CN118364864A - Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard - Google Patents
Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard Download PDFInfo
- Publication number
- CN118364864A CN118364864A CN202410576983.7A CN202410576983A CN118364864A CN 118364864 A CN118364864 A CN 118364864A CN 202410576983 A CN202410576983 A CN 202410576983A CN 118364864 A CN118364864 A CN 118364864A
- Authority
- CN
- China
- Prior art keywords
- layer
- vector
- data
- spraying liquid
- spraying
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000010410 layer Substances 0.000 title claims abstract description 147
- 238000005507 spraying Methods 0.000 title claims abstract description 88
- 238000000034 method Methods 0.000 title claims abstract description 79
- 239000007788 liquid Substances 0.000 title claims abstract description 61
- 239000011087 paperboard Substances 0.000 title claims abstract description 45
- 239000002131 composite material Substances 0.000 title claims abstract description 27
- 239000002344 surface layer Substances 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 103
- 230000008569 process Effects 0.000 claims abstract description 43
- 238000003062 neural network model Methods 0.000 claims abstract description 25
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000004519 manufacturing process Methods 0.000 claims abstract description 9
- 239000000758 substrate Substances 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 31
- 238000012549 training Methods 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 10
- 210000002569 neuron Anatomy 0.000 claims description 10
- 239000002346 layers by function Substances 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000002194 synthesizing effect Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 238000013480 data collection Methods 0.000 abstract description 16
- 238000006243 chemical reaction Methods 0.000 abstract description 6
- 230000000694 effects Effects 0.000 abstract description 4
- 239000005022 packaging material Substances 0.000 abstract description 2
- 238000000576 coating method Methods 0.000 description 18
- 239000011248 coating agent Substances 0.000 description 17
- 238000012544 monitoring process Methods 0.000 description 10
- 239000007921 spray Substances 0.000 description 8
- 239000008199 coating composition Substances 0.000 description 6
- 235000013305 food Nutrition 0.000 description 6
- 239000000203 mixture Substances 0.000 description 6
- 238000004806 packaging method and process Methods 0.000 description 6
- 239000003973 paint Substances 0.000 description 6
- 238000009826 distribution Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000004224 protection Effects 0.000 description 5
- 239000012792 core layer Substances 0.000 description 4
- 239000003814 drug Substances 0.000 description 4
- 238000009472 formulation Methods 0.000 description 4
- 239000011111 cardboard Substances 0.000 description 3
- 230000008602 contraction Effects 0.000 description 3
- 239000002537 cosmetic Substances 0.000 description 3
- 230000006750 UV protection Effects 0.000 description 2
- 230000000844 anti-bacterial effect Effects 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 241000894006 Bacteria Species 0.000 description 1
- 235000013194 Lyophyllum decastes Nutrition 0.000 description 1
- 240000005856 Lyophyllum decastes Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 239000003242 anti bacterial agent Substances 0.000 description 1
- 230000000845 anti-microbial effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 239000011247 coating layer Substances 0.000 description 1
- 235000016213 coffee Nutrition 0.000 description 1
- 235000013353 coffee beverage Nutrition 0.000 description 1
- 235000009508 confectionery Nutrition 0.000 description 1
- 230000032798 delamination Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 235000011869 dried fruits Nutrition 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 235000012020 french fries Nutrition 0.000 description 1
- 239000004519 grease Substances 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Landscapes
- Paper (AREA)
Abstract
The invention belongs to the technical field of packaging material processing, and particularly relates to a surface layer spraying liquid temperature control method and a spraying device of a composite six-layer paperboard, wherein the method comprises the following steps: collecting cross-sectional dimension data and process flow data of five-layer paper boards serving as substrates to obtain a first data set; collecting formula data, performance data and spraying requirement data of the spraying liquid to obtain a second data set; collecting real-time parameters of a position to be sprayed at a set production position, including the temperature and the humidity of the surface, and obtaining a third data set; converting the data set to obtain three vectors respectively; and respectively giving corresponding weights to the three vectors, inputting the result into a neural network model, and outputting a temperature control value of the spraying liquid by the neural network model. According to the invention, through integrating data collection, conversion and neural network model prediction, the technical effect of spraying liquid and controlling temperature on the surface layer of the composite six-layer paperboard is realized, and a high-efficiency and accurate solution is provided for production in the related field.
Description
Technical Field
The invention belongs to the technical field of packaging material processing, and particularly relates to a surface layer spraying liquid temperature control method and a spraying device for a composite six-layer paperboard.
Background
At present, for packaging box paperboards in multiple fields, the set requirement is often required to be achieved through an additional coating on the surface, and the moisture-proof requirement is taken as an example:
food packaging field: is especially suitable for foods which need long-term preservation, such as dried fruits, candies, coffee and the like;
Medicine packaging field: can keep the medicine dry and prolong the effective period;
cosmetic field: some cosmetics require moisture protection, and such paperboard may provide protection;
Electronic product packaging: for some electronic products, the moisture protection function can protect the internal circuit board from moisture.
Of course, in addition to the moisture barrier function described above, by the provision of additional coatings, the functions that can be achieved include, but are not limited to:
oil repellency: for the field of food packaging, in particular to grease food such as french fries, fried chicken and the like, the oil-proof function can prevent the paperboard from being permeated by greasy dirt, and the packaging appearance is kept clean.
Antibacterial/mildew-proof: in the fields of foods, medicines and the like, the coating possibly contains an antibacterial agent or a mildew preventive, so that the product is protected from bacteria and mildew, and the shelf life of goods is prolonged.
UV protection: for certain products that require protection from ultraviolet radiation, such as cosmetics, pharmaceuticals or certain foods, the coating may provide a degree of ultraviolet protection.
For each coating, the coating can be covered on the surface of the existing paperboard structure in an attempt of spraying, different spraying schemes are often needed for different coating formulas and different paperboard structures and parameters serving as spraying substrates, and one important control parameter in the schemes is the temperature of spraying liquid.
For the existing paperboard structure, a three-layer structure and a five-layer structure are more conventional, the three-layer structure comprises two plane surface layers and a corrugated core layer clamped between the two plane surface layers, the five-layer structure further comprises a corrugated core layer and a plane surface layer on the basis of the three-layer structure, and the two more conventional products can obtain additional functional layers by adding a coating layer, so that a four-layer and six-layer composite structure is formed respectively.
In a specific implementation process, compared with a three-layer structure, the spray coating liquid temperature control of the five-layer structure is obviously more difficult, and the specific reason is that:
The heat conduction path increases: there are more levels in the five-layer structure, including additional corrugated core layers and planar skin layers, which increase the path of heat conduction, resulting in non-uniformity of temperature distribution, and therefore, more precise temperature control is required to ensure uniform coverage of the coating over the entire paperboard surface.
Temperature gradient increases: due to the multilayer structure, the temperature gradient between the different layers is also increased. This means that the temperature may vary at different levels during the spraying process, which affects the adhesion and properties of the coating.
Difference of thermal expansion and cold contraction: in the multilayer structure, the coefficient of thermal expansion and contraction may be different for each layer, resulting in different degrees of expansion and contraction between layers when the spray liquid temperature is changed, which may lead to structural instability and even to delamination or deformation between layers.
Based on the above situation, how to accurately control the temperature of the surface spraying of the five-layer paperboard at present, so as to obtain the composite six-layer paperboard with better performance, and the method becomes a technical problem to be solved in the field.
Disclosure of Invention
The invention provides a surface layer spraying liquid temperature control method and a spraying device for a composite six-layer paperboard, which can effectively solve the problems in the background technology.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
a surface spraying liquid temperature control method of a composite six-layer paperboard comprises the following steps:
Collecting cross-sectional dimension data and process flow data of five-layer paper boards serving as substrates to obtain a first data set;
Collecting formula data, performance data and spraying requirement data of the spraying liquid to obtain a second data set;
Collecting real-time parameters of a position to be sprayed at a set production position, including the temperature and the humidity of the surface, and obtaining a third data set;
Converting the first data set, the second data set and the third data set to obtain a first vector, a second vector and a third vector respectively;
respectively giving corresponding weights to the first vector, the second vector and the third vector;
And inputting the weighted vector result into a neural network model, wherein the neural network model is used for outputting a temperature control value of the spraying liquid.
Further, assigning corresponding weights to the first vector, the second vector, and the third vector, respectively, includes:
determining a weight index as a criterion for giving weight to each vector; and calculating the weight by adopting a hierarchical analysis method based on the weight index.
Further, the analytic hierarchy process includes:
Determining a hierarchical structure of the decision problem, wherein the hierarchical structure comprises a target layer, a criterion layer and a sub-criterion layer, and each hierarchy comprises a plurality of factors;
Determining the relative importance of each factor in two layers between the criterion layer and the target layer and between all sub-criterion layers in the criterion layer, and correspondingly filling into a judgment matrix;
carrying out consistency test on each judgment matrix, if the judgment results are consistent, continuing, otherwise, adjusting the judgment matrix;
Calculating the feature vector of each judgment matrix to obtain the weight of each factor;
Synthesizing the weights of all layers to obtain the final sorting result of all factors;
And comprehensively processing the sequencing results of the factors in each vector to obtain weights of the first vector, the second vector and the third vector respectively.
Further, the consistency check of each judgment matrix is realized by calculating a consistency ratio and a random consistency index.
Further, calculating the eigenvector of each judgment matrix adopts an eigenvalue method.
Further, the weight of each layer is synthesized by a weighted average method.
Further, converting the first data set, the second data set and the third data set to obtain a first vector, a second vector and a third vector, respectively, including:
Preprocessing the first data set, the second data set and the third data set;
Feature extraction is carried out on each preprocessed data set, and primary vectors of three vectors are constructed;
Performing feature engineering on each initial vector, and extracting representative and distinguishing features;
And converting the extracted features into vector forms to respectively obtain the first vector, the second vector and the third vector.
Further, the neural network model is a multi-layer perceptron.
Further, the multi-layer sensor includes:
An input layer for receiving the vector result given with the weight;
The hidden layers comprise one or more hidden layers, each hidden layer is composed of a plurality of fully connected nodes, and a Dropout layer is added behind each hidden layer so as to randomly discard part of neurons in the training process;
the Dropout layers are added behind each hidden layer and are used for discarding part of neurons in the hidden layers at a set discarding rate in the training process;
An output layer for generating an output result of the model;
The calculation formula of the discarding rate is as follows:
Wherein Dropoutrate is the discard rate;
w 1,w2,w3 represents the weights of the first, second and third vectors, respectively;
Var (w 1,w2,w3) represents the variance of the three weights;
max (w 1,w2,w3) and Min (w 1,w2,w3) represent the maximum and minimum of the three weights, respectively.
A spraying device is used for spraying functional layers on five-layer paper boards serving as substrates, and in the spraying process of the functional layers, the spraying liquid temperature control method is adopted for controlling the spraying liquid temperature on the surface layer of the composite six-layer paper board.
By the technical scheme of the invention, the following technical effects can be realized:
The invention starts from collecting basic data, through data conversion and weight giving, and finally utilizes the neural network model to output the spraying liquid temperature control, thereby realizing the comprehensive and systemization from data collection to model application; in the implementation process, the data driving and instantaneity of the spraying liquid temperature control scheme are realized by collecting basic data and real-time parameter data and combining with training and prediction of a neural network model, so that the adjustment and optimization are facilitated according to real-time conditions, and the accurate control of the spraying liquid temperature is ensured; the data collection and conversion steps in the method provide flexibility and adjustability for the subsequent neural network model, and the importance degree of the model on various data can be adjusted by giving weights to different data sets, so that the spraying requirements under different conditions can be better met. According to the method, the high efficiency and the accuracy of the temperature control of the spraying liquid are realized by combining the application of the neural network model, and the neural network model can predict the optimal temperature control value of the spraying liquid by learning historical data and real-time parameters, so that the temperature control accuracy and the efficiency of the spraying liquid are effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings may be obtained according to the drawings without inventive effort to those skilled in the art.
FIG. 1 is a flow chart of a method for controlling the temperature of a surface layer spraying liquid of a composite six-layer paperboard;
FIG. 2 is a flow chart of an analytic hierarchy process;
FIG. 3 is a flow chart of obtaining a first vector, a second vector, and a third vector.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
As shown in fig. 1, a method for controlling the temperature of a surface layer spraying liquid of a composite six-layer paperboard comprises the following steps:
S1, collecting cross-section size data and process flow data of five-layer paper boards serving as substrates, namely collecting related data of original five-layer paper boards which are not sprayed by spraying liquid, and obtaining a first data set;
In this step, the cross-sectional dimension data collection involves detailed data of the actual dimensions, thickness and layering structure of the five-layer board, including the thickness of each layer, the overall dimensions of the board, the wavy shape and size of the corrugated core layer, etc., which will be the basis for the temperature control of the spray liquid to adjust the temperature of the spray liquid according to the geometric characteristics of the board. Process flow data collection includes collecting current board production process flow data, such as line speed, temperature and humidity control, etc., which is important to assess factors that affect spray coating temperature control, as these factors may affect coating adhesion and performance.
S2, collecting formula data, performance data and spraying requirement data of the spraying liquid to obtain a second data set;
The formulation data, including composition, concentration, viscosity, etc., of the liquid coating material to be sprayed are critical to adjusting the properties of the coating and the temperature of the spray liquid to ensure uniform distribution and desired function on the surface of the board at different levels. Performance data is primarily directed to, for example, moisture resistance, oil repellency, antimicrobial properties, etc., which will help determine whether the coating meets the desired performance requirements, thereby directing the adjustment of the spray temperature. The spraying requirements are aimed at the spraying requirements of different products and application scenes, such as coating thickness, coating mode, curing conditions and the like, and the data are used for guiding the specific requirements and adjustment of temperature control in the spraying process.
S3, collecting real-time parameters of a position to be sprayed at a set production position, including the temperature and the humidity of the surface, and obtaining a third data set;
The step can be realized by installing a sensor or monitoring equipment so as to ensure the stability of temperature and humidity in the spraying process, and the real-time monitoring of the parameters can help to timely adjust the temperature and spraying parameters of the spraying liquid so as to ensure the quality and performance of the coating;
In the implementation process, the steps S1 and S2 are required to be collected before spraying, but the steps are not required to be carried out in sequence, and can be carried out simultaneously, can be carried out in a crossed mode or can be carried out in an exchange sequence, and are all within the protection scope of the invention. Step S3 is performed in real time during the implementation process, and certainly stable production needs to be established under the condition that the implementation parameters are stable within a set range, and if the real-time parameters are mutated, abnormal situations need to be checked and processed.
S4: converting the first data set, the second data set and the third data set to obtain a first vector, a second vector and a third vector respectively; these vectors will be input as input data into the neural network model;
To better reflect the importance of the data, S5 is performed: corresponding weights are respectively given to the first vector, the second vector and the third vector; these weights can be adjusted according to the actual situation to ensure that the model makes better use of the input data;
s6: the vector result given with the weight is input into a neural network model, the neural network model is used for outputting a temperature control value of the spraying liquid, the model can be a trained neural network, and the optimal spraying liquid temperature control value is predicted according to the input vector data so as to realize the surface spraying liquid temperature control of the composite six-layer paperboard.
By implementing the invention, the following technical effects can be realized:
(1) Overall and systematic: the invention starts from collecting basic data, through data conversion and weight giving, and finally utilizes the neural network model to output the spraying liquid temperature control, thereby realizing the comprehensive and systemization from data collection to model application.
(2) Data driving and real-time: by collecting basic data and real-time parameter data and combining training and prediction of a neural network model, data driving and instantaneity of a spraying liquid temperature control scheme are realized, adjustment and optimization are facilitated according to real-time conditions, and accurate control of spraying liquid temperature is ensured.
(3) Flexibility and adjustability: the data collection and conversion steps in the method provide flexibility and adjustability for the subsequent neural network model, and the importance degree of the model on various data can be adjusted by giving weights to different data sets, so that the spraying requirements under different conditions can be better met.
(4) High efficiency and accuracy: by combining the application of the neural network model, the method realizes the high efficiency and the accuracy of the temperature control of the spraying liquid, and the neural network model can predict the optimal temperature control value of the spraying liquid by learning historical data and real-time parameters, so that the temperature control accuracy and the efficiency of the spraying liquid are effectively improved.
As a preference of the above embodiment, assigning corresponding weights to the first vector, the second vector, and the third vector, respectively, includes:
Determining a weight index as a criterion for giving weight to each vector; based on the weight index, the weight is calculated by adopting a analytic hierarchy process. In the preferred scheme, the relative importance of each vector in solving the problem needs to be considered when determining the weight indexes, and the weight indexes can be factors related to the target, such as the reliability of data, the contribution degree of the target implementation, the influence on the system performance and the like, and in the implementation process, the selection of the weight indexes should be based on the actual situation of the problem so as to ensure the accuracy and the reliability of weight calculation.
Through the use of the analytic hierarchy process, the composite six-layer paperboard spraying liquid temperature control problem can be decomposed into a plurality of layers, including basic data collection, paint formula and performance data collection, real-time parameter monitoring and the like, and the factors of each layer are compared and evaluated, so that the relative importance of each factor in problem solving can be accurately determined, and a decision maker is helped to better distribute resources and energy. The analytic hierarchy process can decompose the complex spraying liquid temperature control problem into a series of relatively independent sub-problems, so that the problem treatment is more systematic and manageable, the relation among the factors can be clearly understood through layer-by-layer comparison and induction, the complexity of the system is reduced, and the problem solving efficiency is improved.
As a preferred embodiment of the above, as shown in fig. 3, the hierarchical analysis method includes:
A1: determining a hierarchical structure of the decision problem, wherein the hierarchical structure comprises a target layer, a criterion layer and a sub-criterion layer, and each hierarchy comprises a plurality of factors; for example:
(1) Target layer: realizing the surface spraying liquid temperature control of composite six-layer paperboard
(2) Criterion layer:
(21) Basic data collection:
Paperboard dimensions and structural data: including the size, thickness, and structural and characteristic data of the various layers.
Process flow data: and collecting process parameters such as production line speed, temperature, humidity control and the like.
(22) Coating formulation and performance data collection:
Coating composition data: the formula, the components, the concentration and the like of the coating.
Coating performance data: the coating has the advantages of moisture resistance, oil resistance, antibacterial property and the like.
(23) And (3) monitoring real-time parameters:
surface temperature monitoring: and monitoring the temperature of the surface of the paperboard.
Surface humidity monitoring: monitoring the humidity of the cardboard surface.
(3) Sub-criteria layer:
(31) Basic data collection:
Cardboard dimension measurement accuracy: accuracy and precision of measuring the dimensions of the cardboard.
Reliability of process flow data: reliability and stability of process data during production.
(32) Coating formulation and performance data collection:
Accuracy of coating composition: accuracy and stability of paint composition data.
Reliability of the paint performance test method: and (5) evaluating the accuracy and the reliability of the paint performance data testing method.
(33) And (3) monitoring real-time parameters:
precision of temperature and humidity sensor: accuracy and precision of the sensor for monitoring surface temperature and humidity.
Stability of the real-time data acquisition system: the stability of the system and the accuracy of data acquisition are monitored in real time.
In this hierarchy, each factor is a key element that must be considered in order to achieve the goal, and further distance these factors can help us better understand the contribution degree and importance of each factor to achieving the goal.
A2: determining the relative importance of each factor in two layers between the criterion layer and the target layer and between all sub-criterion layers in the criterion layer, and correspondingly filling into a judgment matrix;
In the judgment matrix, each element represents the relative importance evaluation of a certain factor to another factor, and the process of filling the judgment matrix can be completed by expert judgment, questionnaires and other modes so as to obtain quantitative evaluation of the relative importance among the factors;
For example, assuming that we want to fill out the judgment matrix between the criterion layer and the target layer of the paint formulation and the performance data collection, we need to consider the contribution degree of the paint formulation and the performance data collection to the realization of the surface spraying liquid temperature control target of the composite six-layer paperboard: first, the opinion may be collected by a set of experts or with a questionnaire, and then the expert opinion or survey results used to populate the judgment matrix, which requires consideration of the relative importance between the two factors, typically with a quantitative score of 1 to 9, where 1 indicates the same importance between the two factors and 9 indicates that the importance of one factor to the other is extremely significant. Assuming that the expert believes that the coating formulation and performance data collection contribute 7 to achieving the spray liquid temperature control objective, this means that the expert believes that this factor is relatively important to achieving the objective.
Similarly, we need to fill in the decision matrix between other criterion layers and the target layer, and between sub-criterion layers inside the criterion layer, in order to fully evaluate the relative importance between the factors.
A3: carrying out consistency test on each judgment matrix, if the judgment results are consistent, continuing, otherwise, adjusting the judgment matrix;
in general, consistency indexes such as a Consistency Ratio (CR) and a random Consistency Index (CI) are adopted for evaluation, if the consistency indexes meet a certain standard, it is indicated that the judgment matrix has higher consistency, and the judgment result is relatively reliable, otherwise, it is indicated that the judgment matrix has larger inconsistency and needs to be adjusted;
the consistency ratio is a quantization index of consistency of the judgment matrix, and represents a proportional relation between consistency and random consistency contained in the matrix, if the consistency ratio is closer to 1, the consistency of the judgment matrix is better, otherwise, the consistency ratio is worse, the random Consistency Index (CI) is calculated firstly, and then the random consistency index is compared with an actual consistency difference. The random uniformity index is a theoretical value used to represent the expected value of random uniformity, and is calculated based on the dimension of a decision matrix, usually determined using a specific table or function, and is obtained by averaging the matrices of random uniformity indexes to decide whether the uniformity level of a given matrix is higher than the random level.
A4: calculating the feature vector of each judgment matrix to obtain the weight of each factor;
each element in the feature vector represents the weight of the corresponding factor, the weights reflect the relative importance of each factor in the decision problem, and the weight of each factor can be obtained by calculating the feature vector for subsequent comprehensive processing;
a5: synthesizing the weights of all layers to obtain the final sorting result of all factors;
in this step, the weights of the factors are synthesized according to the hierarchical structure where the factors are located, so as to obtain the final weights of the factors, and when the method is implemented, the weights of the factors can be weighted and averaged according to the weights of the layers, so as to obtain the final sorting result.
A6: and comprehensively processing the sequencing results of the factors in each vector to obtain weights of the first vector, the second vector and the third vector respectively.
In this step, for the weights of the factors in each vector, further comprehensive processing, such as normalization processing or weighting processing, may be performed according to actual requirements, so as to obtain the weights of the final first vector, the second vector and the third vector, where these weights are used in subsequent decision and analysis processes to guide the surface spraying liquid temperature control of the composite six-layer paperboard.
In the above step, it is preferable that the eigenvector of each judgment matrix is calculated by an eigenvalue method. In the implementation process, the eigenvalue needs to be solved through an eigenvalue, and for a judgment matrix in the form of an n-order square matrix A, the eigenvalue lambda satisfies an eigenvalue |A-lambdaI|=0, wherein I is a unit matrix, and the eigenvalue lambda is obtained by solving the eigenvalue. For each eigenvalue λ, a system of equations (a- λi) x=0 is solved, where X is an eigenvector, the system of equations is solved to obtain an eigenvector corresponding to the eigenvalue λ, and for each eigenvector, normalization processing is performed to make its modulus be 1. This ensures that the sum of the weights of the factors is 1.
Preferably, the weights of the respective layers are combined by a weighted average method. In the specific implementation process, for each layer, calculating the weighted average value of the weights of all factors of the next stage, namely multiplying the weight of each factor by the weight of the layer where the factor is positioned, adding the results to obtain the weighted average weight of the layer, synthesizing the weights layer by layer from the bottommost layer, and multiplying the weighted average weight of the sub-criterion layer by the weight of the criterion layer to obtain the weighted average weight of the criterion layer; and multiplying the weighted average weight of the criterion layer by the weight of the target layer to obtain the weighted average weight of the target layer. And finally obtaining weighted average weights of all the factors under the target layer, wherein the weights are the final sequencing results of all the factors, and the importance degree of all the factors in realizing the target can be determined according to the weights. By the method, the weights of the factors can be synthesized to the target layer by layer, so that a final ordering result is obtained, and the result is obtained based on the whole hierarchical structure and the relative importance among the layers.
In a specific implementation process, the first data set, the second data set and the third data set are converted to obtain a first vector, a second vector and a third vector respectively, and one specific mode includes:
S41: preprocessing the first data set, the second data set and the third data set;
the purpose of the preprocessing is to clean and prepare the data so that it is suitable for subsequent feature extraction and modeling processing; the preprocessing process may include processing missing values, processing outliers, normalizing or normalizing data, and removing duplicate data, etc.; the quality of the pretreatment is important for the subsequent feature extraction and model establishment, because the high-quality data can improve the accuracy and stability of the model;
S42: feature extraction is carried out on each preprocessed data set, and primary vectors of three vectors are constructed;
In the step, extracting features of each preprocessed data set, and extracting important information capable of representing the features of the data from the feature extraction; the feature extraction method may include statistical feature extraction, frequency domain feature extraction, time domain feature extraction, etc., specifically selected according to the nature and characteristics of the data; the extracted features will constitute the primary vector form of each dataset, i.e. the initial forms of the first, second and third vectors;
S43: carrying out feature engineering on each initial vector, and extracting representative and distinguishing features; the primary vector can be further processed by adopting a dimension reduction technology (such as Principal Component Analysis (PCA)), feature selection, feature conversion and other methods, more important and effective features are extracted, and the purpose of feature engineering is to reduce the complexity of data, improve the generalization capability of a model and simultaneously reserve the effective information of the data;
s44: the extracted features are converted into vector form to obtain a first vector, a second vector and a third vector, respectively. The vectors are used as inputs for subsequent steps for weight assignment, neural network model input and the like.
Preferably, the neural network model is a multi-layer sensor. The multi-layer perceptron can process structured and unstructured data and is therefore suitable for use with a variety of collected data types, including paperboard dimensional data, process flow data, spray formulation data, performance data, and real-time parameter data, with the transformed and weighted vectors as inputs to the MLP during implementation, and the MLP can be configured to output desired control-by-temperature values for the spray.
Specifically, the multilayer perceptron comprises:
An input layer for receiving the vector result given with the weight;
The hidden layers comprise one or more hidden layers, each hidden layer is composed of a plurality of fully connected nodes, and a Dropout layer is added behind each hidden layer so as to randomly discard part of neurons in the training process;
the Dropout layer is added behind each hidden layer and is used for discarding part of neurons in the hidden layers at a set discarding rate in the training process;
An output layer for generating an output result of the model;
The calculation formula of the discarding rate is:
Wherein Dropoutrate is the discard rate;
w 1,w2,w3 represents the weights of the first vector, the second vector, and the third vector, respectively;
Var (w 1,w2,w3) represents the variance of the three weights;
max (w 1,w2,w3) and Min (w 1,w2,w3) represent the maximum and minimum of the three weights, respectively.
In the multilayer perceptron, partial neurons in the hidden layer can be randomly discarded in the training process by introducing the discarding rate, so that the overfitting of a neural network model to training data is reduced, the generalization capability of the model is improved, and the model is better represented on unseen data; by discarding some neurons, the model is not overly dependent on certain specific neurons, thus enhancing the robustness of the model, which means that the model is more robust to noise and variations in the input data, helping to improve the performance of the model in practical applications. The discard rate may facilitate the convergence speed of the model because each neuron is discarded with a certain probability during the training process, which means that the parameter space in the network is explored more efficiently, thereby speeding up the training process.
The use of the above drop rate formula can bring about the following effects:
More comprehensive weight distribution information: the variance reflects the dispersion degree of the weight distribution, the difference between the maximum value and the minimum value can reflect the weight range, comprehensive consideration of the two indexes can provide more comprehensive weight distribution information, and the dispersion degree of the weight and the weight range are considered.
More accurate discard rate adjustment: the appropriate discard rate can be determined more accurately using the variance, the difference between the maximum and minimum values in combination. If the variance of the weights is large but the difference between the maximum and minimum is small, indicating that the weight distribution is more diffuse but the range of weights is small, then a higher discard rate can be selected to prevent overfitting; conversely, if the variance of the weights is small but the difference between the maximum and minimum is large, indicating that the weights are concentrated in a certain range, but the range is large, then a lower discard rate may be selected to retain more important information.
The stability and generalization capability of the model are improved: the fitting capacity and the generalization capacity of the model can be balanced better by comprehensively considering the differences of the variance, the maximum value and the minimum value, the complexity of the model can be effectively controlled by properly adjusting the discarding rate, and the stability and the generalization capacity of the model are improved, so that better performance is obtained in different data sets and tasks.
A spraying device is used for spraying a functional layer on five-layer paper boards serving as a substrate, and in the spraying process of the functional layer, the spraying liquid temperature control method is adopted for controlling the spraying liquid temperature on the surface layer of the composite six-layer paper board.
The foregoing has outlined and described the basic principles, features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made without departing from the spirit and scope of the invention, which is defined in the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. The surface layer spraying liquid temperature control method of the composite six-layer paperboard is characterized by comprising the following steps of:
Collecting cross-sectional dimension data and process flow data of five-layer paper boards serving as substrates to obtain a first data set;
Collecting formula data, performance data and spraying requirement data of the spraying liquid to obtain a second data set;
Collecting real-time parameters of a position to be sprayed at a set production position, including the temperature and the humidity of the surface, and obtaining a third data set;
Converting the first data set, the second data set and the third data set to obtain a first vector, a second vector and a third vector respectively;
respectively giving corresponding weights to the first vector, the second vector and the third vector;
And inputting the weighted vector result into a neural network model, wherein the neural network model is used for outputting a temperature control value of the spraying liquid.
2. The method of controlling the surface layer spraying liquid temperature of a composite six-layer paperboard according to claim 1, wherein the applying weights to the first vector, the second vector and the third vector, respectively, comprises:
determining a weight index as a criterion for giving weight to each vector; and calculating the weight by adopting a hierarchical analysis method based on the weight index.
3. The method for controlling the surface layer spraying liquid temperature of the composite six-layer paperboard according to claim 2, wherein the analytic hierarchy process comprises:
Determining a hierarchical structure of the decision problem, wherein the hierarchical structure comprises a target layer, a criterion layer and a sub-criterion layer, and each hierarchy comprises a plurality of factors;
Determining the relative importance of each factor in two layers between the criterion layer and the target layer and between all sub-criterion layers in the criterion layer, and correspondingly filling into a judgment matrix;
carrying out consistency test on each judgment matrix, if the judgment results are consistent, continuing, otherwise, adjusting the judgment matrix;
Calculating the feature vector of each judgment matrix to obtain the weight of each factor;
Synthesizing the weights of all layers to obtain the final sorting result of all factors;
And comprehensively processing the sequencing results of the factors in each vector to obtain weights of the first vector, the second vector and the third vector respectively.
4. A surface spraying liquid temperature control method for composite six-ply paperboards according to claim 3, wherein the consistency check of each of said judgment matrices is achieved by calculating a consistency ratio and a random consistency index.
5. A method of controlling the surface spraying liquid temperature of a composite six-ply paperboard according to claim 3, wherein calculating the eigenvectors of each judgment matrix adopts an eigenvalue method.
6. A surface spraying liquid temperature control method for composite six-layer paper board according to claim 3, wherein the weight of each layer is synthesized by a weighted average method.
7. The method of claim 1, wherein converting the first, second, and third data sets to obtain first, second, and third vectors, respectively, comprises:
Preprocessing the first data set, the second data set and the third data set;
Feature extraction is carried out on each preprocessed data set, and primary vectors of three vectors are constructed;
Performing feature engineering on each initial vector, and extracting representative and distinguishing features;
And converting the extracted features into vector forms to respectively obtain the first vector, the second vector and the third vector.
8. The method for controlling the temperature of the surface layer spraying liquid of the composite six-layer paperboard according to claim 1, wherein the neural network model is a multi-layer sensor.
9. The method of controlling the surface layer spraying liquid temperature of a composite six-layer paperboard according to claim 1, wherein the multi-layer sensor comprises:
An input layer for receiving the vector result given with the weight;
The hidden layers comprise one or more hidden layers, each hidden layer is composed of a plurality of fully connected nodes, and a Dropout layer is added behind each hidden layer so as to randomly discard part of neurons in the training process;
the Dropout layers are added behind each hidden layer and are used for discarding part of neurons in the hidden layers at a set discarding rate in the training process;
An output layer for generating an output result of the model;
The calculation formula of the discarding rate is as follows:
Wherein Dropoutrate is the discard rate;
w 1,w2,w3 represents the weights of the first, second and third vectors, respectively;
Var (w 1,w2,w3) represents the variance of the three weights;
max (w 1,w2,w3) and Min (w 1,w2,w3) represent the maximum and minimum of the three weights, respectively.
10. A spraying device, which is characterized in that the spraying device is used for spraying functional layers on five-layer paper boards serving as substrates, and the spraying liquid temperature control method for the surface layer spraying liquid of the composite six-layer paper board is adopted for carrying out spraying liquid temperature control in the spraying process of the functional layers.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410576983.7A CN118364864A (en) | 2024-05-10 | 2024-05-10 | Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410576983.7A CN118364864A (en) | 2024-05-10 | 2024-05-10 | Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118364864A true CN118364864A (en) | 2024-07-19 |
Family
ID=91876385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410576983.7A Pending CN118364864A (en) | 2024-05-10 | 2024-05-10 | Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118364864A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10041882A1 (en) * | 2000-08-25 | 2002-03-07 | Abb Patent Gmbh | Detecting coating thickness distribution in paint coating involves converting real input parameters into input parameters for model of quasi-3D spray figure in trained neural network |
| US20050096796A1 (en) * | 2003-10-30 | 2005-05-05 | Ford Motor Company | Global paint process optimization |
| CN105805884A (en) * | 2015-08-28 | 2016-07-27 | 北京海登赛思工业智能技术有限公司 | Temperature and humidity control method and control system for air conditioners in coating workshops |
| CN116341389A (en) * | 2023-04-25 | 2023-06-27 | 武汉湾流科技股份有限公司 | Intelligent paint spraying method for automobile parts based on neural network |
| CN116390649A (en) * | 2020-10-23 | 2023-07-04 | 阿比尔技术公司 | Apparatus, system and method for product coating |
| CN116449779A (en) * | 2023-04-03 | 2023-07-18 | 安徽信息工程学院 | Actor-Critic structure-based environmental data analysis method for automobile body spraying |
-
2024
- 2024-05-10 CN CN202410576983.7A patent/CN118364864A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10041882A1 (en) * | 2000-08-25 | 2002-03-07 | Abb Patent Gmbh | Detecting coating thickness distribution in paint coating involves converting real input parameters into input parameters for model of quasi-3D spray figure in trained neural network |
| US20050096796A1 (en) * | 2003-10-30 | 2005-05-05 | Ford Motor Company | Global paint process optimization |
| CN105805884A (en) * | 2015-08-28 | 2016-07-27 | 北京海登赛思工业智能技术有限公司 | Temperature and humidity control method and control system for air conditioners in coating workshops |
| CN116390649A (en) * | 2020-10-23 | 2023-07-04 | 阿比尔技术公司 | Apparatus, system and method for product coating |
| CN116449779A (en) * | 2023-04-03 | 2023-07-18 | 安徽信息工程学院 | Actor-Critic structure-based environmental data analysis method for automobile body spraying |
| CN116341389A (en) * | 2023-04-25 | 2023-06-27 | 武汉湾流科技股份有限公司 | Intelligent paint spraying method for automobile parts based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kim et al. | Predicting the household power consumption using CNN-LSTM hybrid networks | |
| Gumus et al. | A multi-echelon inventory management framework for stochastic and fuzzy supply chains | |
| Sari et al. | Backpropagation on neural network method for inflation rate forecasting in Indonesia | |
| Hurrion | An example of simulation optimisation using a neural network metamodel: finding the optimum number of kanbans in a manufacturing system | |
| US10365085B2 (en) | Method and system for measuring thickness of thin film | |
| Zhang et al. | Comparison of econometric models and artificial neural networks algorithms for the prediction of baltic dry index | |
| CN112597392B (en) | Recommendation system based on dynamic attention and hierarchical reinforcement learning | |
| CN109190834A (en) | Stock price trend prediction method and system neural network based | |
| Singh et al. | Application of Box-Jenkins method and Artificial Neural Network procedure for time series forecasting of prices | |
| Goyal et al. | Simulated neural network intelligent computing models for predicting shelf life of soft cakes | |
| CN106372426A (en) | Multi-response parameter optimization method based on principal component analysis and neural network | |
| Šibalija et al. | Taguchi-based and intelligent optimisation of a multi-response process using historical data | |
| CN118364864A (en) | Surface layer spraying liquid temperature control method and spraying device for composite six-layer paperboard | |
| Dewandra et al. | Journal Unique Visitors Forecasting Based on Multivariate Attributes Using CNN | |
| CN109886488A (en) | Distributed wind power plant layered hybrid short-term prediction method considering wind speed time lag | |
| Chiou et al. | Grey prediction model for forecasting the planning material of equipment spare parts in navy of Taiwan | |
| Li et al. | A comparative analysis of predictive data mining techniques | |
| Akyüz | Future projection and the sales of industrial wood in Turkey: artificial neural networks | |
| Ruhil et al. | A comparative study of kinetic and connectionist modeling for shelf-life prediction of Basundi mix | |
| Purkaystha et al. | Product recommendation: A deep learning factorization method using separate learners | |
| CN115754875A (en) | Voltage transformer error evaluation method and device | |
| Hikmawati et al. | A novel hybrid GSTARX-RNN model for forecasting space-time data with calendar variation effect | |
| Napitupulu | Artificial neural network application in gross domestic product forecasting: an Indonesia case | |
| Yousefpour et al. | The prediction of air pollution by using Neuro-fuzzy GMDH | |
| Sholtanyuk | Comparative Analysis of Neural Networking and Regression Models for Time Series Forecasting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |