the trained remaining life prediction model. The remaining life prediction model predicts the main shaft and the main shaft bearing and outputs the remaining life of the main shaft and the main shaft bearing; the remaining life prediction result is the percentage of the remaining life of the equipment.

[0252] The percentage of the remaining life of the target thin-film evaporator predicted by the remaining life prediction model of this embodiment is of great significance for equipment management and production scheduling. In terms of equipment management, the maintenance team can prepare the corresponding maintenance tools, spare parts, and allocate professional maintenance personnel in advance according to the remaining life prediction result. In this way, when a device failure occurs, it can respond quickly, reduce the device downtime, and reduce the production interruption loss caused by the device failure. For production scheduling, production managers can reasonably adjust the production plan according to the remaining life of the equipment to avoid affecting product delivery due to equipment failure.

[0253] In addition, by accumulating and analyzing the equipment remaining life prediction data in the long term, the potential laws of equipment failures can also be discovered. For example, in a specific production environment or operating conditions, the remaining life of the equipment decays faster and the frequency of failures is higher. Based on these findings, the equipment maintenance strategy and production process can be further optimized to extend the service life of the equipment and improve the stability and reliability of production.

[0254] Embodiment 3

[0255] Based on Embodiment 1 or Embodiment 2, the key equipment status prediction module of this embodiment may further include a health prediction sub-module for a thin-film evaporator. The health prediction sub-module for the thin-film evaporator is used to obtain the current real-time production parameters of the target thin-film evaporator and the current real-time production parameters of the upstream pulper, and then preprocess the collected data, including steps of removing outliers and normalization; then, based on the expert scoring method, weights are assigned to each collected parameter from three dimensions: the degree of parameter fluctuation, the degree of influence of the parameter on the equipment health status, and the severity of the parameter deviation from the normal range, and the sum of the products of each dimension of each parameter and the corresponding weight is calculated to obtain the influence weight of each parameter in health prediction; finally, the weighted data of each parameter is input into the trained health prediction model to predict the health degree of the target thin-film evaporator, and a health index of the target thin-film evaporator is output; wherein, the real-time production parameters of the thin-film evaporator include evaporation chamber pressure, stirring motor current, feed flow rate, feed pipe pressure, dissolving liquid temperature, solution viscosity, and solvent recovery rate; the real-time production parameters of the pulper include discharge flow rate, pulp concentration, and pulp temperature.

[0256] For the above-mentioned health prediction sub-module of the thin-film evaporator, the working principle of the thin-film evaporator is to make the pulp porridge flow in a film shape along the heating pipe wall, evaporate the water therein, and dissolve the cellulose to prepare the spinning sizing agent. The influencing factors of the operating state of the thin-film evaporator not only depend on its own equipment parameters, but also are closely related to the internal product parameters of the equipment. For example, in terms of equipment parameters, the fluctuation of the stirring motor current of the thin-film evaporator can reflect the load condition of the stirring motor, and abnormal conditions may indicate motor overload or failure; the feed pipe pressure reflects the fluid flow state, and abnormal conditions may indicate blockage or pressure loss, etc. In terms of the internal product parameters of the equipment, the dissolving liquid temperature of the thin-film evaporator directly affects the dissolution efficiency of cellulose. When the temperature is abnormal, it may cause uneven pulp, thereby affecting the entire process; the evaporation chamber pressure affects the solvent evaporation rate, and abnormal pressure may cause evaporation failure or vacuum system failure, etc. In addition, as a downstream device of the pulper, the operating state of the thin-film evaporator is also affected by the operating conditions of the pulper. For example, the pulp concentration of the pulper will affect the dissolution uniformity, and too high may cause blockage; the pulp temperature of the pulper directly affects the feed temperature of the thin-film evaporator; the discharge flow rate of the pulper reflects the discharge stability of the pulper and affects the continuity of the subsequent process, etc.

[0257] However, current conventional device health prediction methods are mostly limited to the device level, relying only on the operating data of the device itself (such as vibration, current, pressure, etc.) for health prediction, lacking health correlation analysis at the process level. For the devices on the Lyocell fiber production line, since their operating states are affected by multiple factors, it is particularly prone to false alarms due to short-term transient process changes, resulting in certain parameter over-limit warnings, unnecessary shutdowns, and delays in production progress. For example, assuming that the operating temperature of the thin-film evaporator device exceeds the normal range, looking at the device data alone may determine it as a failure. However, by combining process data analysis, it is found that it is a normal phenomenon due to the current production formula adjustment, so there is no need for false alarm warnings. Therefore, a prediction system lacking process-level analysis is prone to misjudging normal operating conditions as failures or failing to capture real potential failures due to limited prediction accuracy.

[0258] Therefore, the health prediction sub-module of the thin-film evaporator of the present invention mainly performs fault correlation analysis by combining device-level and process-level data to predict the health level of the thin-film evaporator, and finally outputs the health index of the device. It overall improves the prediction accuracy and reliability, optimizes production scheduling, and enhances the overall production efficiency.

[0259] Therefore, the steps for constructing and training the health prediction model involved in the health prediction sub-module are as follows:

[0260] Step D1: Collect production parameters related to the operating state of the thin-film evaporator, including device parameters of the thin-film evaporator and in-process product parameters of the device; specifically, the device parameters of the thin-film evaporator collected include evaporation chamber pressure, stirring motor current, feed flow rate, feed pipe pressure, and the in-process product parameters of the thin-film evaporator include dissolution liquid temperature, solution viscosity, and solvent recovery rate.

[0261] The influences of the above parameters on the operating state of the thin-film evaporator are as follows:

[0262] Evaporation chamber pressure (Pa) - The evaporation chamber pressure affects the evaporation rate. Too low pressure will cause the solution concentration to be too high, affecting NMMO recovery; too high pressure may cause equipment leakage.

[0263] Stirring motor current (A) - The stirring motor current affects the slurry uniformity. Too high current may cause motor failure due to increased load, and lower than the normal range may indicate motor aging or too low viscosity.

[0264] Feed flow rate (m³ / h) - The feed flow rate directly affects the stability of the evaporation and dissolution process. Low flow rate may mean pipeline blockage or abnormal pump operation, and too high flow rate may cause solution dilution, affecting evaporation efficiency.

[0265] Feed pipe pressure (Pa) - The feed pipe pressure reflects the fluid flow state. Abnormalities may indicate blockages or pressure losses; it has an indirect impact on the thin-film evaporator but can be used to monitor abnormalities.

[0266] Dissolved solution temperature (°C) - The dissolved solution temperature directly affects the dissolution efficiency of cellulose. When the temperature is abnormal, it may lead to uneven slurry, thereby affecting the entire process.

[0267] Solution viscosity (Pa·s) - The solution viscosity directly affects the spinning quality. Too low may indicate insufficient solution concentration, affecting the finished product quality; too high may cause equipment blockages.

[0268] Solvent (NMMO) recovery rate (%) - The solvent concentration affects the final dissolution effect. A low recovery rate may mean leakage or vacuum system failure.

[0269] Furthermore, as a downstream device of the pulper, the operating state of the thin-film evaporator is also affected by the operating conditions of the pulper. For example, when the slurry viscosity of the pulper is abnormal, it may affect the dissolution efficiency of the thin-film evaporator, resulting in fluctuations in solution viscosity, and further affecting the subsequent spinning quality. For the convenience of data collection, no additional collection equipment is installed at the feed end of the thin-film evaporator. In this embodiment, the production parameters of the pulper are directly incorporated during the data collection stage. The production parameters of the pulper also include the equipment parameters of the pulper and the in-process product parameters of the equipment.

[0270] Specifically, the equipment parameters of the pulper collected in this embodiment include the discharge flow rate, and the in-process product parameters of the pulper include the slurry concentration and the slurry temperature. The effects of the foregoing parameters on the operating state of the thin-film evaporator are as follows:

[0271] Discharge flow rate (m³ / h) - The discharge flow rate of the pulper reflects the discharge stability of the pulper and affects the continuity of the subsequent process.

[0272] Slurry concentration (%) - Affects the dissolution uniformity. Too high may cause blockages.

[0273] Slurry temperature (°C) - Affects the feed temperature of the thin-film evaporator.

[0274] Finally, the parameters related to the health prediction of the thin-film evaporator collected are shown in Table 1 below.

[0275]

[0276] When collecting the above data, the method of timestamp synchronization is adopted to uniformly add accurate timestamps to all data to ensure the time series alignment of different data sources. It should be noted that since the change in the pulp quality of the pulper does not immediately affect the thin-film evaporator, but there is a lag of about 10 minutes, the parameters of the pulper should be compensated in advance according to the material transportation time (using data about 10 minutes earlier than the thin-film evaporator) to match the operating state of the thin-film evaporator. As shown in Table 2 below.

[0277]

[0278] Step D2: Preprocess the collected original production parameter time series data, including removing outliers and normalizing the data.

[0279] In this embodiment, a combination of physical constraint methods and statistical methods is used to remove outliers. Among them, the physical constraint method refers to setting parameter thresholds based on the operating range of the equipment. For example, the pressure in the evaporation chamber of the thin-film evaporator should be between 400 pa and 600 pa, and values outside this range are regarded as abnormal. The statistical method refers to using the 3σ principle to remove abnormal points based on the distribution of historical data.

[0280] In this embodiment, the purpose of data normalization is to scale the data to a unified range, such as [0, 1], to eliminate the influence of different dimensions. The normalization process is a well-known technology in this field and will not be elaborated in detail here.

[0281] After the data in the original production parameter time series data table is processed by the above normalization formula, the normalized production parameter data table as shown in Table 3 below is obtained.

[0282]

[0283] Step D3: Combining expert knowledge and experience, assign weights to each collected parameter from three dimensions: the degree of production parameter fluctuation, the degree of influence of production parameters on the equipment health status, and the severity of parameter deviation from the normal range, so as to further improve the accuracy of the prediction results.

[0284] The degree of production parameter fluctuation is determined according to the following method:

[0285] Calculate the standard deviation ( σ ) of each production parameter. The larger the standard deviation, the greater the fluctuation of the production parameter and the stronger the possible impact on the equipment state.

[0286] To ensure that the sum of the volatility contribution weights of all production parameters is 1, the normalized standard deviation is used to measure the parameter fluctuation intensity, and the calculation formula is as follows:

[0287] ;

[0288] In the formula, represents the normalized standard deviation of the i th production parameter; represents the standard deviation of the i th production parameter, is the sum of the standard deviations of all production parameters.

[0289] The degree of influence of production parameters on the equipment health status is determined according to the following method:

[0290] In this embodiment, the degree of influence of production parameters on the equipment health status is measured by expert scoring. Combining expert knowledge and experience, importance ranking and scoring (0 - 10 points) are carried out for each parameter. The higher the score, the greater the influence of the parameter on the equipment health status.

[0291] The importance ranking and scores of the above production parameters can be referred to Table 4 - 1 and Table 4 - 2 below.

[0292]

[0293]

[0294] It should be noted that for the parameters related to the pulper equipment in Table 4 - 1 and Table 4 - 2, although the importance rankings of the pulp concentration and feed flow are at the back, according to the physical mechanism analysis and the influence of the actual production process (combining expert knowledge and experience), the set scores are higher than those of the parameters with higher importance rankings.

[0295] Based on the above scores, the normalized expert scores of each parameter are obtained, and the calculation formula is as follows:

[0296] ;

[0297] In the formula, represents the normalized expert score of the i th production parameter; is the expert score of the i th production parameter, is the total score.

[0298] The severity of the parameter deviation from the normal range is determined according to the following method:

[0299] Set the physical constraint weight, considering the severity of the parameter deviation from the normal range, and the calculation formula is as follows:

[0300] ;

[0301] In the formula, represents the iThe severity of deviation of each production parameter from the normal range; For the i number of times that a production parameter exceeds the normal range, and

[0302] ① Determine the final weight of each production parameter under the influence of the foregoing three dimensions.

[0303] The final weight of each production parameter under the influence of considering the foregoing three dimensions is determined according to the following formula:

[0304] ;

[0305] In the formula, represents the weight of the i th production parameter; and , , 。

[0306] In this embodiment, it can be set that 。

[0307] In this step, the method of combining standard deviation + expert experience scoring + physical constraints is adopted to assign weights to each parameter, which can ensure that the importance of key parameters in the prediction model is reasonably reflected. When the data in the early stage is insufficient, more reliance is placed on expert experience to determine the weight values of each dimension (that is, α 1 、 α 2 、 α 3 values); after data accumulation, more reliance can be placed on data-driven parameter weight allocation.

[0308] Step D4: Construct a production parameter feature sequence based on the time series data after weighting each production parameter, and obtain the training sample data for the health prediction model.

[0309] Determine the sliding window size and step length, and convert the above time series data into a sliding window of a fixed length. The sliding window method divides continuous data into subsequences of a fixed length to facilitate the model to learn time dependence. Each input sliding window includes time series features (dissolved liquid temperature, evaporation chamber pressure, etc.) and target labels (normal / fault), and the target label data is obtained from long-term historical data, mainly relying on equipment maintenance records.

[0310] Construct the feature sequence in the above manner to obtain the training sample data for the health prediction model.

[0311] Step D5: Input the obtained training sample data into the constructed LSTM model architecture for training. Use multiple rounds of iteration to optimize the model parameters and adopt an early stopping strategy to prevent overfitting.

[0312] <1> Model Architecture Design

[0313] Since the original data is a continuously collected continuous time series, and the LSTM requires a fixed-length time window as input for training. When training the LSTM, each sample is a (w 3 , F n ) matrix, where w 3 represents the window size, and F n is the total number of features at each time step.

[0314] The model architecture includes:

[0315] Input layer: Accepts time series data with a shape of (time step, number of features).

[0316] LSTM layer: In this embodiment, a single-layer LSTM is used with 64 hidden units to capture long-term dependencies in the time series.

[0317] Fully connected layer: Used to generate prediction results.

[0318] <2> Loss Function and Optimizer

[0319] Classification task: Binary Cross-Entropy loss.

[0320] Loss function:

[0321] ;

[0322] In the formula, is the true fault label, 0 = normal, 1 = fault; is the fault probability predicted by the LSTM model; N 2 is the total number of samples.

[0323] Regression task: Mean Absolute Error (MAE) or Mean Squared Error (MSE).

[0324] Optimizer: Adam optimizer with a learning rate set to 0.001.

[0325] The above-selected loss function in this embodiment is applicable to binary classification problems. When P fault deviates from the true label , the loss will increase, guiding the gradient adjustment.

[0326] <3> Model Training

[0327] Input the training sample data obtained in step D4 into the constructed model architecture, and divide the training sample data into a training set and a validation set at a ratio of 80:20. Through multiple rounds of iteration (set to 100 rounds in this embodiment), optimize the model parameters.

[0328] Specifically, in this embodiment, the early stopping strategy is used to prevent overfitting. The training stop strategy is as follows:

[0329] The training loss converges, that is, when the change in the training loss for consecutive multiple rounds (set to 10 rounds in this embodiment) is less than a certain threshold ϵ (the threshold is set to 10 in this embodiment) -4 , then stop the training. The formula is as follows:

[0330] ;

[0331] where, represents the training loss of this round, is the training loss of the previous round.

[0332] When the training loss tends to be stable, it indicates that the LSTM has learned the optimal parameters and no longer significantly optimizes.

[0333] Step D6, finally, use evaluation metrics to evaluate and verify the model accuracy, and finally obtain a trained prediction model.

[0334] In this embodiment, the model evaluation metrics can adopt precision, recall, F1 score, mean square error (MSE), or mean absolute error (MAE), etc. The model verification method can adopt time series cross-validation or test the model performance with the actual operation data of the device.

[0335] After constructing and training a health prediction model based on the above steps, the health prediction sub-module of the thin film evaporator can predict the health degree of the target thin film evaporator based on the trained health prediction model. The specific process is as follows:

[0336] Obtain the real-time production parameters of the target thin film evaporator and the real-time production parameters of the upstream pulper (the collected real-time production parameters are the same as those in Table 1), then use the same data processing method as in the model training stage of steps D1~D4 to process the collected real-time data, and then input the processed data into the trained health prediction model. Finally, the current health index of the target thin film evaporator can be obtained.

[0337] LSTM processes data through multiple gates (forget gate, input gate, output gate), and finally calculates the fault probability at the output layer Pfault The output of the last time step of the LSTM is the hidden state ℎ 𝑇 Then, the fault probability is calculated through a Dense Layer:

[0338] P fault = σ ( W h* h T + b h );

[0339] In the formula, W h and b h are the parameters learned during LSTM training; h T is the hidden state after the LSTM processes the last time step; σ ( W h* h T + b h ) represents the Sigmoid function.

[0340] In this embodiment, the Sigmoid function is selected as the activation function because the output range of this function is (0, 1), so it is applicable to binary classification problems and represents the probability of equipment failure.

[0341] In this embodiment, the fault probability of the target thin-film evaporator is output through the LSTM model. Then, the health prediction sub-module will calculate the health index of the target thin-film evaporator based on the fault probability output by the model. If the health index is lower than the set safety threshold (in this embodiment, the safety threshold is set to 0.3), it is determined that there is a potential fault in the equipment. The health index formula is as follows:

[0342] ;

[0343] Among them, is the health index of the thin-film evaporator; is the predicted fault probability of the thin-film evaporator;

[0344] If >0.7 (more than 70%), it indicates that the equipment is highly likely to fail and needs maintenance;

[0345] If <0.3 (less than 30%), it indicates that the equipment is in good health;

[0346] If rises gradually, indicating that the device is gradually degrading, falls gradually.

[0347] Based on the above fault probability prediction values and health index, the health prediction sub-module of the thin film evaporator outputs the fault probability prediction values and health index to the warning module. The warning module determines whether to trigger a warning message based on the health index of the device. For example, when the warning module determines that the health index of the current target thin film evaporator is lower than the threshold value, in the device model corresponding to the target thin film evaporator in the 3D digital twin model, it is displayed in red and the predicted health index value is synchronously displayed to achieve the warning effect; if it is determined that the health index of the current target thin film evaporator is higher than the threshold value, it is displayed in green in the device model corresponding to the target thin film evaporator in the 3D digital twin model.

[0348] It should be noted that the health index only reflects the current operating state of the device and does not indicate a specific fault.

[0349] Finally, the warning module can also output the current device state and guiding maintenance suggestions on the operation display interface according to the health index output by the health prediction model, referring to the fault warning and maintenance suggestions, as shown in Table 5 below.

[0350]

[0351] The following uses a comparative example to illustrate the health prediction effect of the health prediction sub-module of the present invention on the thin film evaporator.

[0352] On a certain day, the DCS system monitored the real-time production parameters of the target thin film evaporator for a period of time as shown in Table 6 below:

[0353] Table 6 Real-time production parameters of the target thin film evaporator

[0354]

[0355] The physical constraint ranges of each real-time production parameter are shown in Table 7 below:

[0356]

[0357] According to the parameter physical constraint ranges in Table 7, the following anomalies occurred in the device-level data in Table 6:

[0358] The current of the stirring motor exceeded the upper limit of 40A;

[0359] The pressure in the evaporation chamber approached the upper limit of 600 Pa;

[0360] The power of the stirring motor exceeded the upper limit of 3.5 KW;

[0361] The discharge flow rate exceeds the limit of 30 m³ / h.

[0362] Meanwhile, the DCS system gives a warning message, suggesting that there may be an overload problem in the current mixing system and recommending shutdown for inspection; the vacuum in the evaporation system is too high, which may pose a leakage risk; the discharge flow rate is too high, and there may be problems with pump transportation.

[0363] At this time, analyze the process-level data:

[0364] The slurry concentration is within the normal range of 8 - 12%, indicating that the slurry concentration is stably controlled without being too thick or too thin. And there is no problem of excessive solvent volatilization, ruling out the risk of vacuum system leakage.

[0365] The solution viscosity is within the normal range of 4.5 - 6.5 Pa·s, indicating that the mixing load is normal and no additional maintenance is required.

[0366] The feed flow rate is within the normal range of 40 - 50 m³ / h, indicating stable feeding and ruling out abnormal feeding.

[0367] The solvent recovery rate is within the normal range of 98 - 99%, indicating that the solvent recovery system is operating normally without obvious leakage or excessive volatilization. The pipeline sealing is good, and the solvent loss is controlled within a reasonable range.

[0368] The pressure of the feed pipeline is within the normal range of 48000 - 56000 Pa, indicating normal fluid transportation without blockage or leakage. There is no need to adjust the flow rate or check the pipeline, and the system is operating smoothly.

[0369] The slurry temperature is within the normal range of 55 - 65°C, indicating stable temperature, suitable for the evaporation and dissolution process, and will not cause insufficient dissolution of the slurry.

[0370] Input the real-time production parameters of the target thin-film evaporator in Table 6 above into the health prediction model constructed and trained in this embodiment, and the LSTM outputs the fault probability P fault is 0.18, and the equipment health index HI is 0.82. The corresponding maintenance suggestion is to maintain the current operation and monitor key parameters.

[0371] Meanwhile, according to the daily maintenance records, the operation and maintenance team arrived at the scene for fault troubleshooting, and the result was that no abnormalities were found. And 10 minutes after the fault alarm, the current of the mixing motor dropped to 36A, the pressure dropped back to 550 Pa, and the mixing power recovered to 3.2kW, all returning to the normal range. Through the integration of process-level data analysis, it was confirmed that the short-term fluctuations in equipment-level parameters were caused by feed adjustment and there was no need for shutdown and repair. This verified the correctness of the prediction results of this solution.

[0372] Example 4

[0373] Based on Embodiment 1, this embodiment further explains and illustrates the energy consumption prediction module. In this embodiment, the energy consumption prediction module mainly realizes the prediction of the energy consumption of the key glue-making process in Lyocell fiber production. Figure 5 This is the data processing architecture diagram of the energy consumption prediction module of the present invention. The specific working process of the energy consumption prediction module for energy consumption prediction is as follows:

[0374] Step E1: The process sheet receiving unit in the energy consumption prediction module obtains the process sheet corresponding to the glue-making stage through the process management module of the system, and then transmits the obtained process sheet to the process sheet data extraction unit;

[0375] Step E2: The process sheet data extraction unit extracts the corresponding porridge weight, process parameter group 1, process parameter group 2, and process parameter group 3 according to the current process sheet; then inputs the current porridge weight and process parameter group 1 into the internally stored and trained glue weight calculation model to comprehensively obtain the glue weight; inputs process parameter group 2 into the internally stored and trained glue refractive index calculation model to calculate the glue refractive index; and inputs the current porridge weight and process parameter group 3 into the internally stored and trained glue-making consumption time calculation model to comprehensively obtain the glue-making consumption time;

[0376] Step E3: The process sheet data extraction unit inputs the obtained porridge weight, glue weight, glue refractive index, and glue-making consumption time into the energy consumption prediction model of the energy consumption prediction unit, and the energy consumption prediction model outputs the energy consumption prediction value corresponding to the current glue-making process sheet.

[0377] In the embodiment described in the present invention, the energy consumption data is the sum of the electricity energy consumption and the steam energy consumption; the specific calculation formula for converting the steam energy consumption into electricity energy consumption is as follows:

[0378] ;

[0379] In the formula, is the electricity energy consumption of the steam; is the price of the steam, with the unit of yuan / m 3 ; is the price of the electric energy, with the unit of yuan / KW·h; is the consumption of the steam, with the unit of m 3 .

[0380] In the embodiments depicted by the present invention, the process parameter set 1 includes the motor speed of the thin-film evaporator, the temperatures of the four zones of the thin-film evaporator, the pressure of the vacuum system, and the rotational speed of the vacuum pump; the process parameter set 2 includes the motor speed of the thin-film evaporator, the temperatures of the four zones of the thin-film evaporator, the rotational speed of the vacuum pump, and the outlet temperature of the heat preservation water circulation pump; the process parameter set 3 includes the motor speed of the thin-film evaporator and the rotational speed of the vacuum pump.

[0381] After the above steps E1 to E2, the porridge weight, the glue solution weight, the refractive index of the glue solution, and the glue-making consumption time corresponding to the current glue-making process can be obtained. Based on these four indicators, finally, the energy consumption prediction result corresponding to the current glue-making process can be output through the pre-trained energy consumption prediction model stored in the energy consumption prediction unit.

[0382] Example 5

[0383] In the production and manufacturing of Lyocell, the raw material cost and the equipment maintenance cost are the key factors affecting the enterprise's profit rate and market competitiveness. Among them, the fluctuations in the raw material cost caused by the raw material consumption data involved in the raw material cost and the equipment maintenance cost have led to excessive fluctuations in the process consumption cost. If the key factors such as the raw material cost and the equipment maintenance cost cannot be accurately predicted, and because of the excessive fluctuations in the process consumption cost, the environmental protection cost in the process consumption cost, such as the recycling cost, cannot be confirmed, making it impossible for the enterprise to conduct refined management of the costs involved in the entire process of production and manufacturing.

[0384] Therefore, based on any of the above embodiments, in this embodiment, a cost prediction module is further proposed in the intelligent simulation system to provide a scientific basis for process optimization and resource scheduling, and assist the enterprise in conducting refined cost management in the process of Lyocell production and manufacturing.

[0385] The cost prediction module of this embodiment is composed of a raw material cost prediction sub-module and an equipment maintenance cost prediction sub-module. Figure 6 For the data processing architecture diagram of the cost prediction module, the methods for the two sub-modules to achieve cost prediction are similar. This embodiment takes the raw material cost prediction sub-module as an example to introduce the working process of the raw material cost prediction sub-module.

[0386] For the prediction of the raw material cost in the process of Lyocell fiber production, the specific steps are as follows:

[0387] Step F1, obtain the raw material data in the current Lyocell production and manufacturing process, including wood pulp parameter data, NMMO solution parameter data, and chemical additive parameter data; the wood pulp parameter data includes wood pulp viscosity, wood pulp polymerization degree, and wood pulp methyl cellulose content; the NMMO solution parameter data includes NMMO solution concentration; the chemical additive parameter data includes propylene glycol content and hydroxylamine content.

[0388] In step F1, the raw material data also includes raw material purchase price data, raw material consumption data, and raw material inventory data;

[0389] It can be understood that, for the above raw material data, the wood pulp viscosity, wood pulp polymerization degree and wood pulp methyl fiber content are monitored and collected in real time through online viscometers and optical measuring instruments; the NMMO solution concentration is collected by a liquid phase concentration detection sensor; the propylene glycol content and hydroxylamine content are detected and collected by chemical sensors; the raw material purchase price data, raw material inventory data, and raw material consumption data are collected through the enterprise resource planning system and synchronized with the external market database.

[0390] Step F2: preprocess the raw material data. The specific preprocessing process is as follows:

[0391] Step a: Detect outliers in the raw material data and remove them; then process missing values ​​and remove duplicate data; for the outlier check of numerical data in the raw material data, 3 σ The principle is to detect outliers in the data; for outlier detection of non-normally distributed data in raw material data, the interquartile range method is used to detect outliers and calculate the first quartile Q 1 and the third quartile Q 3. The set exception range is as follows:

[0392] ;

[0393] ;

[0394] Data outside this range are considered outliers and are removed; where: IQR The meaning of is the interquartile range;

[0395] Furthermore, for the processing of missing values ​​of raw material data, the specific processing method is as follows: time series data is filled by linear interpolation method; if the missing ratio of numerical data is less than 20%, k-nearest neighbor interpolation is used to fill it, if the missing ratio exceeds 20%, the sample is eliminated; if the missing ratio of non-normal distribution data is less than 20%, k-nearest neighbor interpolation is used to fill it, if the missing ratio exceeds 20%, the sample is eliminated.

[0396] Step b: Use Min-Max to normalize the raw material data. The calculation method is as follows:

[0397] ;

[0398] In the formula, x norm is the normalized value; x原始 is the original data; x min is the minimum value of the original data; x max is the maximum value of the original data;

[0399] For the data with abnormal distribution, Z-score standardization is adopted, and the calculation method is as follows:

[0400] ;

[0401] In the formula, μ is the mean value of the original data; σ is the standard deviation, x scaled is the data value after Z-score standardization.

[0402] Step c: Use Pearson correlation analysis and principal component analysis to extract key features from the raw material data, specifically as follows:

[0403] Step c1: Calculate the correlation features between data of the same category and different categories, analyze the linear correlation between variables, and then screen out redundant features. Calculate the Pearson correlation coefficient between features. Specifically, calculate the Pearson correlation coefficient between the following features: Inside the raw material data, analyze the correlation between the viscosity of wood pulp, the degree of polymerization of wood pulp, and the methyl cellulose content of wood pulp to judge their linear dependence relationship in the dataset; at the same time, analyze the correlation between the NMMO solution concentration, the content of propylene glycol, and the content of hydroxylamine to evaluate the influence of chemical additives on the production process; in addition, it is also necessary to calculate the correlation between the inventory data of raw materials, the purchase price data of raw materials, and the consumption data of raw materials to analyze the impact of inventory management on costs.

[0404] The specific calculation method is as follows:

[0405] ;

[0406] In the formula, r is the Pearson correlation coefficient, h i is the variable h the i th observation value of H i is the variable H the i th observation value of is the mean value of the variable h and is the mean value of the variable H ;

[0407] When ∣ rWhen ∣>0.8, it is determined that the two features are highly correlated, and one of them is removed;

[0408] Step c2: Perform principal component analysis for dimensionality reduction. First, construct a standardized data matrix G , with dimensions m G × n G , where m G is the number of samples, n G is the number of features, and then calculate the covariance matrix V , and the calculation method is as follows:

[0409] ;

[0410] In the formula, V is the covariance matrix; G T is the transpose matrix of the standardized data matrix G ;

[0411] Step c3: Solve the eigenvalues and eigenvectors of the covariance matrix V . First, calculate the eigenvalues V 1, λ 1, λ 2,..., λn , sort them in descending order of eigenvalues, and calculate the cumulative contribution rate; select the first q eigenvectors whose cumulative contribution rate exceeds 95% to form a dimensionality reduction projection matrix W . Through the dimensionality reduction projection matrix W , obtain the projection data. Finally, reduce the dimension of the data after principal component analysis dimensionality reduction from n G dimensions to q dimensions; The reason for choosing 95% as the cumulative contribution rate threshold is that a good balance can be achieved between information retention and dimensionality reduction effect. If the cumulative contribution rate is too low, it may lead to the loss of important information and affect the accuracy of data analysis; if it is too high, the dimensionality reduction effect is limited and the computational complexity cannot be effectively reduced; Therefore, 95% as an empirical threshold can ensure that most information is retained, while significantly reducing the data dimension, improving the generalization ability and computational efficiency of the model; The specific formula for obtaining the projection data is as follows:

[0412] ;

[0413] In the formula, W is the dimensionality reduction projection matrix, G′ is the feature data after dimensionality reduction.

[0414] Step d: Use the K-means clustering analysis and information entropy screening method to finally screen the raw material data, extract key features, and remove redundant data to improve the computational efficiency and prediction accuracy of the model. For the raw material data, screen the wood pulp viscosity, wood pulp polymerization degree, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, raw material purchase price data, raw material inventory data, and raw material consumption data to determine which raw material parameters have the greatest impact on production costs and energy consumption.

[0415] The specific process is as follows:

[0416] Step d1: Use the K-means clustering analysis. First, perform initialization, set the number of clusters l = 5, select l random samples as the initial cluster centers, and then calculate the Euclidean distance from the samples to the cluster centers. The calculation method is as follows:

[0417] ;

[0418] In the formula, d ( x , c ) represents the Euclidean distance between the sample point x and the cluster center c , that is, it measures the similarity between the data point and the cluster center. The smaller the distance, the closer the data point is to the cluster center; x i is the i th data sample point, C i is the new cluster center of the i th cluster, is n dimensional space;

[0419] Step d2: Reassign the clusters, calculate the distances from each sample point to all cluster centers, and assign the sample points to the nearest cluster;

[0420] Step d3: Update the cluster centers. First, calculate the new cluster centers and continue to iterate until convergence. When the change in the cluster centers is less than the set threshold, stop the iteration; the calculation method for calculating the new cluster centers is:

[0421] ;

[0422] In the formula, C i is the new cluster center of the i th cluster, ∣ C i ∣ is the number of sample points in the cluster C i ;

[0423] Step d4, filter out representative data, calculate the average distance from each cluster to the cluster center, and select 10% of the data around the cluster center in each cluster as training samples;

[0424] Step d5: Although the most representative samples have been selected through the intra-cluster distance in step d4, the importance and information content of different clusters may still vary greatly; some clusters may contain too much redundant information, while some clusters may contribute more to the data distribution; therefore, it is necessary to further evaluate the information content of each cluster through information entropy screening and calculate the information entropy of each cluster. H ( C i ), calculated as follows:

[0425] ;

[0426] In the formula, P B Cluster C i Medium Category B The probability of M is the number of categories within the cluster;

[0427] Step d6: Filter the cluster with the highest information entropy, calculate the information entropy values ​​of all clusters, and select the top 50% clusters with the highest information entropy. Among the selected clusters, retain the samples around the center of the cluster and remove outliers; finally, obtain the preprocessed raw material data.

[0428] Step F3: input the raw material data preprocessed in step F2 into the trained raw material consumption prediction model, and output the raw material consumption prediction values ​​corresponding to different machine learning algorithm models.

[0429] The overall raw material consumption prediction model of step F3 is composed of a first raw material consumption prediction sub-model based on random forest, a second raw material consumption prediction sub-model based on support vector machine, and a third raw material consumption prediction sub-model based on long short-term memory network.

[0430] Step F4: weighted fusion of the predicted values ​​of each machine learning algorithm model to obtain the final raw material consumption prediction value, which is calculated as follows:

[0431] ;

[0432] In the formula, Indicates the final forecast value of raw material consumption; represents the raw material consumption forecast value of the first raw material consumption forecast sub-model; represents the raw material consumption forecast value of the second raw material consumption forecast sub-model; Represents the predicted value of raw material consumption of the third raw material consumption prediction sub - model; The meaning of is the contribution of the first raw material consumption prediction sub - model in weighted fusion, that is, the weight value of the first raw material consumption prediction sub - model; The meaning of is the contribution of the second raw material consumption prediction sub - model in weighted fusion, that is, the weight value of the second raw material consumption prediction sub - model; The meaning of is the contribution of the third raw material consumption prediction sub - model in weighted fusion, that is, the weight value of the third raw material consumption prediction sub - model;

[0433] Step F5, finally, combine the raw material price and the final predicted value of raw material consumption to obtain the predicted cost of raw materials, and the calculation method is as follows:

[0434] ;

[0435] In the formula, Is the predicted value of the total cost of raw materials; Is the final predicted value of wood pulp raw material consumption; Is the unit price of wood pulp raw material; Is the final predicted value of NMMO solution raw material consumption; Is the unit price of NMMO solution raw material; Is the final predicted value of propylene glycol raw material consumption; Is the unit price of propylene glycol raw material; Is the final consumption value of hydroxylamine raw material; Is the unit price of hydroxylamine raw material.

[0436] In the present invention, for the prediction of raw material consumption, it includes the final predicted value of wood pulp consumption, the final predicted value of NMMO solution consumption, the final predicted value of propylene glycol consumption, and the final predicted value of hydroxylamine consumption.

[0437] Therefore, based on the above prediction process, it can be known that the predicted values of raw material consumption finally predicted by the present invention include the predicted value of wood pulp consumption, the predicted value of NMMO solution consumption, and the predicted value of chemical auxiliary consumption; the predicted value of chemical auxiliary consumption includes the predicted value of propylene glycol consumption and the predicted value of hydroxylamine consumption.

[0438] The predicted value of the total cost of raw materials finally predicted by the present invention is composed of the predicted value of wood pulp raw material cost, the predicted value of NMMO solution raw material cost, the predicted value of propylene glycol raw material cost, and the predicted value of hydroxylamine raw material cost.

[0439] In the embodiment depicted in the present invention, the construction and training processes of the above three raw material consumption prediction sub - models will be described below, specifically as follows:

[0440] (I) Construct a first raw material consumption prediction sub-model for raw material consumption prediction based on random forest, and then collect historical raw material data in the production and manufacturing process of Lyocell, including wood pulp viscosity, wood pulp degree of polymerization, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw materials, consumption data of raw materials, consumption of wood pulp, consumption of NMMO solution, consumption of propylene glycol, and consumption of hydroxylamine; use the above collected data as the data set of the first raw material consumption prediction sub-model, perform data processing in the same manner as in step F2 above, divide the processed data set into a training set, a test set, and a validation set, train, test, and validate the constructed first raw material consumption prediction sub-model, and finally obtain a trained first raw material consumption prediction sub-model; among them, when training the model, wood pulp viscosity, wood pulp degree of polymerization, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw materials, and consumption data of raw materials are used as input features of the training set, and the consumption of wood pulp, the consumption of NMMO solution, the consumption of propylene glycol, and the consumption of hydroxylamine are used as output features of the training set.

[0441] (II) Construct a second raw material consumption prediction sub-model for raw material consumption prediction based on support vector machine. For the collected historical wood pulp viscosity, wood pulp degree of polymerization, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw materials, consumption data of raw materials, consumption of wood pulp, consumption of NMMO solution, consumption of propylene glycol, and consumption of hydroxylamine, perform data processing in the same manner as in step F2, divide the processed data set into a training set, a test set, and a validation set, train, test, and validate the constructed second raw material consumption prediction sub-model, and finally obtain a trained second raw material consumption prediction sub-model; among them, when training the model, wood pulp viscosity, wood pulp degree of polymerization, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw materials, and consumption data of raw materials are used as input features of the training set, and the consumption of wood pulp, the consumption of NMMO solution, the consumption of propylene glycol, and the consumption of hydroxylamine are used as output features of the training set.

[0442] (III) Construct a third raw material consumption prediction sub-model for raw material consumption prediction based on the long short-term memory network. For the collected historical data of pulp viscosity, pulp polymerization degree, pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw material warehouses, raw material consumption data, pulp consumption, NMMO solution consumption, propylene glycol consumption, and hydroxylamine consumption, data processing is carried out in the same manner as in step F2. The processed data set is divided into a training set, a test set, and a validation set. The constructed third raw material consumption prediction sub-model is trained, tested, and validated, and finally a trained third raw material consumption prediction sub-model is obtained. Among them, when training the model, pulp viscosity, pulp polymerization degree, pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, purchase price of raw materials, inventory data of raw materials, and raw material consumption data are used as input features of the training set, and pulp consumption, NMMO solution consumption, propylene glycol consumption, and hydroxylamine consumption are used as output features of the training set.

[0443] After the above steps (I)-(III), finally, a trained first raw material consumption prediction sub-model, a second raw material consumption prediction sub-model, and a third raw material consumption prediction sub-model can be obtained. The three trained sub-models can be used for raw material consumption prediction.

[0444] It should be noted that for the first raw material consumption prediction sub-model, during the training process, in order to prevent overfitting and underfitting, according to experience, the random forest model consists of 100 decision trees, and the maximum depth of each tree is set to 15 to prevent overfitting. In the training stage, first, the Bootstrap sampling method is used to randomly extract sub-samples from the training data set, that is, m samples are randomly and with replacement extracted from the original data set as the training data set for each tree. Each sample may be extracted multiple times or not selected, thus forming multiple different data subsets. The main purpose of doing this is to improve the generalization ability of the model, reduce overfitting, and enhance the stability of the model. In addition, when each decision tree splits nodes, randomly select some features for decision-making to ensure the diversity of the model, further reduce the dependence on a single feature, and improve the overall prediction effect.

[0445] Moreover, for regression tasks, such as raw material consumption prediction, energy consumption prediction, or equipment maintenance prediction, the mean squared error (MSE) is used as the loss function to minimize the error between the predicted value and the true value.

[0446] The final output result of the random forest is calculated through the integration of multiple decision trees to obtain the final predicted value, and the final predicted value is the weighted average of all decision trees.

[0447] For the second raw material consumption prediction sub-model, during the training process, the optimization objective of support vector regression is to minimize the ε-insensitive loss function, that is, through the combination of the regularization term and the error penalty term, to ensure the generalization ability of the model.

[0448] For the third raw material consumption prediction sub-model, during the training process, LSTM inputs the training data into the model for training. The LSTM model structure includes an LSTM layer, a fully connected layer, an optimizer, and a loss function. Among them, the LSTM layer contains 64 LSTM units, and the Tanh activation function is used to enhance the non-linear expression ability; the fully connected layer includes two layers of neural networks, consisting of 128 neurons and 64 neurons respectively, and the ReLU activation function is used for non-linear transformation to improve the model fitting ability. During the training process, each batch contains 32 samples (Batch Size = 32); the forward propagation is used to calculate the output. LSTM processes the input data step by step through time steps, and finally the fully connected layer calculates the predicted value.

[0449] Furthermore, for step F4, the determination method of the weighted fusion value of each machine learning algorithm model is as follows:

[0450] Assume that the weights of the three prediction sub-models included in the total raw material prediction model follow Beta distribution:

[0451] ;

[0452] In the formula, is the weight coefficient of each prediction sub-model included in the total raw material consumption prediction model; α 0 and β 0 are set as initial parameters, generally taking:

[0453] ;

[0454] In the formula, C 0 is a positive number;

[0455] According to the mean square errors of the first raw material consumption prediction sub-model, the second raw material consumption prediction sub-model, and the third raw material consumption prediction sub-model, the prediction errors of the three raw material consumption prediction sub-models are obtained, and the calculation method is as follows:

[0456] ;

[0457] ;

[0458] ;

[0459] In the formula, represents the mean square error of the first raw material consumption prediction sub-model; represents the mean square error of the second raw material consumption prediction sub-model; represents the mean square error of the third raw material consumption prediction sub-model; N 3 represents the total number of training samples of the first raw material consumption prediction sub-model, the second raw material consumption prediction sub-model, and the third raw material consumption prediction sub-model; represents the i true value of the th raw material data sample; i represents the predicted value of the first raw material consumption prediction sub-model for the th raw material data sample; i represents the predicted value of the second raw material consumption prediction sub-model for the th raw material data sample; i represents the predicted value of the third raw material consumption prediction sub-model for the

[0460] Then, update the Beta distribution according to the mean square error:

[0461] ;

[0462] ;

[0463] In the formula, represents a distribution parameter in the calculation of the fusion weight of the total raw material consumption prediction model; represents another distribution parameter in the calculation of the fusion weight of the total raw material consumption prediction model; represents the mean square error of the total raw material consumption prediction model;

[0464] Introduce Gaussian process regression to calculate the uncertainty of the total raw material consumption prediction model, and the calculation method is as follows:

[0465] ;

[0466] ;

[0467] In the formula, means the uncertainty of the total raw material consumption prediction model; means the prediction variance of the total raw material consumption prediction model on all samples; represents the predicted value of the total raw material consumption prediction model for the i th raw material data sample; represents the average predicted value of the total raw material consumption prediction model for all raw material data samples;

[0468] Introduce the uncertainty of the total raw material consumption prediction model into the update Beta distribution:

[0469] ;

[0470] ;

[0471] In the formula, represents a distribution parameter in the calculation of the fusion weight of the updated total raw material consumption prediction model; represents another distribution parameter in the calculation of the fusion weight of the updated total raw material consumption prediction model;

[0472] Finally, use variational Bayesian inference to calculate the final weight of the total raw material consumption prediction model:

[0473] ;

[0474] In the formula, is the model weight after the t +1 round of iteration of the total raw material consumption prediction model; represents the contribution of the actual data;

[0475] Then, calculate the fusion weights of the three raw material consumption prediction sub-models respectively:

[0476] ;

[0477] ;

[0478] ;

[0479] In the formula, is a distribution parameter in the calculation of the fusion weight of the updated first raw material consumption prediction sub-model; is another distribution parameter in the calculation of the fusion weight of the updated first raw material consumption prediction sub-model; is a distribution parameter in the calculation of the fusion weight of the updated second raw material consumption prediction sub-model; is another distribution parameter in the calculation of the fusion weight of the updated second raw material consumption prediction sub-model; is a distribution parameter in the calculation of the fusion weight of the updated third raw material consumption prediction sub-model; is another distribution parameter in the calculation of the fusion weight of the updated third raw material consumption prediction sub-model.

[0480] In the embodiments described in the present invention, for the prediction of raw material costs, the time series data includes the inventory data of raw materials and the consumption data of raw materials;

[0481] The numerical data includes wood pulp viscosity, wood pulp polymerization degree, wood pulp methyl cellulose content, NMMO solution concentration, propylene glycol content, hydroxylamine content, and the purchase price data of raw materials;

[0482] The non-normal distribution data includes the purchase price data of raw materials, the inventory data of raw materials, and the consumption data of raw materials.

[0483] In the embodiments described in the present invention, the raw material data can be obtained through the material management module of the system.

[0484] Example 6

[0485] On the basis of Example 5, the cost prediction module of this embodiment further includes an equipment maintenance cost prediction sub-module. The equipment maintenance cost prediction sub-module is used to predict the equipment maintenance costs involved in Lyocell fiber production. The equipment maintenance cost prediction sub-module is also composed of a first equipment maintenance prediction sub-model constructed based on a random forest, a second equipment maintenance prediction sub-model constructed based on a support vector machine, and a third equipment maintenance prediction sub-model constructed based on a long short-term memory network. The specific data processing methods and model construction and training methods of each sub-model are the same as those of the three prediction sub-models in the raw material cost prediction sub-module, and will not be elaborated in detail here.

[0486] The difference lies in the data used. Specifically, for the prediction of equipment maintenance costs, when training the three equipment maintenance prediction sub-models, the equipment maintenance data collected and used includes equipment operation time, maintenance cost, number of failures, equipment failure rate, equipment vibration data, and equipment maintenance history. Among them, the equipment operation time, maintenance cost, number of failures, equipment failure rate, equipment vibration data, and equipment maintenance history data are used as input features of the model training set during the training of the equipment maintenance prediction sub-model, and the equipment failure rate and number of maintenance times are used as output features of the model training set.

[0487] In the embodiments described in the present invention, after the training of the equipment maintenance prediction sub-model is completed, the equipment operation time, maintenance cost, number of failures, equipment failure rate, equipment vibration data, and equipment maintenance history during the production process are collected. After processing the above data using the same preprocessing method as in step F2, the data is input into each trained equipment maintenance prediction sub-model. Finally, each equipment maintenance prediction sub-model outputs the equipment failure rate and the number of maintenance times as the prediction values of the model. Then, the prediction values of the three equipment maintenance prediction sub-models are fused in the same way of weighted fusion to obtain the final equipment maintenance prediction value. The weight values of the three equipment maintenance prediction sub-models can be determined by referring to the method of the three raw material consumption prediction sub-models.

[0488] Further, for the equipment maintenance prediction sub-module, the time series data includes the equipment operation time and the number of failures; the numerical data includes the maintenance cost and the equipment failure rate; the non-normal distribution data includes the maintenance cost, the number of failures, the equipment failure rate, the equipment vibration data, and the equipment maintenance history;

[0489] For the key feature extraction of equipment maintenance data using Pearson correlation analysis and principal component analysis, calculate the correlation between the equipment operation time, the number of failures, and the equipment failure rate to evaluate the impact of equipment aging on the failure rate. At the same time, analyze the relationship between the maintenance cost and the equipment operation time and the number of failures to optimize the maintenance budget and repair plan.

[0490] When using the K-means clustering analysis and information entropy screening method to finally screen the equipment maintenance data, in terms of the equipment maintenance data, focus on screening the equipment operation time, the maintenance cost, the number of failures, and the equipment failure rate to analyze the impact of the equipment status on the production efficiency and the maintenance cost.

[0491] Finally, the final equipment maintenance cost prediction value can be calculated through the following formula:

[0492] ;

[0493] In the formula, represents the final equipment maintenance prediction value; represents the equipment maintenance prediction value of the first equipment maintenance prediction sub-model; represents the equipment maintenance prediction value of the second equipment maintenance prediction sub-model; represents the equipment maintenance prediction value of the third equipment maintenance prediction sub-model; The meaning of The meaning is the contribution of the second device maintenance prediction sub-model in weighted fusion, that is, the weight value of the second device maintenance prediction sub-model; The meaning is the contribution of the third device maintenance prediction sub-model in weighted fusion, that is, the weight value of the third device maintenance prediction sub-model;

[0494] Furthermore, the final device maintenance prediction cost can be calculated through the following formula:

[0495] ;

[0496] C 设备维护 is the predicted value of the final device maintenance cost; C 维护 is the fixed maintenance cost of the device; C 单次故障 is the cost of single fault repair; Y final-故障 is the final number of device failures.

[0497] Embodiment 7

[0498] Based on any of the above embodiments, the intelligent simulation system of this embodiment may further include production operation management modules or production comprehensive management modules such as a user management module, a production management module, a material management module, and a device management module. These modules further improve the intelligent simulation system of the present invention, enhancing the practicality and reliability of the system, enabling it to better meet the actual needs of enterprise production management. By realizing the comprehensive digital management of the production process, production efficiency, product quality, and resource utilization rate are also improved, while production costs and risks are reduced.

[0499] Specifically, the above functional modules will be introduced and described in detail below.

[0500] The user management module is mainly used to create, maintain, and log in user accounts, and set and maintain user account permissions; specifically, the user management module can be composed of the following sub-modules, including:

[0501] The account management sub-module is used for the system administrator to maintain user accounts and create new users; for creating new users, it usually includes operations such as entering user basic information and setting an initial login password.

[0502] The login management module is used to verify the user's login process, including verifying the entered account and password; if the information is correct, login is allowed; if incorrect, prompt to re-enter and record the number of errors. When the number of errors reaches a certain level, the account is locked to enhance system security.

[0503] The permission management module is used to set and maintain permissions for user accounts. When a user attempts to access a function beyond their permissions, the system will immediately block and prompt the user that they have insufficient permissions.

[0504] It can be understood that the user management module can also communicate and interact with other modules of the system. For example, it can transfer the user's identity and permission information to the device management module, production management module, etc. If the current user's position is an inspector, then in the device management module, the user can only view device information related to the inspection task; in the production management module, the user can only view the production plan and cannot perform modification operations, ensuring the standardization of system operations and the security of data.

[0505] The production management module is used to formulate, adjust, approve, query, and issue production plans. The production plans consist of annual production plans, monthly production plans, and daily production plans. The production management module is usually mostly used by the enterprise's production management department. The production management department formulates production plans through the production management module based on market orders and the enterprise's production capacity, clarifying the output targets and production time arrangements for each production line. After the plan is formulated, the approval personnel approve the plan in the system. If the approval is passed, the production management module issues the production plan to each relevant module.

[0506] For example, after receiving the production plan transmitted by the production management module, the material management module calculates the required quantity of raw materials based on the production requirements and raw material consumption quotas in the plan, and arranges raw material procurement and inventory allocation. The device management module reasonably arranges the maintenance and repair time of the devices according to the production plan to ensure the normal operation of the devices during production. For example, it arranges in-depth maintenance of the thin-film evaporator during periods with relatively light production tasks.

[0507] Moreover, the production management module can also communicate with mobile terminal devices. After the production plan is formulated and approved, the formulated production plan is issued to the corresponding staff's mobile terminal devices through the wireless transmission module.

[0508] For the material management module, in the raw material procurement link, the procurement department determines the procurement requirements based on the production plan issued by the production management module and the inventory data of the material management module. After the procurement is completed, the raw materials arrive at the warehouse and are put into storage. The warehouse management staff enters information such as the batch number, quantity, and quality inspection report of the raw materials through the warehousing function of the material management module. The system automatically updates the inventory ledger and increases the inventory quantity.

[0509] Furthermore, when the workshop staff pick up materials according to the production plan in the production workshop, they submit a material requisition application in the material management module, specifying the types, quantities, and uses of the required raw materials. The material management module reviews the application and checks whether the inventory is sufficient. If the inventory meets the demand, the material requisition application is approved, and the inventory ledger is updated to reduce the inventory quantity. If the inventory is insufficient, the information is fed back to the production management module and the warning module.

[0510] Generally speaking, during the production process, the material management module monitors the consumption of materials in real time. By linking with the equipment management module and the production management module, it obtains equipment operation data and production progress data, and calculates the actual consumption rate of materials. When abnormal material consumption (such as too fast or too slow consumption) is detected, the abnormal information is transmitted to the warning module, and through the warning module, relevant production management personnel are notified in a timely manner for inspection and verification to ensure the normal progress of the production process and avoid affecting product quality and production efficiency due to material problems. At the same time, the material management module regularly takes inventory of the inventory, compares the inventory results with the system inventory data, and if there are differences, promptly finds out the reasons and makes adjustments to ensure the accuracy of the inventory data.

[0511] For the warning module of this embodiment, during the production process of Lyocell fiber, the material management module of the system monitors the raw material inventory in real time. When the inventory of a certain raw material (such as wood pulp) is lower than the set safety threshold, the material management module transmits the inventory data to the warning module. After receiving the data, the warning module determines that the inventory is insufficient and triggers a material inventory warning.

[0512] At the same time, when the system is running, the key equipment status prediction module of the system continuously monitors the status of key equipment such as thin film evaporators, and transmits the predicted remaining life or health index data of the target thin film evaporator to the warning module; the warning module determines whether the remaining life or health index of the target thin film evaporator is lower than the safety threshold and achieves a warning effect.

[0513] Generally speaking, the warning module issues alarms in various ways to notify relevant personnel. For example, a prominent identifier pops up on the system interface, a text message is sent to relevant personnel (equipment maintenance personnel, production supervisors, etc.), and a warning is given in the three-dimensional digital twin model of the production line.

[0514] Moreover, it can be understood that on the operation interface of the entire intelligent simulation system, the operator can intuitively view the detailed information of the above-mentioned insufficient material inventory and potential equipment failures, including inventory quantity, name and location of the faulty equipment, etc. After receiving the warning information transmitted by the warning module, the production management personnel can adjust the production plan in a timely manner through the production management module, reduce the dependence on raw materials with insufficient inventory, or suspend relevant production links to avoid production interruption caused by material shortage or equipment failure. Moreover, the equipment management module can arrange maintenance personnel to prepare for maintenance work according to the equipment failure warning, prepare maintenance tools and spare parts in advance to ensure that equipment failures can be handled in a timely manner and maintain the continuity of production.

[0515] Furthermore, for the equipment management module of this embodiment, the equipment management module is generally composed of an equipment information maintenance sub-module, an equipment inspection sub-module, and an equipment repair and maintenance sub-module. Among them,

[0516] In the equipment information maintenance sub-module, the management personnel input the detailed information of newly purchased equipment (such as a new spinning machine), including basic data such as equipment name, model, manufacturer, installation location, technical parameters, etc., and assign a unique number to the equipment.

[0517] The equipment inspection sub-module formulates an inspection plan according to the equipment type, operating conditions, and maintenance requirements. For example, it is stipulated that a comprehensive inspection of the spinning machine is carried out once a week. The inspection plan is sent to the terminal devices of the inspection personnel (such as mobile phones or tablets) through the system. The inspection personnel perform inspection tasks according to the plan, scan the equipment QR code with the terminal device, record information such as inspection time, equipment operating status (such as parameters such as temperature, vibration, pressure, etc.), and whether any abnormalities are found, and upload it to the equipment management module, and finally form an equipment inspection log.

[0518] If equipment abnormalities are found during the inspection process, after receiving the data, the equipment management module transmits the abnormality information to the warning module for warning.

[0519] The equipment repair and maintenance sub-module is used to formulate an equipment repair and maintenance plan and arrange maintenance personnel for repair. After the maintenance personnel complete the repair, they record information such as repair content, replaced parts, and repair time through the equipment management module to generate a repair and maintenance log. Moreover, the equipment management module also communicates with the production management module. When the equipment needs to be repaired and may affect the production plan, it timely notifies the production management module to adjust the production arrangement to ensure the coordinated progress of production and equipment maintenance work. At the same time, the equipment management module provides the operation data and maintenance records of the equipment to the cost prediction module as the basis for equipment maintenance cost prediction.

[0520] Finally, for the process management module of this embodiment, the process engineer prepares a new Lyocell fiber production process document in the process management module, which details information such as the production process, process parameters of each process (such as dissolution temperature, spinning speed, draw ratio, etc.), and quality control standards. After the process document is prepared, it is submitted for approval. After the approval is passed, the process management module publishes the process document to relevant modules such as the production management module, equipment management module, and quality control module.

[0521] At the same time, the process management module is linked with the 3D modeling and digital twin module to simulate the production process of the new process in a virtual environment, discover potential problems in advance, verify the feasibility and stability of the process, and reduce risks in actual production.

[0522] It can be understood that the process management module assists the energy consumption prediction module in realizing the energy consumption prediction of the glue-making process by providing the process sheet of the Lyocell fiber production glue-making process to the energy consumption prediction module.

[0523] The above is only a preferred embodiment of the present invention, and does not impose any form of limitation on the present invention. Any simple modification or equivalent change made to the above embodiments based on the technical essence of the present invention falls within the protection scope of the present invention.

Claims

1. A lyocell fiber production line intelligent simulation system based on digital twin, characterized in that: The system includes: 3D modeling and digital twin module, key equipment status prediction module, energy consumption prediction module and early warning module; 3D modeling and digital twin module, which is used to build a 3D model of the Lyocell fiber production line, associate the real-time production data of the production line with the 3D model, and generate a 3D digital twin model of the Lyocell fiber production line based on digital twin technology; A key equipment status prediction module includes a remaining life prediction submodule for a thin film evaporator, wherein the remaining life prediction submodule for the thin film evaporator obtains real-time vibration data collected by a vibration sensor on a target thin film evaporator and real-time spindle speed transmitted by a central control platform of the target thin film evaporator; the real-time vibration data is segmented according to a set time interval to obtain a plurality of data segments at set time intervals, and then each data segment is analyzed in the time domain and frequency domain to obtain the root mean square, kurtosis and frequency center of gravity of each data segment; the spindle fault characteristic frequency is obtained according to the obtained real-time spindle speed, and then the spindle bearing fault characteristic frequency is obtained according to the real-time spindle speed and the size parameters of the spindle bearing; finally, the obtained root mean square, kurtosis, frequency center of gravity, spindle fault characteristic frequency and spindle bearing fault characteristic frequency are input into a remaining life prediction model, and finally the remaining life prediction result is output to the early warning module; The energy consumption prediction module extracts the porridge weight, process parameter group 1, process parameter group 2 and process parameter group 3 corresponding to the current process list according to the process list of the glue making process provided by the process management module; then obtains the glue liquid weight based on the porridge weight and process parameter group 1, obtains the glue making consumption time based on the porridge weight and process parameter group 3, and obtains the glue liquid refractive index based on process parameter group 2; finally, the porridge weight, glue liquid weight, glue liquid refractive index and glue making consumption time are input into the trained energy consumption prediction model, and the energy consumption prediction value corresponding to the glue making process is output; wherein, the process parameter group 1 includes the motor speed of the thin film evaporator, the temperature of the four zones of the thin film evaporator, the vacuum system pressure and the vacuum pump speed; the process parameter group 2 includes the motor speed of the thin film evaporator, the temperature of the four zones of the thin film evaporator, the vacuum pump speed and the outlet temperature of the insulation water circulation pump; the process parameter group 3 includes the motor speed of the thin film evaporator and the vacuum pump speed; The early warning module is used to issue an early warning in a device model corresponding to the target thin film evaporator in the three-dimensional digital twin model according to the remaining life prediction result of the target thin film evaporator.

2. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The real-time vibration data includes acceleration signals in three directions respectively collected by vibration sensors at the upper and lower ends and the middle area of ​​the target thin film evaporator cylinder; wherein the vibration sensor at the upper end of the cylinder is used to monitor the fault condition of the upper end bearing of the main shaft, the vibration sensor at the lower end of the cylinder is used to monitor the fault condition of the lower end bearing of the main shaft, and the vibration sensor in the middle area of ​​the cylinder is used to monitor the fault condition of the main shaft.

3. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The construction process of the remaining life prediction model of the thin film evaporator is as follows: Obtain historical vibration data and historical spindle speed data of the entire process from the normal operation of the thin film evaporator to the failure of the spindle and / or spindle bearing resulting in the failure of the thin film evaporator to operate normally; The historical vibration data includes acceleration signals in three directions respectively collected by vibration sensors at the upper and lower ends and the middle area of ​​the thin film evaporator cylinder; wherein the vibration sensor at the upper end of the cylinder is used to monitor the fault condition of the upper end bearing of the main shaft, the vibration sensor at the lower end of the cylinder is used to monitor the fault condition of the lower end bearing of the main shaft, and the vibration sensor in the middle area of ​​the cylinder is used to monitor the fault condition of the main shaft; The acceleration signal is segmented at set time intervals to obtain multiple data segments, and then each data segment is analyzed in the time domain and the frequency domain to extract the root mean square, kurtosis and frequency center of gravity of each data segment; the fault characteristic frequency of the spindle is obtained according to the historical spindle speed; the fault characteristic frequency of the spindle bearing is obtained according to the historical spindle speed and the spindle bearing size parameters; After vector addition of acceleration signals in three directions of each vibration sensor, multiple vibration vector data are obtained. A one-dimensional convolutional neural network model including convolution, pooling, activation function and full connection layer is built for each vibration vector data. The response trend of the vibration response of each vibration vector data changing with time is extracted through the one-dimensional convolutional neural network model of each vibration vector data, and the vibration vector data characteristic curves of the upper and lower end bearings and the main shaft are obtained. Multiple high-order exponential functions are used to fit the vibration vector data characteristic curves of the upper and lower end bearings and the main shaft, respectively, to form the remaining life characteristic curves of the upper end bearing of the main shaft, the remaining life characteristic curves of the lower end bearing of the main shaft and the remaining life characteristic curves of the main shaft, and then expressed in percentage form; Taking the set time interval as the time resolution, the data set of the remaining life prediction model is formed based on the root mean square, kurtosis, frequency centroid, fault characteristic frequency of the spindle bearing, fault characteristic frequency of the spindle, and the data on the remaining life characteristic curve of the upper end bearing of the spindle, the remaining life characteristic curve of the lower end bearing of the spindle, and the remaining life characteristic curve of the spindle; wherein the data on the remaining life characteristic curve is the label data; The obtained data set is divided into a training set, a validation set and a test set; the constructed remaining life prediction model is trained, verified and tested respectively using the divided data sets, and finally a trained remaining life prediction model is obtained; the remaining life prediction model is a one-dimensional convolutional neural network model.

4. A lyocell fiber production line intelligent simulation system based on digital twin according to claim 1 or 3, characterized in that: The characteristic frequencies of spindle bearing faults include: Inner race fault characteristic frequency: ； Outer race fault characteristic frequency: ； Rolling element failure characteristic frequency: ； Cage failure characteristic frequency: ； Where: Q is the number of rolling elements, d The diameter of the rolling element, D Bearing pitch diameter, β is the contact angle, f r The rotation frequency of the main axis.

5. A Lyocell fiber production line intelligent simulation system based on digital twin according to claim 1 or 3, characterized in that: The characteristic frequency of spindle fault is: ； Where: n 1 is the multiple number, f r is the rotation frequency of the main axis; is the main axis disturbance frequency.

6. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 3, characterized in that: The calculation formula expressed in percentage form after forming the remaining life characteristic curve of the upper end bearing of the main shaft, the remaining life characteristic curve of the lower end bearing of the main shaft and the remaining life characteristic curve of the main shaft is as follows: ； Where: , , , , is an adjustable parameter; is the horizontal coordinate of the vibration data, i.e., the system time of the vibration data; The time limit for the equipment to go from fully normal to abnormal is defined; is the remaining life percentage; e Is the natural base.

7. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The construction process of the energy consumption prediction model is as follows: Collect the historical raw data of evaporation system parameter data, glue and sealing system parameter data, insulation water and vacuum system parameter data, slurry weight, glue weight, glue refractive index, glue consumption time and energy consumption data during the Lyocell fiber glue making process, and perform data preprocessing to obtain historical data; Calculate the process similarity of the corresponding historical data according to the process sheet of the Lyocell fiber glue making process, and rearrange the historical data according to the process similarity to obtain the historical sorted data; A reference group sequence is selected from the historical sorting data, and the process similarity of each reference group sequence is calculated to obtain the historical sorting data sequence, and the threshold interval of the process similarity is obtained; then, the historical sorting data sequence within the threshold interval is used to calculate the energy consumption prediction value sequence; Compare the energy consumption forecast value sequence with the historical energy consumption data of the corresponding reference group sequence, and record their absolute errors and X; Increase the threshold interval percentage P of process similarity to obtain the threshold interval of the current process similarity until the threshold interval of the current process similarity reaches the set maximum value, and obtain the threshold interval percentage P of process similarity corresponding to the minimum absolute error and X, and mark P as P s ; Build an energy consumption prediction model and use P s The historical screening and sorting data are obtained by screening, and the porridge weight, glue liquid weight, glue liquid refractive index and glue making time in the historical screening and sorting data are used as the data set of the energy consumption prediction model. The energy consumption prediction model is trained using the data set to finally obtain a trained energy consumption prediction model.

8. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: Historical original data include historical original process parameter data, historical original energy consumption data and historical original intermediate parameter data; historical data include historical process parameter data, historical energy consumption data and historical intermediate parameter data; historical sorted data include historical sorted process parameter data, historical sorted intermediate parameter data and historical sorted energy consumption data; historical filtered and sorted data include historical filtered and sorted process parameter data, historical filtered and sorted intermediate parameter data and historical filtered and sorted energy consumption data.

9. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 8, characterized in that: The process sheet includes process parameter data; the process parameter data include evaporation system parameter data, glue liquid and sealing system parameter data, insulation water and vacuum system parameter data and porridge weight; the evaporation system parameter data include evaporation heating system parameter data, thin film evaporator parameter data and evaporation condensation water system parameter data; the glue liquid and sealing system parameter data include sealing liquid system parameter data and glue liquid transportation parameter data; the insulation water and vacuum system parameter data include insulation water system parameter data and vacuum system parameter data.

10. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 9, characterized in that: The evaporation heating system parameter data includes the heating system temperature, the heating system liquid level and the heating system pressure; the thin film evaporator parameter data includes the four-zone temperature of the thin film evaporator, the thin film evaporator motor speed, the thin film evaporator bearing lubricating oil circulation temperature, the reducer bearing speed, the cooling fan speed, the thin film evaporator liquid level, the thin film evaporator outlet pressure and the thin film evaporator bottom outlet temperature; the four-zone temperature of the thin film evaporator includes the jacket temperature of the four evaporation zones and the glue temperature of the four evaporation zones; the glue delivery parameter data includes the temperature of the insulation water heat exchanger in the pump, the flow rate of the insulation water heat exchanger in the pump, and the insulation water heat exchanger outside the pump. The data include the temperature of the sealing liquid tank, the flow rate of the sealing liquid tank and the liquid level of the sealing liquid tank; the data include the outlet temperature of the insulation water circulation pump and the hot water pressure of the insulation system; the data include the temperature of the evaporative condenser, the liquid level of the evaporative condenser, the pressure of the evaporative condenser, the flow rate of the cooling water to the evaporative condenser and the outlet flow rate of the evaporative condensate pump; the data include the vacuum system liquid level, the pressure of the vacuum system and the speed of the vacuum pump.

11. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 8, characterized in that: The intermediate parameter data include glue liquid weight, glue liquid refractive index and glue making time.

12. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: The data preprocessing method is: based on 3 σ In principle, outliers are removed; after outliers are removed, the missing values ​​are filled by taking the weighted average of the K neighboring data values ​​near the missing value.

13. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: The specific method of selecting a reference group from historical sorting data is: Z The row data constitutes the reference group sequence ; By reference group sequence For each sequence in the reference group, we obtain the historical sorting data corresponding to each sequence in the reference group. , thus forming the historical sorting data sequence of the reference group sequence , and record the maximum process similarity value in the process ; Among them, the threshold interval of process similarity is , is the minimum process similarity value.

14. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: Using the historical sorted data sequence within the threshold range, the energy consumption forecast value sequence is calculated. The specific method is as follows: The porridge weight of the historically sorted intermediate parameter data and the historically sorted process parameter data in the historically sorted data sequence of the reference group sequence is used as the training set of the neural network model to obtain the energy consumption prediction value sequence of the reference group sequence; the process similarity corresponding to the historically sorted intermediate parameter data belongs to the threshold interval range of the process similarity.

15. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: The energy consumption prediction value sequence of the reference group sequence is compared with the historical energy consumption data of the corresponding reference group sequence, and their absolute errors and X are recorded. The specific method is: Compare the energy consumption prediction value sequence of the reference group sequence with the historical energy consumption data of the corresponding reference group sequence, record their absolute error values, and traverse the historical sorting data sequence. Z Sorting historical data to obtain the absolute error value sequence of energy consumption prediction , and then accumulate the energy consumption prediction absolute error value sequence to obtain the absolute error and X and record it.

16. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: The porridge weight, glue weight, glue refractive index and consumption time in the historical screening and sorting data are used as the data set of the energy consumption prediction model. The energy consumption prediction model is trained using the data set, and finally a trained energy consumption prediction model is obtained, including: The glue liquid weight, glue liquid refractive index, consumption time and porridge weight in the historical screening and sorting of intermediate parameter data are used as input features of the training set of the energy consumption prediction model, and the energy consumption data of historical screening and sorting are used as the target output of the energy consumption prediction model. The model is trained using the above data to obtain a trained energy consumption prediction model; the energy consumption prediction model uses a three-layer BP neural network.

17. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 7, characterized in that: The energy consumption data is the sum of electricity consumption and steam consumption.

18. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The key equipment status prediction module also includes a health prediction submodule for a thin film evaporator; the health prediction submodule for the thin film evaporator obtains the current production parameters of the target thin film evaporator and the current production parameters of the upstream pulper, and then pre-processes the collected data, including the step of removing outliers and normalization processing; then based on the expert scoring method, the collected parameters are weighted from three dimensions: the degree of parameter fluctuation, the degree of influence of the parameters on the health status of the equipment, and the severity of the deviation of the parameters from the normal range, and the sum of the products of each dimension of each parameter and the corresponding weight is calculated to obtain the influence weight of each parameter in the health prediction; finally, the weighted data of each parameter are input into the trained health prediction model to predict the health of the target thin film evaporator, and the health index of the target thin film evaporator is output; wherein, the real-time production parameters of the thin film evaporator include evaporation chamber pressure, stirring motor current, feed flow rate, feed pipeline pressure, dissolving liquid temperature, solution viscosity and solvent recovery rate; the real-time production parameters of the pulper include discharge flow rate, slurry concentration and slurry temperature.

19. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: When collecting the real-time production parameters of the target thin-film evaporator and the real-time production parameters of the pulper, the parameters are synchronized in time to ensure the timing alignment of different data sources; among them, the parameters of the pulper are compensated in advance according to the material transportation time to match the operating status of the thin-film evaporator.

20. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 19, characterized in that: Outliers are eliminated by combining physical constraint methods and statistical methods. The physical constraint method refers to setting parameter thresholds based on the equipment operating range, and exceeding the range is considered an abnormality; the statistical method refers to eliminating abnormal points based on the distribution of historical data using the 3σ principle.

21. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: The degree of parameter fluctuation is measured by normalized standard deviation, which is calculated as follows: ； In the formula, Indicates i Normalized standard deviation of production parameters; For the i The standard deviation of the production parameters, is the sum of the standard deviations of all production parameters.

22. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: The degree of influence of parameters on the health status of equipment is measured by normalized expert score, which is calculated as follows: ； In the formula, Indicates i Normalized expert scores for production parameters; For the i Expert ratings of production parameters, For the total score.

23. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: The severity of a parameter deviation from the normal range is measured using the physical constraint weight, which is calculated as follows: ； In the formula, Indicates i The severity of the deviation of a production parameter from the normal range; For production parameters i The number of times beyond the normal range, is the total number of samples of production parameters.

24. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: The health index is calculated as follows: ； In the formula, is the health index of the thin film evaporator; is the predicted failure probability of the thin film evaporator.

25. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 18, characterized in that: The health prediction model is a long short-term memory artificial neural network, which includes an input layer, an LSTM layer, a fully connected layer and an output layer. The input layer is used to accept time series feature data, and its shape is determined by the step size and the total number of features at each step size; the LSTM layer is a single layer with 64 hidden units, which is used to capture long-term dependencies in the time series; the fully connected layer is used to generate prediction results; and the output layer is used to output the failure probability.

26. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The system also includes a cost prediction module, which includes a raw material cost prediction submodule. The raw material cost prediction submodule obtains raw material data of the lyocell fiber production process and preprocesses it, inputs the preprocessed raw material data into the trained raw material consumption prediction overall model, outputs the raw material consumption prediction values ​​corresponding to different machine learning algorithm models, and then obtains the final raw material consumption prediction value by weighted fusion of the prediction values ​​of each machine learning algorithm model; finally, the raw material prediction cost is obtained in combination with the raw material price; wherein the raw material data includes wood pulp parameter data, NMMO solution parameter data, and chemical additive parameter data; the wood pulp parameter data includes wood pulp viscosity, wood pulp polymerization degree and wood pulp methyl fiber content; the NMMO solution parameter data includes NMMO solution concentration; the chemical additive parameter data includes propylene glycol content and hydroxylamine content.

27. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 26, characterized in that: The raw material consumption forecast value includes the wood pulp consumption forecast value, the NMMO solution consumption forecast value, the propylene glycol consumption forecast value and the hydroxylamine consumption forecast value; the raw material forecast cost includes the wood pulp forecast cost, the NMMO solution forecast cost, the propylene glycol forecast cost and the hydroxylamine forecast cost.

28. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 26, characterized in that: The raw material consumption prediction overall model includes a first raw material consumption prediction sub-model constructed based on random forest, a second raw material consumption prediction sub-model constructed based on support vector machine, and a third raw material consumption prediction sub-model constructed based on long short-term memory network.

29. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 26, characterized in that: Preprocess the raw material data, specifically: Adoption 3 σ In principle, outlier detection is performed on the numerical data in the raw material data, and the interquartile range method is used to detect outliers on the non-normal distribution data in the raw material data, and the outliers are removed; then missing value processing and duplicate data removal are performed; Min-Max was used to normalize the raw material data; Pearson correlation analysis and principal component analysis were used to extract key features of raw material data; K-means cluster analysis and information entropy screening methods were used to conduct final screening of raw material data.

30. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 29, characterized in that: The specific method for handling missing values ​​is as follows: time series data is filled using linear interpolation; if the missing ratio of numerical data is less than 20%, k-nearest neighbor interpolation is used to fill it, and if the missing ratio exceeds 20%, the sample is eliminated; if the missing ratio of non-normal distribution data is less than 20%, k-nearest neighbor interpolation is used to fill it, and if the missing ratio exceeds 20%, the sample is eliminated.

31. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 26, characterized in that: The final raw material consumption prediction value is obtained by weighted fusion of the predicted values ​​of each machine learning algorithm model. Specifically, the initial weight distribution of the total raw material consumption prediction model is set, and the mean square error of each machine learning algorithm model is calculated; then the Beta distribution parameters are updated according to the mean square error, and Gaussian process regression is introduced to calculate the uncertainty of each machine learning algorithm model in the total model, and then the Beta distribution parameters are updated again according to the uncertainty; finally, the fusion weights of each machine learning algorithm model are obtained according to the updated Beta distribution parameters combined with variational Bayesian inference calculation.

32. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The system also includes a production management module, which is used to formulate, adjust, approve, query and issue annual production plans, monthly production plans and daily production plans.

33. The digital twin-based lyocell fiber production line intelligent simulation system according to claim 1, characterized in that: The system also includes an equipment management module, which is used to create and manage the equipment information of the production line, configure and issue inspection plans, and configure and issue equipment maintenance strategies; the equipment management module includes an equipment information maintenance submodule, an equipment inspection submodule, and an equipment repair and maintenance submodule; wherein, The equipment information maintenance submodule is used to create and manage production line equipment information, including basic data such as the name, type and number of the configured equipment. The equipment inspection submodule is used to configure the inspection plan and send it to the employee terminal, and collect the inspection results to generate the corresponding inspection log; The equipment inspection and maintenance submodule is used to configure the maintenance strategy and send it to the employee terminal, and collect equipment inspection and maintenance to generate corresponding inspection and maintenance logs.

Images