An energy consumption prediction method and application for the sizing process of Lyocell fiber

Abstract

The present invention relates to the technical field of energy consumption prediction, and particularly relates to an energy consumption prediction method and application for the viscose preparation process of lyocell fiber; historical original data is collected and processed to obtain historical data; the process similarity is calculated according to the process sheet to obtain historical sorting data; the reference group sequences are selected, the process similarity and the threshold interval of the process similarity of each reference group sequence are calculated, and then the energy consumption prediction value sequence of the reference group sequence is calculated; the energy consumption prediction value sequence is compared with the historical energy consumption data, and the sum of absolute errors is recorded; the percentage of the threshold interval of the process similarity is increased to obtain the percentage of the threshold interval of the process similarity corresponding to the minimum sum of absolute errors; the historical screening and sorting data is used as the training set of the neural network to train and obtain an energy consumption prediction model; the process parameter data and the intermediate parameter data are input into the energy consumption prediction model to obtain an energy consumption prediction value; the present invention controls the scale of the training set and improves the credibility and accuracy of model prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of energy consumption prediction, and particularly relates to an energy consumption prediction method and application for a viscose preparation process of Lyocell fibers. Background Art

[0002] As a new type of cellulose fiber, Lyocell fibers are made from natural raw materials such as wood and bamboo. The manufacturing process is green and environmentally friendly. Although it started relatively late in China, it has developed rapidly. Its viscose preparation process is a key link in the production process of Lyocell fibers, directly affecting the quality and performance of the fibers. Steam and electricity energy consumption account for a relatively large proportion in the cost of the Lyocell viscose preparation process. Accurate prediction of it helps enterprises optimize the production process, reduce costs and improve market competitiveness. At present, there are few studies on energy consumption prediction in the production process of Lyocell fibers, and no perfect technical methods have been formed. For the direction of energy consumption prediction, it mostly focuses on the fields of buildings and equipment, and the prediction object is mainly electricity. At the same time, these energy consumption prediction methods often regard energy consumption data as time series data, combine environmental parameters, and perform real-time dynamic prediction through time prediction algorithms, but lack attention to the static energy consumption prediction under a given process sheet. The lack of energy consumption prediction obtained under a given process sheet makes it impossible for the power supply and gas supply departments of enterprises to provide future electricity and gas usage plan reports to cope with sudden power outages and gas cuts during peak electricity and gas usage times in the future, and more energy consumption is required.

[0003] In the related technologies of energy consumption prediction in the prior art, a large amount of historical data is input as a training set for the model to train, and then the energy consumption prediction result is obtained using the model. However, in the technical field of the Lyocell fiber viscose preparation process, if the energy consumption prediction of the Lyocell fiber viscose preparation process is trained with all the collected historical parameters as the model training set, it will lead to too many historical parameters with too low similarity to the current process sheet of the Lyocell fiber viscose preparation process, resulting in low model prediction accuracy and unable to perform accurate energy consumption prediction.

[0004] For example, a Chinese patent with the publication number CN115062847A, publication date of September 16, 2022, and invention title of "A Method and System for Predicting the Energy Consumption of Carbon Fiber Heating Based on Logistic Regression Algorithm" has the following specific technical solution: Real-time collection of the energy consumption of carbon fiber electric heating equipment and environmental data in the target area to obtain factor data, where the factor data is used to characterize the data that affects the electricity consumption of the carbon fiber electric heating equipment; Loading the factor data into a pre-established and trained energy consumption prediction model to obtain the heating energy consumption prediction result; Among them, the establishment and training process of the energy consumption prediction model specifically includes: Obtaining factor data for training; Inputting the factor data for training into the prediction function, and updating the parameters in the prediction function according to the gradient descent method until the training is completed; Using the trained prediction function as the energy consumption prediction model to predict the energy consumption of the factor data collected in real time.

[0005] The above patent obtains the factor data for training through collection, completes the training of the energy consumption prediction model, and inputs the factor data into the energy consumption prediction model to obtain the final prediction result. However, the above patent does not process the factor data for training, which will lead to a decrease in the accuracy of the trained energy consumption prediction model and a low accuracy of the obtained energy consumption prediction result. Summary of the Invention

[0006] In order to solve the problems existing in the above-mentioned prior art, the present invention provides an energy consumption prediction method and application for the Lyocell fiber sizing process that can ensure the accuracy of energy consumption prediction, control the scale of the training set, and reduce the input parameters during the neural training process, thereby preventing overfitting.

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

[0008] In the first aspect, an energy consumption prediction method for the Lyocell fiber sizing process includes the following specific steps:

[0009] Step 1: Collect the evaporation system parameter data, glue liquid and sealing 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 of the historical raw data during the Lyocell fiber sizing process, and perform data preprocessing to obtain historical data;

[0010] Step 2: Calculate the process similarity of the corresponding historical data according to the current process sheet, and rearrange the historical data according to the process similarity to obtain historical sorted data;

[0011] Step 3: 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;

[0012] Step 4: Compare the energy consumption prediction value sequence with the historical energy consumption data of the corresponding reference group sequence respectively, and record the absolute error sum X;

[0013] Step 5: Increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat Steps 3 to 4 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. Mark P as P s ;

[0014] Step 6: Screen the historical sorting data through P s to obtain the historical screened sorting data, and use the pulp porridge weight, glue liquid weight, glue liquid refractive index, and consumption time therein as the training set of the neural network. After training, obtain the energy consumption prediction model;

[0015] Step 7: Extract the pulp porridge weight, process parameter group 1, process parameter group 2, and process parameter group 3 from the process sheet, and input them into the constructed calculation model to obtain the pulp porridge weight, glue liquid weight, glue liquid refractive index, and consumption time, and then input them into the energy consumption prediction model to obtain the energy consumption prediction value.

[0016] Further, the specific method of extracting the pulp porridge weight, process parameter group 1, process parameter group 2, and process parameter group 3 from the process sheet in Step 7, inputting them into the constructed calculation model to obtain the pulp porridge weight, glue liquid weight, glue liquid refractive index, and consumption time, and then inputting them into the energy consumption prediction model to obtain the energy consumption prediction value is as follows: Extract the pulp porridge weight from the process sheet. Pass the pulp porridge weight and the process parameter group 1 extracted from the process sheet through the glue liquid weight calculation model constructed based on the BP neural network algorithm to obtain the glue liquid weight. Pass the pulp porridge weight and the process parameter group 3 extracted from the process sheet through the glue consumption time calculation model constructed based on the least squares method for fitting nonlinear parameters to obtain the consumption time. Pass the process parameter group 2 extracted from the process sheet through the glue liquid refractive index calculation model constructed based on the BP neural network algorithm to obtain the glue liquid refractive index. Then input the pulp porridge weight, glue liquid weight, glue liquid refractive index, and consumption time into the energy consumption prediction model to finally obtain the energy consumption prediction value.

[0017] Further, 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 sorted data includes historically sorted process parameter data, historically sorted intermediate parameter data, and historically sorted energy consumption data; the historical filtered and sorted data includes historically filtered and sorted process parameter data, historically filtered and sorted intermediate parameter data, and historically filtered and sorted energy consumption data.

[0018] Further, the process sheet includes process parameter data.

[0019] Furthermore, the process parameter data includes evaporation system parameter data, glue solution and sealing system parameter data, insulation water and vacuum system parameter data, and gruel weight; the evaporation system parameter data includes evaporation heating system parameter data, evaporation dissolver 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.

[0020] Furthermore, the evaporation heating system parameter data includes heating system temperature, heating system liquid level, and heating system pressure; the evaporation dissolver parameter data includes evaporation dissolver four-zone temperature, evaporation dissolver motor speed, evaporation dissolver bearing lubricating oil circulation temperature, reducer bearing speed, cooling fan speed, evaporation dissolver liquid level, evaporation dissolver outlet pressure, and evaporation dissolver bottom outlet temperature; the evaporation dissolver four-zone temperature includes four evaporation zone jacket temperatures and four evaporation zone glue solution temperatures; the glue solution conveying parameter data includes pump internal insulation water heat exchanger temperature, pump internal insulation water heat exchanger flow rate, pump external insulation water heat exchanger temperature, pump external insulation water heat exchanger flow rate, glue solution conveying pump speed, glue solution conveying pump outlet pressure, and glue solution conveying pump outlet temperature; the sealing liquid system parameter data includes sealing liquid tank temperature, sealing liquid flow rate, and sealing liquid tank liquid level; the insulation water system parameter data includes insulation water circulation pump outlet temperature and insulation system hot water pressure; the evaporation condensate water system parameter data includes evaporation condenser temperature, evaporation condenser liquid level, evaporation condenser pressure, cooling water flow rate to evaporation condenser, and evaporation condensate liquid pump outlet flow rate; the vacuum system parameter data includes vacuum system liquid level, vacuum system pressure, and vacuum pump speed.

[0021] Further, the specific method steps of step seven are as follows:

[0022] Step a1: Based on the BP neural network algorithm, construct a glue weight calculation model; the specific construction method of the glue weight calculation model is as follows: take the slurry weight, the motor speed of the evaporation and dissolution machine, the temperature of the four zones of the evaporation and dissolution machine, the pressure of the vacuum system, and the pump speed of the vacuum pump in the historical screened and sorted process parameter data as the input features of the training set of the glue weight calculation model, and take the glue weight in the historical screened and sorted intermediate parameter data as the target output of the training set of the glue weight calculation model;

[0023] Step b1: Based on the BP neural network algorithm, construct a glue refractive index calculation model; the specific construction method of the glue refractive index calculation model is as follows: take the motor speed of the evaporation and dissolution machine, the temperature of the four zones of the evaporation and dissolution machine, the pump speed of the vacuum pump, and the outlet temperature of the heat preservation water circulation pump in the historical screened and sorted process parameter data as the input features of the training set of the glue refractive index calculation model, and take the glue refractive index in the historical screened and sorted intermediate parameter data as the target output of the training set of the glue refractive index calculation model;

[0024] Step c1: Based on the least squares fitting nonlinear parameter algorithm, construct a glue-making consumption time calculation model; the specific construction method of the glue-making consumption time calculation model is as follows: take the slurry weight, the motor speed of the evaporation and dissolution machine, and the pump speed of the vacuum pump in the historical screened and sorted process parameter data as the input features of the glue-making consumption time calculation model, and take the consumption time in the historical screened and sorted intermediate parameter data as the target output of the glue-making consumption time calculation model;

[0025] Step d1: Extract the slurry weight, process parameter group 1, process parameter group 2, and process parameter group 3 from the process sheet; input the slurry weight and process parameter group 1 into the glue weight calculation model to obtain the glue weight, input process parameter group 2 into the glue refractive index calculation model to obtain the glue refractive index, and input the slurry weight and process parameter group 3 into the glue-making consumption time calculation model to obtain the consumption time. Finally, input the slurry weight, glue weight, glue refractive index, and consumption time into the energy consumption prediction model to obtain the energy consumption prediction value; the process parameter group 1 includes the motor speed of the evaporation and dissolution machine, the temperature of the four zones of the evaporation and dissolution machine, 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 evaporation and dissolution machine, the temperature of the four zones of the evaporation and dissolution machine, 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 evaporation and dissolution machine and the pump speed of the vacuum pump.

[0026] Furthermore, in the step a1, the number of input layer neurons of the glue weight calculation model is set to 8, and the number of output layer neurons is 1; there are two hidden layers, the number of neurons in the first layer is 5, and the number of neurons in the second layer is 4; the activation function is the tanh activation function; the loss function is the MSE; the initial weight is 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; and the precision is set to three decimal places.

[0027] Furthermore, in the step b1, the number of input layer neurons of the glue refractive index calculation model is set to 8, and the number of output layer neurons is set to 1; there are two hidden layers, the number of neurons in the first layer is 6, and the number of neurons in the second layer is 4; the activation function is the tanh activation function; the loss function is the MSE; the initial weight is in the range of -1 to 1; the initial bias value is zero; the learning rate is set to 0.001; and the precision is set to four decimal places.

[0028] Furthermore, the expression of the calculation model of the glue consumption time is as follows:

[0029] ;

[0030] In the formula, t is the time consumed by the glue making process; a, b and e are parameters to be fitted; is the motor speed of the evaporation dissolution machine; is the vacuum pump speed; It is the weight of porridge.

[0031] Furthermore, the intermediate parameter data include glue weight, glue refractive index and consumption time.

[0032] Furthermore, the specific steps of data preprocessing in step 1 are as follows:

[0033] Step a: Eliminate outliers based on the 3σ principle;

[0034] Step b: After outliers are removed, the missing values ​​are filled by taking the weighted average of the K neighboring data values ​​near the missing value.

[0035] Furthermore, in step a, outliers are eliminated based on the 3σ principle, and the expression is as follows:

[0036] ;

[0037] In the formula, is the mean of each column of historical original data, is the standard deviation of each column of historical raw data. The specific calculation formula is as follows:

[0038] ;

[0039] ;

[0040] In the formula, is the number of historical original data in each column, is the value of the historical original data in each column.

[0041] Furthermore, the process similarity in step two is calculated using cosine similarity, and its specific calculation formula is:

[0042] ;

[0043] In the formula, is the vector composed of the process parameter data of the th row of historical data; is the vector composed of the current process sheet data.

[0044] Furthermore, the specific method for selecting the reference group from the historical sorting data in step three is: the first m rows of data in the historical sorting data form a reference group sequence ; through each sequence in the reference group sequence , the historical sorting data corresponding to each reference group sequence is obtained respectively, so as to form the historical sorting data sequence of the reference group sequence, and the maximum process similarity value in the process is recorded; the threshold interval of the process similarity in step three, and the specific calculation formula for calculating the minimum process similarity value of this threshold interval is as follows:

[0045] ;

[0046] In the formula, represents the initial value of the percentage of the threshold interval of the set process similarity;

[0047] Furthermore, the specific method for calculating the energy consumption prediction value sequence using the historical sorting data sequence within its threshold interval in step three is: 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 are used 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 range of the threshold interval of the process similarity.

[0048] Furthermore, before using the porridge weight of the historically sorted intermediate parameter data and the historically sorted 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 homogenization processing. The specific calculation formula is:

[0049] ;

[0050] In the formula, It is the result of the uniform processing of the i-th row and j-th column of the training set; is the original data of row i and column j of the training set; is the minimum value of the process parameters in the jth column of the training set; is the maximum value of the process parameters in the jth column in the training set.

[0051] Furthermore, the neural network model specifically uses a three-layer BP neural network, with 4 neurons in the input layer; 4 neurons in the output layer; 1 hidden layer with 4 neurons; the Tanh activation function is used as the activation function; the MSE is used as the loss function; the initial weight is in the range of -1 to 1; the initial bias value is zero; the learning rate is set to 0.001; and the precision is set to five decimal places.

[0052] Furthermore, the specific calculation formula for increasing the threshold interval percentage P of process similarity in step 5 is:

[0053] ;

[0054] 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; n represents the current step number;

[0055] The threshold interval of the current process similarity is :

[0056] ;

[0057] In the formula, Indicates the current minimum process similarity value; Indicates the maximum process similarity value.

[0058] Furthermore, in step six, the training set is homogenized to eliminate the differences between the dimensions of the input data.

[0059] Further, in step six, the porridge weight, glue solution weight, refractive index of the glue solution, and consumption time in the historical screening and sorting data are used as the training set of the neural network. The specific method for obtaining the energy consumption prediction model through training is 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 porridge 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 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.

[0060] Further, 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:

[0061] ;

[0062] 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 .

[0063] Further, the specific method in step four is: Compare the energy consumption prediction value sequences of the reference group sequences with the historical energy consumption data of their respective corresponding reference group sequences, record their absolute error values, traverse the m rows of historical sorting data in the historical sorting data sequence, and obtain the 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.

[0064] In the second aspect, an application of an energy consumption prediction method for the process of preparing glue for Lyocell fiber is applicable to the energy consumption prediction of the process of preparing glue for Lyocell fiber.

[0065] According to the above technical solutions, the beneficial effects of the present application are as follows:

[0066] 1. Since the steam energy consumption and the electricity energy consumption account for a relatively large proportion in the cost of the process of preparing glue for Lyocell fiber, accurately predicting the energy consumption data by adopting the method of the present invention helps to achieve the technical effects of assisting the enterprise in optimizing the production process of Lyocell fiber and reducing the cost of the process of preparing glue for Lyocell fiber.

[0067] 2. The method adopted by the present invention controls the scale of the training set by calculating the percentage of the most appropriate threshold interval of the process similarity, improving the credibility and accuracy of the model prediction.

[0068] 3. The method adopted in the present invention screens out historical screening and sorting data, and uses the intermediate parameter data of historical screening and sorting and the energy consumption parameter data of historical screening and sorting in the historical screening and sorting data as the training set of the neural network. After training the neural network, an energy consumption prediction model is obtained. Compared with the prior art in which a large amount of data is used 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.

[0069] 4. The method adopted in the present invention extracts and processes the glue solution weight, the refractive index of the glue solution, and the consumption time from the current process sheet by constructing a glue solution weight calculation model, a glue solution refractive index calculation model, and a glue production 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.

[0070] 5. The method adopted in 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.

[0071] 6. The method adopted in the present invention selects the training feature set based on the mechanism of analyzing the influence on 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 model.

[0072] 7. The method adopted in the present invention constructs a glue solution weight calculation model and a glue solution refractive index calculation model through the BP neural network algorithm, and constructs a glue production consumption time calculation model based on the least squares fitting nonlinear parameter algorithm, ensuring the accuracy of the finally obtained glue solution weight, glue solution refractive index, and consumption time, making the final energy consumption prediction result more accurate.

[0073] 8. The method adopted in the present invention obtains the energy consumption prediction value. The obtained energy consumption prediction value helps the power supply and gas supply departments of the enterprise to provide future electricity and gas usage plan reports to cope with the sudden power outages and gas outages during peak electricity and gas usage times in the future, and the need for more energy consumption.

[0074] 9. The method adopted in the present invention obtains the energy consumption prediction value. Relevant staff judge whether the process and process equipment are unstable based on the obtained energy consumption prediction value. If an unstable situation occurs, relevant staff can check and judge the specific problems according to the unstable 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. Description of the Drawings

[0075] Figure 1 It is a block diagram of the energy consumption prediction structure.

[0076] Figure 2 It is a schematic flow chart of Steps 1 to 3 of the method of the present invention.

[0077] Figure 3 It is a schematic flow chart of Steps 4 to 7 of the method of the present invention. Specific embodiments

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application.

[0079] Embodiment 1

[0080] An energy consumption prediction method for a Lyocell fiber sizing process, comprising the following specific steps:

[0081] Step 1: Collect the evaporation system parameter data, sizing solution and sealing system parameter data, heat preservation water and vacuum system parameter data, pulp weight, sizing solution weight, sizing solution refractive index, consumption time, and energy consumption data in the historical raw data of the Lyocell fiber sizing 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 these methods can be used to eliminate outliers in the historical raw data and fill in the missing values in the historical raw data;

[0082] Step 2: Calculate the process similarity of the historical process parameter data corresponding to the current process sheet in the historical data, and rearrange the historical data in descending order of process similarity to obtain historical sorted data;

[0083] Step 3: Select a reference group sequence from the historical sorted data, calculate the process similarity of each reference group sequence to obtain a historical sorted data sequence of the reference group sequence, and obtain a threshold interval of the process similarity during the process of obtaining the historical sorted data sequence of the reference group sequence, and then use the historical sorted data sequence of the reference group sequence within the threshold interval of the process similarity to calculate and obtain an energy consumption prediction value sequence of the reference group sequence; the method for solving the process similarity in Step 3 is the same as that in Step 2;

[0084] Step 4: 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 sum X;

[0085] Step 5: Gradually increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat Steps 3 to 4 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 sum of absolute errors X during the process of repeating Steps 3 to 4, and mark P as P s ; where 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 is taken as half of the maximum similarity.

[0086] Step 6: Screen the historical sorted data through the range set by the threshold interval percentage P of the process similarity to obtain the historical screened and sorted data, and use the gruel weight, glue solution weight, refractive index of the glue solution, and consumption time in the historical screened and sorted data as the training set of the neural network, and obtain the energy consumption prediction model through neural network training; s As shown in

[0087] Step 7: Extract the gruel weight from the process sheet, obtain the glue solution weight by passing the gruel weight and the process parameter group 1 extracted from the process sheet through the glue solution weight calculation model constructed based on the BP neural network algorithm, obtain the consumption time by passing the gruel weight and the process parameter group 3 extracted from the process sheet through the glue-making consumption time calculation model constructed based on the least squares method for fitting nonlinear parameters algorithm, obtain the refractive index of the glue solution by passing the process parameter group 2 extracted from the process sheet through the refractive index of the glue solution calculation model constructed based on the BP neural network algorithm, and then input the gruel weight, glue solution weight, refractive index of the glue solution, and consumption time into the energy consumption prediction model to finally obtain the energy consumption prediction value. Figure 1 Example 2

[0088] An energy consumption prediction method for the glue-making process of Lyocell fibers includes the following specific steps:

[0089] Step 1: Collect the evaporation system parameter data, glue solution and seal system parameter data, insulation water and vacuum system parameter data, gruel weight, glue solution weight, refractive index of the glue solution, consumption time, and energy consumption data in the historical raw data during the glue-making process of Lyocell fibers, and perform data preprocessing to obtain historical data;

[0090] Step 2: Calculate the process similarity of the historical process parameter data in the historical data corresponding to the current process sheet, and rearrange the historical data in descending order of the process similarity to obtain the historical sorted data;

[0091]

[0092] ​Step 3: 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 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 and obtain the energy consumption prediction value sequence of the reference group sequence; the method for solving the process similarity in Step 3 is the same as that in Step 2;

[0093] Step 4: Compare the energy consumption prediction value sequence of the reference group sequence with the historical energy consumption data of its corresponding reference group sequence respectively, and record their absolute error and X;

[0094] Step 5: Gradually increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat Steps 3 to 4 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 3 to 4, and mark P as P s ; where 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 is taken as half of the maximum similarity;

[0095] Step 6: Screen the historical sorting data through the range set by the threshold interval percentage P of the process similarity to obtain the historical screened sorting data, and use the gruel weight, glue solution weight, glue solution refractive index, and consumption time in the historical screened sorting data as the training set of the neural network. After neural network training, obtain the energy consumption prediction model; s

[0096] Step 7: Extract the gruel weight from the process sheet. Use the gruel weight and process parameter group 1 extracted from the process sheet to obtain the glue solution weight through the glue solution weight calculation model constructed based on the BP neural network algorithm. Use the gruel weight and process parameter group 3 extracted from the process sheet to obtain the consumption time through the gelatinization consumption time calculation model constructed based on the least squares method for fitting nonlinear parameters. Use process parameter group 2 extracted from the process sheet to obtain the glue solution refractive index through the glue solution refractive index calculation model constructed based on the BP neural network algorithm. Then, input the gruel weight, glue solution weight, glue solution refractive index, and consumption time into the energy consumption prediction model to finally obtain the energy consumption prediction value.

[0097] ​Historical original data includes historical original process parameter data, historical original energy consumption data, and historical original intermediate parameter data; historical data includes historical process parameter data, historical energy consumption data, and historical intermediate parameter data; historical sorted data includes historically sorted process parameter data, historically sorted intermediate parameter data, and historically sorted energy consumption data; historical filtered and sorted data includes historically filtered and sorted process parameter data, historically filtered and sorted intermediate parameter data, and historically filtered and sorted energy consumption data; intermediate parameter data includes glue liquid weight, glue liquid refractive index, and consumption time; energy consumption data includes electricity energy consumption and steam energy consumption.

[0098] The process sheet includes process parameter data; process parameter data includes evaporation system parameter data, glue liquid and sealing system parameter data, insulation water and vacuum system parameter data, and gruel weight; evaporation system parameter data includes evaporation heating system parameter data, evaporation dissolver parameter data, and evaporation condensate water system parameter data; glue liquid and sealing system parameter data includes sealing liquid system parameter data and glue liquid conveying parameter data; insulation water and vacuum system parameter data includes insulation water system parameter data and vacuum system parameter data.

[0099] Evaporation heating system parameter data includes heating system temperature, heating system liquid level, and heating system pressure; evaporation dissolver parameter data includes evaporation dissolver four-zone temperature, evaporation dissolver motor speed, evaporation dissolver bearing lubricating oil circulation temperature, reducer bearing speed, cooling fan speed, evaporation dissolver liquid level, evaporation dissolver outlet pressure, and evaporation dissolver bottom outlet temperature; evaporation dissolver four-zone temperature includes four evaporation zone jacket temperatures and four evaporation zone glue liquid temperatures; glue liquid conveying parameter data includes pump internal insulation water heat exchanger temperature, pump internal insulation water heat exchanger flow rate, pump external insulation water heat exchanger temperature, pump external insulation water heat exchanger flow rate, glue liquid conveying pump speed, glue liquid conveying pump outlet pressure, and glue liquid conveying pump outlet temperature; sealing liquid system parameter data includes sealing liquid tank temperature, sealing liquid flow rate, and sealing liquid tank liquid level; insulation water system parameter data includes insulation water circulation pump outlet temperature and insulation system hot water pressure; evaporation condensate water system parameter data includes evaporation condenser temperature, evaporation condenser liquid level, evaporation condenser pressure, cooling water to evaporation condenser flow rate, and evaporation condensate liquid pump outlet flow rate; vacuum system parameter data includes vacuum system liquid level, vacuum system pressure, and vacuum pump speed.

[0100] Example 3

[0101] An energy consumption prediction method for the process of making glue from Lyocell fiber includes the following specific steps:

[0102] Step 1: Collect the evaporation system parameter data, glue solution and seal system parameter data, heat preservation water and vacuum system parameter data, pulp porridge weight, glue solution weight, glue solution refractive index, consumption time, and energy consumption data in the historical original data during the preparation process of Lyocell fiber glue, and perform data preprocessing to obtain historical data;

[0103] Step 2: Calculate the process similarity of the historical process parameter data corresponding to the current process sheet in the historical data, and rearrange the historical data in descending order of process similarity to obtain historical sorted data;

[0104] Step 3: Select a reference group sequence from the historical sorted data, calculate the process similarity of each reference group sequence to obtain a historical sorted data sequence of the reference group sequence, and obtain the threshold interval of the process similarity during the process of obtaining the historical sorted data sequence of the reference group sequence. Then, use the historical sorted 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 3 is the same as that in Step 2;

[0105] Step 4: Compare the energy consumption prediction value sequence of the reference group sequence with the historical energy consumption data of their respective corresponding reference group sequences, and record their absolute error sum X;

[0106] Step 5: Gradually increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat Steps 3 to 4 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 3 to 4, and 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, resulting in overfitting and prolonging the calculation time. In practice, the set maximum value of the threshold interval is taken as half of the maximum similarity.

[0107] Step 6: Screen the historical sorted data through the range set by the threshold interval percentage P of the process similarity s to obtain historical screened sorted data, and use the pulp porridge weight, glue solution weight, glue solution refractive index, and consumption time in the historical screened sorted data as the training set of the neural network, and obtain an energy consumption prediction model through neural network training;

[0108] Step 7: Extract the weight of the pulp porridge from the process sheet. Use the weight of the pulp porridge and the process parameter group 1 extracted from the process sheet to obtain the weight of the glue solution through the glue solution weight calculation model constructed based on the BP neural network algorithm. Use the weight of the pulp porridge and the process parameter group 3 extracted from the process sheet to obtain the consumption time through the glue-making consumption time calculation model constructed based on the least squares method for fitting nonlinear parameters. Use the process parameter group 2 extracted from the process sheet to obtain the refractive index of the glue solution through the glue solution refractive index calculation model constructed based on the BP neural network algorithm. Then input the weight of the pulp porridge, the weight of the glue solution, the refractive index of the glue solution, and the consumption time into the energy consumption prediction model to finally obtain the energy consumption prediction value.

[0109] 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 sorted data includes historically sorted process parameter data, historically sorted intermediate parameter data, and historically sorted energy consumption data; the historical filtered and sorted data includes historically filtered and sorted process parameter data, historically filtered and sorted intermediate parameter data, and historically filtered and sorted energy consumption data; the intermediate parameter data includes the weight of the glue solution, the refractive index of the glue solution, and the consumption time; the energy consumption data includes electricity consumption and steam consumption.

[0110] 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 the weight of the pulp porridge; the evaporation system parameter data includes evaporation heating system parameter data, evaporation dissolver parameter data, and evaporation condensate 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.

[0111] The parameter data of the evaporation heating system include the heating system temperature, the heating system liquid level, and the heating system pressure; the parameter data of the evaporation dissolver include the temperatures of the four zones of the evaporation dissolver, the motor speed of the evaporation dissolver, the circulating temperature of the bearing lubricating oil of the evaporation dissolver, the speed of the reducer bearing, the speed of the cooling fan, the liquid level of the evaporation dissolver, the outlet pressure of the evaporation dissolver, and the bottom outlet temperature of the evaporation dissolver; the temperatures of the four zones of the evaporation dissolver 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 include 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 include 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 include 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 system include 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 include the liquid level of the vacuum system, the pressure of the vacuum system, and the speed of the vacuum pump.

[0112] As Figure 1 shown, extract the weight of the pulp porridge from the process sheet, and through the glue solution weight calculation model constructed based on the BP neural network algorithm with the pulp porridge weight and the process parameter group 1 extracted from the process sheet, obtain the glue solution weight. Through the glue consumption time calculation model constructed based on the least squares method for fitting non-linear parameters with the pulp porridge weight and the process parameter group 3 extracted from the process sheet, obtain the consumption time. The process parameter group 2 extracted from the process sheet passes through the glue solution refractive index calculation model constructed based on the BP neural network algorithm to obtain the glue solution refractive index. Then, input the pulp porridge weight, glue solution weight, glue solution refractive index, and consumption time into the energy consumption prediction model. The specific method steps for finally obtaining the energy consumption prediction value are as follows:

[0113] Step a1: Based on the BP neural network algorithm, construct a glue solution weight calculation model; the specific construction method of the glue solution weight calculation model is: use the pulp porridge weight, the motor speed of the evaporation dissolver, the temperatures of the four zones of the evaporation dissolver, the pressure of the vacuum system, and the speed of the vacuum pump in the historically screened and sorted process parameter data as the input features of the training set of the glue solution weight calculation model, and use the glue solution weight in the historically screened and sorted intermediate parameter data as the target output of the training set of the glue solution weight calculation model; as Figure 1 shown, Model 2 is the glue solution weight calculation model;

[0114] Step b1: Based on the BP neural network algorithm, construct a glue refractive index calculation model. The specific construction method of the glue refractive index calculation model is as follows: Use the motor speed of the evaporation and dissolution machine, the temperature of the fourth zone of the evaporation and dissolution machine, the vacuum pump speed, and the outlet temperature of the heat preservation water circulation pump in the historical screened and sorted process parameter data as the input features of the training set of the glue refractive index calculation model, and use the glue refractive index in the historical screened and sorted intermediate parameter data as the target output of the training set of the glue refractive index calculation model. As Figure 1 shown, Model 3 is the glue refractive index calculation model;

[0115] Step c1: Based on the least squares fitting non-linear parameter algorithm, construct a glue-making consumption time calculation model. The specific construction method of the glue-making consumption time calculation model is as follows: Use the pulp porridge weight, the motor speed of the evaporation and dissolution machine, and the vacuum pump speed in the historical screened and sorted process parameter data as the input features of the glue-making consumption time calculation model, and use the consumption time in the historical screened and sorted intermediate parameter data as the target output of the glue-making consumption time calculation model. As Figure 1 shown, Model 4 is the glue-making consumption time calculation model;

[0116] Step d1: Extract the pulp porridge weight, Process Parameter Group 1, Process Parameter Group 2, and Process Parameter Group 3 from the process sheet. Input the pulp porridge weight and Process Parameter Group 1 into the glue weight calculation model to obtain the glue weight, input Process Parameter Group 2 into the glue refractive index calculation model to obtain the glue refractive index, and input the pulp porridge weight and Process Parameter Group 3 into the glue-making consumption time calculation model to obtain the consumption time. Finally, input the pulp porridge weight, glue weight, glue refractive index, and consumption time into the energy consumption prediction model to obtain the energy consumption prediction value. The Process Parameter Group 1 includes the motor speed of the evaporation and dissolution machine, the temperature of the fourth zone of the evaporation and dissolution machine, the vacuum system pressure, and the vacuum pump speed. The Process Parameter Group 2 includes the motor speed of the evaporation and dissolution machine, the temperature of the fourth zone of the evaporation and dissolution machine, the vacuum pump speed, and the outlet temperature of the heat preservation water circulation pump. The Process Parameter Group 3 includes the motor speed of the evaporation and dissolution machine and the vacuum pump speed. In Step a1, the number of neurons in the input layer of the glue 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 uses the tanh activation function. The loss function uses 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.

[0117] In step b1, the number of input layer neurons of the glue refractive index calculation model is set to 8, and the number of output layer neurons is set to 1; there are two hidden layers, the number of neurons in the first layer is 6, and the number of neurons in the second layer is 4; the tanh activation function is selected as the activation function; the MSE is selected as the loss function; the initial weight is taken in the range between -1 and 1; the initial bias value is zero; the learning rate is set to 0.001; the precision is set to four decimal places.

[0118] The expression of the glue making time consumption calculation model is as follows:

[0119] ;

[0120] In the formula, t is the consumption time of the glue making process; a, b, and e are all parameters to be fitted; is the motor speed of the evaporation dissolver; is the vacuum pump speed; is the weight of the pulp porridge; The nonlinear parameter fitting algorithm based on the least squares method is a mature existing technology algorithm in the field, and its specific fitting process is as follows:

[0121] Each time a new process sheet is input, it needs to be fitted once. When initializing for the first time, the initial values of a, b, and e are all initialized to 1, otherwise they are initialized to the parameters of the previous fit;

[0122] Calculate of the Jacobian matrix. The Jacobian matrix J is the partial derivative matrix of the objective function with respect to the model parameters. The Jacobian matrix element The specific calculation formula is as follows:

[0123] ;

[0124] In the formula, is the data in the i-th row and j-th column of the Jacobian matrix J, k represents the current iteration number, is the input feature data of the observation data set, is an element in; The core of nonlinear fitting through the Jacobian matrix lies in using the local information provided by the Jacobian matrix to gradually adjust the parameters through iteration to minimize the error function;

[0125] Calculate the error vector:

[0126] ;

[0127] In the formula, k represents the current iteration number, is the input feature data of the observation data set, the error of the i-th observation data, The consumption time of the glue making process of the i-th observation data, t is the consumption time of the glue making process;

[0128] in ; In the formula, the observed data has n data rows, i refers to a row, and r is the column vector;

[0129] Solve for the increment:

[0130] Solving Linear Equations Get the increment of the parameter ; Where the Jacobian matrix , is the Jacobian matrix The transpose of

[0131] Update the parameters and continuously update the parameters to be fitted to make them close to the correct values:

[0132] ;

[0133] Determine the termination conditions:

[0134] Repeat the above steps until Up to 3 decimal places.

[0135] Example 4

[0136] Based on Example 3, the specific steps of data preprocessing in step 1 are as follows:

[0137] Step a: Eliminate outliers based on the 3σ principle;

[0138] Step b: After outliers are removed, the missing values ​​are filled by taking the weighted average of the K neighboring data values ​​near the missing value.

[0139] In step a, outliers are eliminated based on the 3σ principle, and the expression is as follows:

[0140] ;

[0141] In the formula, is the mean of each column of historical original data, is the standard deviation of each column of historical raw data. The specific calculation formula is as follows:

[0142] ;

[0143] ;

[0144] In the formula, is the number of historical raw data in each column, It is the numerical value of each column of historical original data.

[0145] The process similarity in Step 2 It is calculated using cosine similarity, and its specific calculation formula is:

[0146] ;

[0147] In the formula, is the vector composed of the process parameter data of the th row of historical data; is the vector composed of the current process sheet data.

[0148] The specific method for selecting the reference group from the historical sorting data in Step 3 is as follows: The first m rows of data in the historical sorting data form a reference group sequence ; The value of m is set artificially and should not be too large, as it will increase the calculation amount. m takes 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 for each reference group sequence is obtained respectively, thus forming a historical sorting data sequence of the reference group sequence , and the maximum process similarity value in the process is recorded; The threshold interval of the process similarity, and the specific calculation formula for obtaining the minimum value of the threshold interval is as follows:

[0149] ;

[0150] In the formula, represents the initial value of the percentage of the threshold interval of the set process similarity;

[0151] In Step 4, the energy consumption prediction value sequence of the reference group sequence is compared with the historical energy consumption data of their respective corresponding reference group sequences, and the specific method for recording their absolute error sum X is as follows: The energy consumption prediction value sequence of the reference group sequence is compared with the historical energy consumption data of their respective corresponding reference group sequences, and their absolute error values are recorded. Traverse the m rows of historical sorting data in the historical sorting data sequence to obtain the energy consumption prediction absolute error value sequence , and accumulate the values in the energy consumption prediction absolute error value sequence to obtain the absolute error sum X and record it.

[0152] In Step 3, 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: The intermediate parameter data of the historical sorting and the porridge weight of the process parameter data of the historical sorting in the historical sorting data sequence of the reference group sequence are used 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.

[0153] Before using the intermediate parameter data of the historical sorting and the porridge weight of the process parameter data of the historical sorting as the training set of the neural network model for training, it is necessary to eliminate the differences between the input data dimensions through normalization processing. The specific calculation formula is:

[0154] ;

[0155] In the formula, is the result after normalization processing of the i-th row and j-th column of the training set; is the original data of the i-th row and j-th column of the training set; is the minimum value of the j-th column of process parameters in the training set; is the maximum value of the j-th column of process parameters in the training set.

[0156] The neural network model specifically selects a three-layer BP neural network. The number of neurons in the input layer is 4; the number of neurons in the output layer is 4; there is 1 hidden layer, and the number of neurons is 4; the activation function selects the Tanh activation function; the loss function selects MSE; the initial weight takes values in the range between -1 and 1; the initial bias value is zero; the learning rate is set to 0.001; the precision is set to five decimal places.

[0157] The specific calculation formula for gradually increasing the threshold interval percentage P of the process similarity in Step 5 is:

[0158] ;

[0159] 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; n represents the current step number;

[0160] The current threshold interval of the process similarity is :

[0161] ;

[0162] In the formula, represents the current minimum process similarity value; represents the maximum process similarity value.

[0163] As Figure 1 shown, Model 1 is an energy consumption prediction model; in Step 6, the training set eliminates the differences in the dimensions of the input data through normalization processing.

[0164] In Step 6, the gruel weight, glue liquid weight, refractive index of the glue liquid, and consumption time in the historical screening and sorting data are used as the training set of the neural network. The specific method for obtaining the energy consumption prediction model through neural network training is as follows: the glue liquid weight, refractive index of the glue liquid, 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 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; there is 1 hidden layer, and the number of neurons in this layer is 4.

[0165] Since 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: ;

[0166] 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 .

[0167] Example 5

[0168] The flow chart of this method is as Figure 2 and Figure 3 shown, Figure 2 Steps 1 to 3 in the figure are the flow schematic diagrams: collect the historical raw data in the process of preparing glue for Lyocell fiber, and perform data preprocessing on the historical raw data to obtain historical data; calculate the process similarity of the historical process parameter data corresponding to the current process sheet, and rearrange the historical data in descending order of the process similarity to obtain historical sorted data; select the reference group sequence from the historical sorted data, calculate the process similarity of each reference group sequence to obtain the historical sorted data sequence of the reference group sequence, and obtain the threshold interval of the process similarity during the process of obtaining the historical sorted data sequence of the reference group sequence, and then use the historical sorted data sequence of the reference group sequence to calculate the energy consumption prediction value sequence of the reference group sequence;

[0169] Figure 3Steps 4 to 7 in the process schematic diagram: Compare the energy consumption prediction value sequences of the reference group sequences with the historical energy consumption data of their respective corresponding reference group sequences, and record their absolute error sum X; incrementally increase the threshold interval percentage P of the process similarity to obtain the current threshold interval of the process similarity; repeat Steps 3 to 4 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 3 to 4 s ; Screen the historical sorting data through the threshold interval percentage P of the process similarity s Set the range to obtain the historical screened and sorted data, and use the historical screened and sorted data as the training set of the neural network. After training the neural network, obtain the energy consumption prediction model; extract the pulp porridge weight, glue liquid weight, glue liquid refractive index, and consumption time from the process sheet, and input them into the energy consumption prediction model to finally obtain the energy consumption prediction value.

[0170] Example 6

[0171] Based on Example 4, an application of an energy consumption prediction method for the viscose preparation process of lyocell fibers, which is applicable to the energy consumption prediction of the viscose preparation process of lyocell fibers.

[0172] The above description is a detailed description of the preferred feasible embodiments of the present application, but the embodiments are not used to limit the patent application scope of the present application. Any equivalent changes or modifications completed under the technical spirit disclosed in the present application shall fall within the patent scope covered by the present application.

Claims

1. A method for predicting energy consumption in a lyocell fiber glue making process, characterized in that: The specific steps include: Step 1: 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, consumption time and energy consumption data during the Lyocell fiber glue making process, and perform data preprocessing to obtain historical data; Step 2: Calculate the process similarity of the corresponding historical data according to the current process order, and rearrange the historical data according to the process similarity to obtain historical sorted data; Step 3: Select a reference group sequence from the historical sorting data, calculate 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 the threshold interval to calculate the energy consumption prediction value sequence; Step 4: Compare the energy consumption forecast value sequence with the historical energy consumption data of the corresponding reference group sequence, and record the absolute error and X; Step 5: Increase the threshold interval percentage P of process similarity to obtain the threshold interval of current process similarity; Repeat steps 3 to 4 until the threshold interval of the current process similarity reaches the set maximum value, and obtain the threshold interval percentage P of the process similarity corresponding to the minimum absolute error and X, and mark P as P s ; Step 6: Pass historical sorting data through P s The historical screening and sorting data are obtained by screening, and the porridge weight, glue liquid weight, glue liquid refractive index and consumption time are used as the training set of the neural network, and the energy consumption prediction model is obtained after training; Step 7: Extract the porridge weight, process parameter group 1, process parameter group 2 and process parameter group 3 from the process list, and input them into the constructed calculation model to obtain the porridge weight, glue weight, glue refractive index and consumption time, and input them into the energy consumption prediction model to obtain the energy consumption prediction value; In the step seven, the porridge weight, process parameter group 1, process parameter group 2 and process parameter group 3 are extracted from the process list and input into the constructed calculation model to obtain the porridge weight, glue liquid weight, glue liquid refractive index and consumption time, and input into the energy consumption prediction model. The specific method to obtain the energy consumption prediction value is as follows: extract the porridge weight from the process list, compare the porridge weight with the process parameter group 1 extracted from the process list, and obtain the glue liquid weight through the glue liquid weight calculation model constructed based on the BP neural network algorithm, compare the porridge weight with the process parameter group 3 extracted from the process list, and obtain the consumption time through the glue making consumption time calculation model constructed based on the least squares fitting nonlinear parameter algorithm, obtain the glue liquid refractive index through the glue liquid refractive index calculation model constructed based on the BP neural network algorithm from the process parameter group 2 extracted from the process list, and then input the porridge weight, glue liquid weight, glue liquid refractive index and consumption time into the energy consumption prediction model to finally obtain the energy consumption prediction value.

2. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, 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.

3. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The process sheet includes process parameter data.

4. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 2 or 3, characterized in that: 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, evaporation dissolution machine 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.

5. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 4, characterized in that: The parameter data of the evaporation heating system include 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 evaporation dissolving machine include the temperature of the four zones of the evaporation dissolving machine, the motor speed of the evaporation dissolving machine, the circulation temperature of the lubricating oil of the bearing of the evaporation dissolving machine, the speed of the reducer bearing, the speed of the cooling fan, the liquid level of the evaporation dissolving machine, the outlet pressure of the evaporation dissolving machine and the outlet temperature of the bottom of the evaporation dissolving machine; the temperature of the four zones of the evaporation dissolving machine include the jacket temperature of the four evaporation zones and the glue temperature of the four evaporation zones; the glue delivery parameter data include the temperature of the insulation water heat exchanger in the pump, the flow rate of the insulation water heat exchanger in the pump, 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.

6. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The specific steps of step seven are as follows: Step a1: Based on the BP neural network algorithm, a glue liquid weight calculation model is constructed; the specific construction method of the glue liquid weight calculation model is: the porridge weight, the evaporation dissolution machine motor speed, the evaporation dissolution machine four-zone temperature, the vacuum system pressure and the vacuum pump speed in the historical screening and sorting process parameter data are used as the input features of the glue liquid weight calculation model training set, and the glue liquid weight in the historical screening and sorting intermediate parameter data is used as the target output of the glue liquid weight calculation model training set; Step b1: Based on the BP neural network algorithm, a glue liquid refractive index calculation model is constructed; the specific construction method of the glue liquid refractive index calculation model is: the motor speed of the evaporation dissolution machine, the temperature of the four zones of the evaporation dissolution machine, the speed of the vacuum pump and the outlet temperature of the insulation water circulation pump in the historical screening and sorting process parameter data are used as the input features of the glue liquid refractive index calculation model training set, and the glue liquid refractive index in the historical screening and sorting intermediate parameter data is used as the target output of the glue liquid refractive index calculation model training set; Step c1: constructing a glue-making time calculation model based on a least squares fitting nonlinear parameter algorithm; the specific construction method of the glue-making time calculation model is as follows: using the porridge weight, the evaporation dissolution machine motor speed and the vacuum pump speed in the historically screened and sorted process parameter data as input features of the glue-making time calculation model, and using the consumption time in the historically screened and sorted intermediate parameter data as the target output of the glue-making time calculation model; Step d1: extracting the porridge weight, process parameter group 1, process parameter group 2 and process parameter group 3 from the process list; The porridge weight and process parameter group 1 are input into the glue liquid weight calculation model to obtain the glue liquid weight, the process parameter group 2 is input into the glue liquid refractive index calculation model to obtain the glue liquid refractive index, and the porridge weight and process parameter group 3 are input into the glue making consumption time calculation model to obtain the consumption time, and finally the porridge weight, glue liquid weight, glue liquid refractive index and consumption time are input into the energy consumption prediction model to obtain the energy consumption prediction value; the process parameter group 1 includes the evaporation dissolution machine motor speed, the evaporation dissolution machine four-zone temperature, the vacuum system pressure and the vacuum pump speed; the process parameter group 2 includes the evaporation dissolution machine motor speed, the evaporation dissolution machine four-zone temperature, the vacuum pump speed and the insulation water circulation pump outlet temperature; the process parameter group 3 includes the evaporation dissolution machine motor speed and the vacuum pump speed.

7. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 6, characterized in that: In the step a1, the number of input layer neurons of the glue weight calculation model is set to 8, and the number of output layer neurons is 1; there are two hidden layers, the number of neurons in the first layer is 5, and the number of neurons in the second layer is 4; the tanh activation function is selected as the activation function; the MSE is selected as the loss function; the initial weight is 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; and the precision is set to three decimal places.

8. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 6, characterized in that: In the step b1, the number of input layer neurons of the glue refractive index calculation model is set to 8, and the number of output layer neurons is set to 1; there are two hidden layers, the number of neurons in the first layer is 6, and the number of neurons in the second layer is 4; the tanh activation function is selected as the activation function; the MSE is selected as the loss function; the initial weight is in the range of -1 to 1; the initial bias value is zero; the learning rate is set to 0.001; and the precision is set to four decimal places.

9. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 6, characterized in that: The expression of the calculation model of glue consumption time is as follows: ； In the formula, t is the time consumed by the glue making process; a, b and e are parameters to be fitted; is the motor speed of the evaporation dissolution machine; is the vacuum pump speed; It is the weight of porridge.

10. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 2, characterized in that: The intermediate parameter data include glue liquid weight, glue liquid refractive index and consumption time.

11. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The specific steps of data preprocessing in step 1 are as follows: Step a: Eliminate outliers based on the 3σ principle; Step b: After outliers are removed, the missing values ​​are filled by taking the weighted average of the K neighboring data values ​​near the missing value.

12. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 11, characterized in that: In step a, outliers are eliminated based on the 3σ principle, and the expression is as follows: ； In the formula, is the mean of each column of historical original data, is the standard deviation of each column of historical raw data. The specific calculation formula is as follows: ； ； In the formula, is the number of historical raw data in each column, It is the numerical value of each column of historical original data.

13. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: Process similarity in step 2 The cosine similarity is used for calculation, and the specific calculation formula is: ； In the formula, It is A vector consisting of the process parameter data of the row history; It is a vector consisting of the current process order data.

14. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The specific method of selecting the reference group from the historical sorting data in step 3 is: the first m rows of the historical sorting data form a 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 ; The threshold interval of process similarity in step 3 , calculate the minimum process similarity value of the threshold interval The specific calculation formula is as follows: ； In the formula, Indicates the initial value of the threshold interval percentage of the set process similarity.

15. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: In the step three, the historical sorted data sequence within the threshold interval is used to calculate the energy consumption prediction value sequence. The specific method is: the porridge weight of the historical sorted intermediate parameter data and the historical sorted process parameter data in the historical sorted data sequence of the reference group sequence is used 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 historical sorted intermediate parameter data belongs to the threshold interval of the process similarity.

16. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 15, characterized in that: Before using the porridge weight of the historically sorted intermediate parameter data and the historically sorted process parameter data as the training set of the neural network model for training, it is necessary to eliminate the difference between the dimensions of the input data through homogenization processing. The specific calculation formula is: ； In the formula, It is the result of the uniform processing of the i-th row and j-th column of the training set; is the original data of row i and column j of the training set; The minimum value of the process parameters in the jth column of the training set; The maximum value of the process parameters in the jth column in the training set.

17. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 16, characterized in that: The neural network model specifically uses a three-layer BP neural network, with 4 neurons in the input layer; 4 neurons in the output layer; 1 hidden layer with 4 neurons; the Tanh activation function is used as the activation function; the MSE is used as the loss function; the initial weight is in the range of -1 to 1; the initial bias value is zero; the learning rate is set to 0.001; and the precision is set to five decimal places.

18. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The percentage of the threshold interval for increasing the process similarity in step 5 The specific calculation formula is: ； 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; n represents the current step number; The threshold interval of the current process similarity is : ； In the formula, Indicates the current minimum process similarity value; Indicates the maximum process similarity value.

19. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 18, characterized in that: In step six, the training set is homogenized to eliminate the differences between the dimensions of the input data.

20. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: In step six, the porridge weight, glue weight, glue refractive index and consumption time in the historical screening and sorting data are used as the training set of the neural network. The specific method of obtaining the energy consumption prediction model after training is as follows: the glue weight, glue refractive index and consumption time in the intermediate parameter data of the historical screening and sorting and the porridge 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 are used as the target output of the energy consumption prediction model; the energy consumption prediction model uses a three-layer BP neural network, the number of neurons in the input layer is 4, 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.

21. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The energy consumption data is the sum of electricity energy consumption and steam energy consumption; the specific calculation formula for converting steam energy consumption into electricity energy consumption is as follows: ； In the formula, The electricity consumption for steam; is the price of steam, in yuan / m 3 ; is the price of electric energy, in Yuan / KW·h; is the amount of steam used, in m 3 .

22. The method for predicting energy consumption in a lyocell fiber glue making process according to claim 1, characterized in that: The specific method in step 4 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, traverse the m rows of historical sorted data sequence, and obtain the energy consumption prediction absolute error value sequence , accumulate the energy consumption prediction absolute error value sequence to obtain the absolute error and X and record it.

23. An application of an energy consumption prediction method for a Lyocell fiber glue making process, characterized in that: The energy consumption prediction method for the lyocell fiber glue making process as described in any one of claims 1 to 22 is suitable for predicting the energy consumption of the lyocell fiber glue making process.

Images