Process and production line matching degree calculation method, optimization method and device

Abstract

This invention provides a method, optimization method, and apparatus for calculating the matching degree between a process and a production line. The calculation method includes: obtaining the key process parameters and probability distributions of each of the M unit operations constituting the process flow, where M is an integer greater than 1; establishing a process flow model, which includes M unit operation models; inputting the key process parameters and their corresponding probability distributions of each unit operation into the process flow model based on Monte Carlo simulation to obtain output information, including output load distribution, output volume distribution, and output product retention time distribution; and calculating the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters. This invention's method for calculating the matching degree between the process and the production line fills the gap in the lack of quantitative indicators for matching degree as a reference. The optimization method optimizes the mismatch degree without changing the existing production line layout.

Description

Technical Field

[0001] This invention relates primarily to the manufacturing field, and more particularly to a method, computing device, optimization method, optimization device, and computer-readable medium for calculating the matching degree of processes and production lines. Background Technology

[0002] In recent years, improvements in cell lines, culture medium compositions, and feeding strategies have led to a significant increase in the cell culture titers of monoclonal antibodies (mAbs). Higher titers can present facility adaptation challenges for traditional biopharmaceutical purification equipment, whose capabilities were initially designed for lower titers. Bottlenecks caused by equipment size mismatches, coupled with process variability after scale-up, can result in the discard of expensive products.

[0003] A common current strategy is to use production line adaptation software (such as Superpro Designer) to calculate the matching rate between the process and the production line. Traditional adaptation software typically simulates the process flow using a single value (maximum or minimum), without taking into account the randomness of the process flow, and may fail to determine the correct matching rate. Certain worst-case combinations may cause production to exceed equipment capacity, resulting in the need to discard expensive products. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method, optimization method, apparatus and computer-readable medium for calculating the matching degree of process and production line, which can improve the matching accuracy.

[0005] To address the aforementioned technical problems, this invention provides a method for calculating the matching degree between a process and a production line, comprising: obtaining key process parameters and probability distributions for each of the M unit operations constituting the process flow, where M is an integer greater than 1; establishing a process flow model, the process flow model including M unit operation models; inputting the key process parameters and their corresponding probability distributions for each unit operation into the process flow model based on Monte Carlo simulation to obtain output information, the output information including output load distribution, output volume distribution, and output product retention time distribution; and calculating the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters.

[0006] Optionally, key process parameters for each unit operation can be obtained by fitting the distribution of expert knowledge or historical data.

[0007] Optionally, the step of obtaining the key process parameters of each unit operation based on the distribution fitting of historical data includes: determining whether the project to be produced has complete process flow information; if so, obtaining the key process parameters of each unit operation from the historical process flow information of the project; if not, obtaining the key process parameters of each unit operation from the historical process flow information of projects similar to the project.

[0008] Optionally, the step of obtaining the probability distribution of each key process parameter includes matching the most suitable distribution type among a variety of mathematical distributions based on maximum likelihood estimation.

[0009] Optionally, the step of inputting the key process parameters of each unit operation and their corresponding probability distributions into the process flow model based on Monte Carlo simulation includes: Step a: Inputting the key process parameters of each unit operation and their probability distributions into the corresponding unit operation model based on Monte Carlo simulation, with each unit operation model outputting a set of output information; Step b: Repeating step a N times, with each unit operation model outputting N sets of output information, and the process flow model outputting N*M sets of output information, where N is a positive integer greater than 1.

[0010] Optionally, the step of calculating the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters includes: calculating the matching degree between the process flow and the production line in multiple dimensions based on the output information and the production line's limiting parameters; and calculating the overall matching degree between the process and the production line based on the matching degree in the multiple dimensions.

[0011] Optionally, the matching degree of the multiple dimensions includes load capacity matching degree, product collection volume matching degree, and product placement time matching degree.

[0012] Optionally, the overall matching degree between the process and the production line can be calculated using the following formula:

[0013]

[0014] Where P is the total matching degree, P a The matching degree for each dimension, where m is the total number of matching dimensions.

[0015] Optionally, the limiting parameters of the production line include the maximum load. The step of calculating the load matching degree based on the output information and the limiting parameters of the production line includes: obtaining N predicted loads based on the output load distribution of the N sets of output information for each unit operation; comparing the N predicted loads with the maximum load and counting the number of loads that do not exceed the maximum load; dividing the number of loads that do not exceed the maximum load by N to obtain the sub-load matching degree for each unit operation; and multiplying the sub-load matching degrees of M unit operations to obtain the load matching degree.

[0016] Optionally, the limiting parameters of the production line include the maximum collection volume. The step of calculating the product collection volume matching degree based on the output information and the limiting parameters of the production line includes: obtaining N predicted product volumes based on the output volume distribution of the N sets of output information for each unit operation; comparing the N predicted product volumes with the maximum collection volume and counting the number of volumes that do not exceed the maximum collection volume; dividing the number of volumes that do not exceed the maximum collection volume by N to obtain the sub-product collection volume matching degree for each unit operation; and multiplying the sub-product collection volume matching degrees of M unit operations to obtain the product collection volume matching degree.

[0017] Optionally, the limiting parameters of the production line include the maximum placement time. The step of calculating the product placement time matching degree based on the output information and the limiting parameters of the production line includes: calculating N predicted product placement times based on the output product retention time distribution of the N sets of output information for each unit operation; comparing the N predicted product placement times with the maximum placement time and counting the number of times that do not exceed the maximum placement time; dividing the number of times that do not exceed the maximum placement time by N to obtain the sub-product placement time matching degree for each unit operation; and multiplying the sub-product placement time matching degrees of M unit operations to obtain the product placement time matching degree.

[0018] Optionally, the predicted product placement time includes process consumption time, settling time, and rotation time.

[0019] Optionally, the predicted product placement time can be calculated using the following formula:

[0020]

[0021]

[0022]

[0023] Among them, t total The predicted product placement time is denoted as 'a', where 'a' is the step in which the product begins to enter the collection container, and 'n' is the step in which product placement ends. It is the time consumed in the j-th step. It is the settling time for the j-th step. V is the rotation time of the j-th step. j Q is the predicted product volume in the j-th step. j It is the flow rate in the j-th step. It is an estimation factor for the rotation time of the j-th step.

[0024] Optionally, the calculation method further includes: determining whether the total matching degree meets the requirements; if so, selecting the production line for project production; if not, switching to other production lines to calculate the matching degree or optimizing the total matching degree of the production line.

[0025] To address the aforementioned technical problems, this invention provides a method for optimizing the matching degree of a process and a production line, comprising: calculating the matching degree of each unit operation in each dimension using the calculation method described above; calculating the mismatch degree in each dimension based on the matching degree of each dimension; sorting all the mismatch degrees and selecting the unit operation corresponding to the highest mismatch degree as the first optimized unit operation; locating the process bottleneck parameter and the first mismatch parameter range of the process bottleneck parameter in the first optimized unit operation; calculating the first optimized parameter range of the process bottleneck parameter based on the first mismatch parameter range and the allowable range of the process bottleneck parameter; and optimizing the first optimized unit operation based on the first optimized parameter range.

[0026] Optionally, the step of locating the process bottleneck parameter and the first mismatch parameter range in the first optimization unit operation includes: constructing an optimal binary decision tree and visualizing the optimal binary decision tree; and locating the process bottleneck parameter and the first mismatch parameter range in the first optimization unit operation based on the selection operation of the optimal binary decision tree.

[0027] Optionally, the steps of constructing the optimal binary decision tree include: establishing a dataset consisting of features and a target, wherein the features include key process parameters for each unit operation, and the target has a value of 0 or 1; constructing a binary decision tree based on the classification tree in the CART algorithm, and selecting the binary decision tree with the highest accuracy as the optimal binary decision tree.

[0028] Optionally, the value of the target is determined by the following steps: inputting the feature into the process flow model based on Monte Carlo simulation, calculating the mismatch degree of the first optimization unit operation; determining whether the mismatch degree of the first optimization unit operation is greater than a first threshold; if so, setting the value of the target to 1; otherwise, setting the value of the target to 0.

[0029] Optionally, the step of locating the process bottleneck parameter and the first mismatch parameter range of the process bottleneck parameter in the first optimization unit operation according to the optimal binary decision tree includes: locating the branch from the node at the bottom of the optimal binary decision tree containing the most mismatch data to the root node, and taking all the optimal splitting features and optimal splitting points on the branch as the process bottleneck parameter and the first mismatch parameter range, respectively.

[0030] Optionally, the Gini index is used to determine the optimal segmentation feature and the optimal segmentation point.

[0031] Optionally, the features may be filtered before constructing a binary decision tree based on the classification tree in the CART algorithm.

[0032] Optionally, the steps for filtering the features include filtering based on process principles, filtering based on correlation, and manual filtering based on specific rules.

[0033] Optionally, the correlation-based filtering includes calculating the correlation between each feature and the target, and deleting features whose correlation is lower than a second threshold.

[0034] Optionally, the step of optimizing the first optimization unit operation according to the first optimization parameter range includes: optimizing the value of the process bottleneck parameter according to the first optimization parameter range, inputting the optimized value of the process bottleneck parameter into the unit operation model, calculating the optimization mismatch degree of the first optimization unit operation; determining whether the difference between the highest mismatch degree and the optimization mismatch degree is greater than or equal to a third threshold, and if so, the first optimization unit operation is optimized.

[0035] Optionally, the optimization method further includes: taking the unit operation corresponding to one of the other mismatch degrees as the second optimization unit operation; locating the process bottleneck parameter in the second optimization unit operation and the second mismatch parameter range of the process bottleneck parameter; calculating the second optimization parameter range based on the second mismatch parameter range and the allowable range of the process bottleneck parameter; if the process bottleneck parameter in the first optimization unit operation and the process bottleneck parameter in the second optimization unit operation are the same, then taking the intersection of the first optimization parameter range and the second optimization parameter range as the final optimization parameter range, and optimizing the first optimization unit operation and the second optimization unit operation based on the final optimization parameter range.

[0036] To address the aforementioned technical problems, the present invention provides a device for calculating the matching degree of process flow and production line, comprising: a memory for storing instructions executable by a processor; and a processor for executing the instructions to implement the calculation method described above.

[0037] To address the aforementioned technical problems, the present invention provides a device for optimizing the matching degree of process flow and production line, comprising: a memory for storing instructions executable by a processor; and a processor for executing the instructions to implement the optimization method described above.

[0038] To address the aforementioned technical problems, the present invention provides a computer-readable medium storing computer program code, which, when executed by a processor, implements the calculation and optimization methods described above.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] 1. Compared with the traditional method of using the upper and lower limits of all process parameters as inputs to the same process flow model, the matching degree calculation method of the process and production line of the present invention uses Monte Carlo simulation of the process flow model as input, which can better simulate the randomness of the process. Moreover, the output of the process flow model of the present invention can provide a more informative probability distribution, and the possibility of extreme parameter combinations can be controlled by the number of Monte Carlo simulations.

[0041] 2. The matching degree calculation method of the process and production line of the present invention visualizes the matching degree problem of process and production line into a specific multi-dimensional mismatch problem, and fills the gap of no quantitative indicator as a reference for matching degree by combining the input of Monte Carlo simulation process flow model.

[0042] 3. The process and production line matching optimization method of the present invention can improve the matching degree between the process and the production line by restricting key process parameters without changing the existing production line layout.

[0043] 4. The process and production line matching optimization method of the present invention locates the process bottleneck parameters that cause production line mismatch by using a decision tree approach. By adjusting or limiting the values ​​of these process bottleneck parameters, the production line mismatch problem can be effectively solved. Attached Figure Description

[0044] The accompanying drawings are included to provide a further understanding of this application; they are incorporated into and constitute a part of this application. The drawings illustrate embodiments of this application and, together with this specification, serve to explain the principles of the invention. In the drawings:

[0045] Figure 1 This is a schematic diagram of the hierarchical structure of the process according to an embodiment of the present invention;

[0046] Figure 2 This is a flowchart of a method for calculating the matching degree between a process and a production line according to an embodiment of the present invention;

[0047] Figure 3 It corresponds Figure 1 A system block diagram of the process flow model for the medium-sized process;

[0048] Figure 4 This is a schematic diagram of the input and output quantities of a process flow model according to an embodiment of the present invention;

[0049] Figure 5 This is a flowchart of a method for optimizing the matching degree between process and production line according to an embodiment of the present invention;

[0050] Figure 6This is an embodiment of the matching degree optimization method between the process and production line of the present invention;

[0051] Figure 7 This is a schematic diagram of a dataset for constructing an optimal binary decision tree according to an embodiment of the present invention;

[0052] Figure 8 This is a schematic diagram of an optimal binary decision tree according to an embodiment of the present invention;

[0053] Figure 9 This refers to the process bottleneck parameter and the range of the first mismatch parameter in the first optimization unit operation of an embodiment of the present invention;

[0054] Figure 10 This refers to the process bottleneck parameter and the range of the second mismatch parameter in the second optimization unit operation of an embodiment of the present invention;

[0055] Figure 11 yes Figure 9 and Figure 10 A schematic diagram of the final optimized parameter range;

[0056] Figure 12 This is a schematic diagram illustrating the production line capacity according to an embodiment of the present invention;

[0057] Figure 13 This is a system block diagram of a process and production line matching degree calculation device according to an embodiment of the present invention. Detailed Implementation

[0058] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this application. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0059] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0060] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0061] Furthermore, it should be noted that the use of terms such as "first" and "second" to define components is merely for the purpose of distinguishing the corresponding components. Unless otherwise stated, these terms have no special meaning and therefore should not be construed as limiting the scope of protection of this application. In addition, although the terminology used in this application is selected from commonly known and used terms, some terms mentioned in this application's specification may have been chosen by the applicant according to his or her judgment, and their detailed meanings are explained in the relevant sections of this description. Moreover, this application should be understood not only through the actual terms used, but also through the meaning implied by each term.

[0062] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more steps may be removed from these processes.

[0063] A process in a project often comprises multiple unit operations, each consisting of several process steps. A unit operation refers to a series of basic operations in the chemical industry and other process industries that cause the desired physical changes in materials, such as crushing, conveying, heating, cooling, mixing, and separating. Various unit operations, based on different physicochemical principles and using appropriate equipment, achieve their respective process objectives. For example, distillation, based on the differences in the volatility of components in a liquid mixture, can achieve the separation of components or the purification of a specific component. Figure 1 This is a schematic diagram of the hierarchical structure of the process in an embodiment of this application. For example... Figure 1As shown, process 1 includes unit operations 11, 12, 13, 14, 15, and 16. Unit operation 13 includes process steps 131, 132, 133, 134, and 135. Each unit operation may have one or more critical process parameters (CCPs).

[0064] A production line is a combination of equipment used to implement processes for production. For example, the product might be a chemical product, especially a pharmaceutical one. Production lines have limiting parameters determined by their design characteristics, such as load capacity, total tank capacity, and hold time limits. These limiting parameters may not meet the critical process parameters of each unit operation within the process. Therefore, it is necessary to calculate the fit between the process and the production line.

[0065] Figure 2 This is a flowchart of a method for calculating the matching degree between a process and a production line according to an embodiment of the present invention. For example... Figure 2 As shown, the method 200 for calculating the matching degree between the process and the production line includes the following steps:

[0066] Step S21: Obtain the key process parameters of each unit operation in the M unit operations that make up the process flow and the probability distribution of each key process parameter, where M is an integer greater than 1;

[0067] Step S22: Establish a process flow model, which includes M unit operation models;

[0068] Step S23: Based on Monte Carlo simulation, input the key process parameters of each unit operation and their corresponding probability distribution into the process flow model to obtain output information, including output load distribution, output volume distribution and output product retention time distribution.

[0069] Step S24: Calculate the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters.

[0070] The following provides a detailed explanation of steps S21 to S24.

[0071] In step S21, the key process parameters for each unit operation can be obtained based on expert knowledge. For example, a template for the key process parameters of each unit operation can be pre-set and saved based on expert knowledge, and the key process parameters of each unit operation can be obtained by retrieving the template. This method is useful for projects lacking complete process information in the early stages, as it allows for the calculation of the production line with the highest matching rate based on currently known information, thereby improving the success rate. The key process parameters for each unit operation can also be obtained by fitting the distribution of historical data. The probability distribution of each key process parameter can be obtained by matching the most suitable distribution type among various mathematical distributions based on maximum likelihood estimation. In this embodiment, the probability distribution of each key process parameter is obtained by fitting the distribution. Compared with relying solely on personal experience, this method better supports the calculation of probability distributions in subsequent process flow simulations and better assigns statistical significance to the simulation results.

[0072] In some embodiments, the step of fitting the distribution of historical data to obtain the key process parameters of each unit operation includes: determining whether the project to be produced has complete process flow information; if so, fitting the actual data distribution in the process flow information to obtain the key process parameters of each unit operation; otherwise, fitting the historical data distribution of similar projects to obtain the key process parameters of each unit operation.

[0073] In step S22, based on the law of conservation of mass, a process flow model consisting of unit operation models is established. The number of unit operation models corresponds to the number of unit operations. Figure 3 It corresponds Figure 1 A system block diagram of the process flow model for the medium-sized process. For example... Figure 3 As shown, the process flow model 300 includes unit operation model 31, unit operation model 32, unit operation model 33, unit operation model 34, unit operation model 35 and unit operation model 36.

[0074] In step S23, the step of inputting the key process parameters and their corresponding probability distributions for each unit operation into the process flow model based on Monte Carlo simulation includes: inputting the key process parameters and their probability distributions for each unit operation into the corresponding unit operation model based on Monte Carlo simulation, and each unit operation model outputting a set of output information. This step is repeated N times, with each unit operation model outputting N sets of output information and the process flow model outputting N*M sets of output information, where N is a positive integer greater than 1. Figure 4 This is a schematic diagram of the input and output quantities of a process flow model according to an embodiment of the present invention. Figure 4As shown, the inputs to the process flow model 400 are the key process parameters and their probability distributions for each unit operation. The probability distributions for the key process parameters include distributions for key process parameter 1 (411), key process parameter 2 (412), and key process parameter 3 (413). The inputs to the process flow model also include an input volume distribution (414) and an input mass distribution (415). These distributions are either the initial inputs or the outputs of the previous unit operation model. The process flow model 400 outputs a set of information, including but not limited to output volume distribution (421), output mass distribution (422), output load distribution (423), and output product retention time distribution (424).

[0075] By controlling the number of Monte Carlo simulations (N), the likelihood of extreme parameter combinations can be controlled, resulting in a more realistic production simulation. The output of a process flow model using the Monte Carlo method will have a corresponding numerical probability distribution. This numerical probability distribution can provide a more accurate estimate of the output in the process flow, including but not limited to upper and lower limits, expected values, etc., providing a quantitative indicator based on probability calculations for subsequent production line matching.

[0076] In step S24, the steps for calculating the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters are as follows: First, the matching degree between the process flow and the production line is calculated in multiple dimensions based on the output information and the production line's limiting parameters. Here, the matching degree in multiple dimensions includes, but is not limited to, load capacity matching degree, product collection volume matching degree, and product placement time matching degree. Then, the overall matching degree between the process and the production line is calculated based on the matching degree in multiple dimensions. The overall matching degree between the process and the production line is calculated using the following formula:

[0077]

[0078] Where P is the total matching degree, P a The matching degree for each dimension, where m is the total number of matching dimensions.

[0079] Load capacity can be further divided into volumetric load capacity and mass load capacity, commonly used in calculations for unit operations related to chromatography and filtration. When the product volume and mass of the unit operation input exceed the load capacity range, the product exceeding the maximum load capacity will be lost. In some embodiments, the production line's limiting parameters include the maximum load capacity, and the step of calculating the load capacity matching degree based on the output information and the production line's limiting parameters includes:

[0080] (1) Obtain N predicted loads based on the output load distribution of N sets of output information for each unit operation;

[0081] (2) Compare each of the N predicted loads with the maximum load and count the number of loads that exceed the maximum load.

[0082] (3) Divide the number of operations that do not exceed the maximum load by N to obtain the sub-load matching degree for each unit operation. For example, it can be expressed by the following formula:

[0083]

[0084] Among them, P i N represents the subload matching degree of the i-th unit operation. i NN represents the number of predicted loads exceeding the maximum load in the i-th unit operation. i The number of predicted loads that do not exceed the maximum load in the i-th unit operation, where N is the number of Monte Carlo simulations, which is equal to the number of output information groups.

[0085] (4) Multiply the sub-load matching degrees of the M unit operations together to obtain the load matching degree. This can be expressed by the following formula:

[0086]

[0087] Where P1 is the load matching degree, P i Let M be the subload matching degree of the i-th unit operation, and M be the number of unit operations.

[0088] In some embodiments, the production line's limiting parameters include a custom first load, which is less than the maximum load. The first load has a safety margin compared to the maximum load. The step of calculating the load matching degree based on the output information and the production line's limiting parameters includes: obtaining N predicted loads based on N sets of output information for each unit operation; comparing the N predicted loads with the first load and counting the number of loads that do not exceed the first load; dividing the number of loads that do not exceed the first load by N to obtain the sub-load matching degree for each unit operation; and multiplying the sub-load matching degrees of M unit operations to obtain the total load matching degree.

[0089] Due to limitations in production line design, the size and number of product collection containers within the space where unit operations are performed are restricted; this restriction is referred to as the product collection container volume. The maximum collection volume within its space is the sum of the volumes of all containers that can be placed there. When the product volume output by the unit operation exceeds the maximum collection volume supported by the space, the excess product volume is lost. In some embodiments, the production line's limiting parameters include the maximum collection volume. The step of calculating the product collection volume matching degree based on the output information and the production line's limiting parameters includes:

[0090] (1) Based on the output volume distribution of N sets of output information for each unit operation, N predicted product volumes are obtained;

[0091] (2) Compare the N predicted product volumes with the maximum collection volume, and count the number of products that do not exceed the maximum collection volume;

[0092] (3) Divide the number of units that do not exceed the maximum collection volume by N to obtain the sub-product collection volume matching degree for each unit operation. For example, it can be expressed by the following formula:

[0093]

[0094] Among them, P j Collect volume matching degree for the sub-product of the j-th unit operation, N j Let NN be the number of predicted product volumes exceeding the maximum collection volume in the j-th unit operation. j Let N be the number of predicted product volumes that do not exceed the maximum collection volume in the j-th unit operation, and let N be the number of Monte Carlo simulations, which is equal to the number of output information groups.

[0095] (4) The product collection volume matching degree is obtained by multiplying the sub-product collection volume matching degrees of the M unit operations. This can be expressed by the following formula:

[0096]

[0097] Where P2 represents the product collection volume matching degree, P j The volume matching degree is collected for the sub-product of the j-th unit operation, and M is the number of unit operations.

[0098] In some embodiments, the production line limiting parameters include a custom first collection volume, which is smaller than the maximum collection volume. The first collection volume has a safety margin compared to the maximum collection volume. The step of calculating the product collection volume matching degree based on the output information and the production line limiting parameters includes: obtaining N predicted product volumes based on N sets of output information for each unit operation; comparing the N predicted product volumes with the first collection volume and counting the number of volumes that do not exceed the first collection volume; dividing the number of volumes that do not exceed the first collection volume by N to obtain the sub-product collection volume matching degree for each unit operation; and multiplying the sub-product collection volume matching degrees of M unit operations to obtain the product collection volume matching degree.

[0099] The hold time limit for product stability is determined during the process development phase by testing the chemical stability of the product and identifying the longest possible holding time to maintain stability. In a unit operation, the holding time is timed from the moment a product enters the product collection container in the first process step and ends when the product enters the next process step. A unit operation involves multiple process steps. The time consumed by each process step consists of process time and settling time. The total time of a unit operation includes the sum of the process step times, plus time for operations such as rotation between process steps. This time can be calculated by statistically analyzing historical data from different production lines performing different types of unit operations and converting it into an estimation factor. In some embodiments, the production line's limiting parameters include the maximum holding time. The step of calculating the product holding time matching degree based on the output information and the production line's limiting parameters includes:

[0100] (1) Calculate N predicted product placement times based on the output product retention time distribution of N sets of output information for each unit operation. The predicted product placement time includes process consumption time, settling time, and rotation time. The process consumption time can be obtained by dividing the predicted volume by the flow rate. The predicted product placement time is calculated using the following formula:

[0101]

[0102]

[0103]

[0104] Among them, t total The predicted product placement time is denoted as 'a', where 'a' is the step in which the product begins to enter the collection container, and 'n' is the step in which product placement ends. It is the time consumed in the j-th step. It is the settling time for the j-th step. V is the rotation time of the j-th step. j Q is the predicted product volume in the j-th step. j It is the flow rate in the j-th step. This is an estimation factor for the turnaround time of the j-th step. The value of this estimation factor is determined by the actual production line conditions and the types of unit operations involved, and is calculated based on historical data. Some process steps are pre-processing steps, which are independent of the product; let's assume these are process steps 1 to 3. In process step 4, a product enters the collection container, and the product retention time begins. Similarly, some process steps are post-processing steps, which are independent of the product; let's assume this is process step 21. Therefore, the actual product retention time is the time required for process steps 4 to 20.

[0105] (2) Compare the predicted placement time of N products with the maximum placement time, and count the number of products that do not exceed the maximum placement time;

[0106] (3) Divide the number of items that have not exceeded the maximum placement time by N to obtain the sub-product placement time matching degree for each unit operation. For example, it can be expressed by the following formula:

[0107]

[0108] Among them, P k Collect volume matching degree for the sub-product of the k-th unit operation, N k NN represents the number of products whose placement time exceeds the maximum placement time in the k-th unit operation. k N represents the number of predicted product placement times that do not exceed the maximum placement time in the k-th unit operation, and N is the number of Monte Carlo simulations, which is equal to the number of output information groups.

[0109] (4) Multiply the product placement time matching degrees of the M unit operations together to obtain the product placement time matching degree. This can be expressed by the following formula:

[0110]

[0111] Where P3 represents the product placement time matching degree, P k Let M be the sub-product placement time matching degree of the k-th unit operation, and M be the number of unit operations.

[0112] In some embodiments, the production line's limiting parameters include a custom first placement time, which is less than the maximum placement time. The first placement time has a safety margin compared to the maximum placement time. The step of calculating the product placement time matching degree based on the output information and the production line's limiting parameters includes: calculating N predicted product placement times based on N sets of output information for each unit operation; comparing the N predicted product placement times with the first placement time and counting the number of times that do not exceed the first placement time; dividing the number of times that do not exceed the first placement time by N to obtain the sub-product placement time matching degree for each unit operation; and multiplying the sub-product placement time matching degrees of M unit operations to obtain the final product placement time matching degree.

[0113] In some embodiments, the method for calculating the matching degree between the process and the production line further includes the steps of: determining whether the total matching degree meets the requirements; if so, selecting a production line for project production; if not, switching to another production line to calculate the matching degree or optimizing the total matching degree of the production line.

[0114] Compared to the traditional approach of using the upper and lower limits of all process parameters as input to the same process flow model, this embodiment uses Monte Carlo simulation of the process flow model as input, which can better simulate the randomness of the process. Furthermore, the output of the process flow model in this embodiment can provide a more informative probability distribution, and the probability of extreme parameter combinations can be controlled by the number of Monte Carlo simulations performed. The process-production line matching degree calculation method in this embodiment concretizes the process-production line matching degree problem into a specific multi-dimensional mismatch problem, and, combined with the input of the Monte Carlo simulation process flow model, fills the gap where there are no quantitative indicators for matching degree as a reference.

[0115] Figure 5 This is a flowchart of a method for optimizing the matching degree between a process and a production line according to an embodiment of the present invention. For example... Figure 5 As shown, the process and production line matching optimization method 500 includes the following steps:

[0116] Step S51: Calculate the matching degree of each unit operation in each dimension using the matching degree calculation method of process and production line mentioned above, and calculate the mismatch degree in each dimension based on the matching degree of each dimension.

[0117] Step S52: Sort all mismatches and select the unit operation corresponding to the highest mismatch as the first optimization unit operation.

[0118] Step S53: Locate the process bottleneck parameter and the first mismatch parameter range of the process bottleneck parameter in the first optimization unit operation.

[0119] Step S54: Calculate the first optimized parameter range of the process bottleneck parameter based on the first mismatch parameter range and the allowable range of the process bottleneck parameter, and optimize the first optimized unit operation based on the first optimized parameter range.

[0120] Figure 6 This is an embodiment of the matching degree optimization method between the process and production line of the present invention. The following is combined with... Figure 5 and Figure 6 Steps S51-S54 are explained in detail.

[0121] In step S51, assume P a To calculate the mismatch for each dimension based on the matching degree of each dimension, the formula is: 1-P a The results of the mismatch degree for each dimension of each unit operation after calculation are as follows: Figure 6 As shown, the process flow model 600 includes unit operation models 61, 62, 63, 64, 65, and 66. Each unit operation model includes product collection volume mismatch, load mismatch, and product placement time mismatch.

[0122] In step S52, all mismatches are sorted from high to low. The sorting result is: product collection volume mismatch of unit operation model 65 > product collection volume mismatch of unit operation model 66 > product collection volume mismatch of unit operation model 64. The mismatches of other dimensions are all 0. Based on the sorting result, unit operation 5 corresponding to unit operation model 65 is designated as the first optimized unit operation, denoted as S1; unit operation 6 corresponding to unit operation model 66 is designated as the second optimized unit operation, denoted as S2; and unit operation 4 corresponding to unit operation model 64 is designated as the third optimized unit operation, denoted as S3. First, the first optimized unit operation S1 is optimized. A common optimization measure is to change the production line layout and add collection containers. However, in actual production, it is difficult to change the production line layout due to space constraints or equipment quantity limitations. This embodiment describes optimizing the mismatch without changing the existing production line layout. Specifically, the process bottleneck parameters causing the production line mismatch are located by using a decision tree approach. Adjusting or limiting the values ​​of these process bottleneck parameters can effectively solve the production line mismatch problem.

[0123] In step S53, the steps for locating the process bottleneck parameters and the first mismatch parameter range of the process bottleneck parameters in the first optimization unit operation are as follows: First, construct the optimal binary decision tree and visualize it. Second, locate the process bottleneck parameters and the first mismatch parameter range of the first optimization unit operation based on the selection operation of the optimal binary decision tree.

[0124] In one example, the steps to construct an optimal binary decision tree include:

[0125] a. Establish a dataset consisting of features and a target. Features include key process parameters for each unit operation, and the target value is either 0 or 1. The target value is determined through the following steps: input the features into the process flow model based on Monte Carlo simulation, calculate the mismatch degree of the first optimized unit operation; determine whether the mismatch degree of the first optimized unit operation is greater than a first threshold. If it is, set the target value to 1; otherwise, set the target value to 0. Figure 7 This is a schematic diagram of the dataset used to construct the optimal binary decision tree according to an embodiment of the present invention. Figure 7 As shown, dataset 700 includes features X and the target Y. Feature X consists of key process parameters for each unit operation. The number of features is determined by the number of key process parameters corresponding to all unit operations involved in the process, and the number of rows in feature X equals the number of Monte Carlo simulations. That is, 'a' in feature X... 11 For the key process parameters of unit operation 1, a 1n For the key process parameters of unit operation n, a mn These are the key process parameters for unit operation n during the m-th Monte Carlo simulation. Each row of values ​​in the objective Y corresponds to a Monte Carlo random parameter combination for each row of the feature X. Values ​​are assigned to specific matching problems (such as the first optimization unit S1), with a value of 0 for a match and 1 for a mismatch.

[0126] In some embodiments, to achieve faster training and higher accuracy, and to eliminate the interference of irrelevant features on the binary decision tree, feature filtering is included before constructing the binary decision tree based on the classification tree in the CART algorithm. The feature filtering steps include, but are not limited to, process-based filtering, correlation-based filtering, and manual filtering based on specific rules. Process-based filtering refers to filtering features based on mechanisms and process principles. For example, when unit operation 5 is the first optimized unit operation S1, all key process parameters of unit operation 6 can be removed from feature X. Since the first optimized unit operation S1 to be optimized is related to the collection volume, key process parameters in all unit operations that do not participate in the collection volume calculation process, such as the flow rate for calculating placement time, can be removed from feature X. Correlation-based filtering includes calculating the correlation between each feature and the target, and deleting features with a correlation below a second threshold. Specifically, feature X and target Y are put into a logistic regression model, and hypothesis testing is performed on all remaining key process parameters and the target to calculate the p-value. Key process parameters with p-values ​​greater than a second threshold (typically 0.05) are removed because this indicates that the key process parameter has no statistical correlation with the target Y, and thus these key process parameters can be removed. This correlation is not limited to linear correlation and is not subject to a normal distribution. Manual screening based on specific rules can be performed after completing screening based on process principles and correlation, reviewing the remaining key process parameters, and manually removing those that are difficult to control or cannot be modified.

[0127] After feature selection, training and validation sets are created based on the dataset. When creating the training set, attention should be paid to its balance, which helps improve the accuracy of the decision tree and better pinpoint bottleneck process parameters and corresponding thresholds. Specifically, in 100,000 Monte Carlo simulations, the probability of mismatch in unit operation 5 is 3.5%, meaning there are 3,500 rows with a target Y value of 1. In this case, 3,500 rows should be selected from the remaining 96,500 feature combinations with a target Y value of 0 to form a 7,000-row training set, making the ratio of 0 to 1 close to 50:50, instead of using a training set containing all 100,000 rows. Otherwise, even if the binary decision tree predicts the target as 0 for all feature combinations, it can still achieve an accuracy of 96.5%, which clearly cannot define accurate bottleneck process parameters in this situation.

[0128] b. Construct a binary decision tree based on the classification tree in the CART (Classification and Regression Tree) algorithm, and select the binary decision tree with the highest accuracy as the optimal binary decision tree. The binary decision tree is selected using the Gini index to determine the optimal splitting feature and the optimal splitting point, and K-fold cross-validation is used to test the robustness of the model. The specific steps are as follows:

[0129] i. Divide the filtered dataset into k subsets, use k-1 subsets as the training set of the model, and use 1 subset as the validation set of the model;

[0130] ii. Based on the training set, recursively construct a binary decision tree for each node, starting from the root node, by performing the following steps:

[0131] 1) Let the dataset entering the node be T. Calculate the Gini index of each feature X with respect to dataset T based on each feature X and each of its possible values ​​x. Select the feature X with the smallest Gini index. opt and its value x opt As the optimal splitting feature and the optimal splitting point, two child nodes are generated from the current node, and the dataset T is split into datasets T1 and T2 according to the value of the optimal splitting feature being greater than or less than or equal to the optimal splitting point, and then distributed to the two child nodes.

[0132] 2) Repeat the previous steps for each child node to perform further splitting until the decision tree depth is controlled by the minimum number of samples required under the leaf node, the maximum number of leaf nodes, or the minimum amount of impurity reduction required to perform the splitting. The depth is usually 3 layers.

[0133] iii. Replace a subset of the training set with the subset used as the validation set, repeat ii, and output the average prediction accuracy after k times.

[0134] iv. Adjust the hyperparameters, repeat steps i, ii and iii, compare the average prediction accuracy, and select the binary decision tree with the highest accuracy for visualization.

[0135] After visualizing the binary decision tree for the user, the user can select the optimal binary decision tree. The method in this embodiment can locate the process bottleneck parameter and the range of the first mismatch parameter in the first optimization unit operation based on the selection operation of the optimal binary decision tree. Specifically, it locates the branch from the node at the bottom of the optimal binary decision tree containing the most mismatched data to the root node, and uses all the optimal splitting features and optimal splitting points on the branch as the process bottleneck parameter and the range of the first mismatch parameter, respectively. Figure 8This is a schematic diagram of an optimal binary decision tree according to an embodiment of the present invention. Figure 8 As shown, in the optimal binary decision tree 800, the branch from the node with the most mismatched data at the bottom to the root node is the branch with A > 40 and B > 2.85. Then, the optimal splitting features A and B are both process bottleneck parameters. The range of the first mismatched parameter is denoted as U1, and the value of U1 is A > 40 and B > 2.85.

[0136] Continue back to Figure 1 As shown, in step S54, the first optimized parameter range of the process bottleneck parameter is calculated based on the first mismatch parameter range and the allowable range of the process bottleneck parameter. Specifically, the first optimized parameter range of the process bottleneck parameter is the difference between the allowable range of the process bottleneck parameter and the first mismatch parameter range. Figure 9 This refers to the range of process bottleneck parameters and first mismatch parameters in the first optimization unit operation of an embodiment of the present invention. For example... Figure 9 As shown, the process bottleneck parameters of the first optimization unit operation S1 are A and B. The allowable range of process bottleneck parameter A is [0, 50], and the allowable range of process bottleneck parameter B is [2, 5]. The value of the first mismatch parameter range U1 is {A > 40 and B > 2.85}, then the value of the first optimization parameter range G1 is {0 ≤ A ≤ 40 and 2 ≤ B ≤ 2.85}.

[0137] In some embodiments, the steps for optimizing the first optimization unit operation according to the first optimization parameter range are as follows: First, optimize the value of the process bottleneck parameter according to the first optimization parameter range, input the optimized process bottleneck parameter value into the unit operation model, and calculate the optimization mismatch degree of the first optimization unit operation. Then, determine whether the difference between the highest mismatch degree and the optimized mismatch degree is greater than or equal to a third threshold. If so, the first optimization unit operation is optimized. In other words, determine whether there is a significant decrease between the optimized mismatch degree and the highest mismatch degree (unoptimized). If there is a significant decrease, it means that the optimization is successful. If not, it is necessary to return to the feature filtering step to re-filter features. In practice, it has been found that in most cases, the mismatch degree can be significantly reduced by simultaneously restricting multiple bottleneck process parameters. After optimizing the value of the process bottleneck parameter according to the first optimization parameter range G1 and inputting the optimized process bottleneck parameter value into the unit operation model, the mismatch degree of the first optimization unit operation S1 decreased from 3.5% to 0.07%.

[0138] In some embodiments, the process and production line matching optimization method further includes the following steps: First, the unit operation corresponding to one of the other mismatches is taken as the second optimized unit operation. Second, the process bottleneck parameter and the second mismatch parameter range of the process bottleneck parameter in the second optimized unit operation are located. Then, the second optimized parameter range is calculated based on the second mismatch parameter range and the allowable range of the process bottleneck parameter. For example, as... Figure 6 As shown, the product collection volume mismatch degree of unit operation model 66 is the second highest. Unit operation 6 corresponding to unit operation model 66 is taken as the second optimized unit operation, denoted as S2. An optimal binary decision tree is constructed for the second optimized unit operation S2. Based on the optimal binary decision tree, the process bottleneck parameter and the range of the second mismatch parameter of the process bottleneck parameter in the second optimized unit operation are located. Figure 10 This refers to the range of process bottleneck parameters and second mismatch parameters in the second optimization unit operation of an embodiment of the present invention. For example... Figure 10 As shown, the bottleneck parameters of the second optimization unit operation S2 are A, B, and D. The allowable range of bottleneck parameter A is [0, 50], the allowable range of bottleneck parameter B is [2, 5], and the allowable range of bottleneck parameter D is [0, 5]. The value of the second mismatch parameter range U2 is {A > 30 and B > 4.25 and D > 2.24}, then the value of the second optimization parameter range G2 is {0 ≤ A ≤ 30 and 2 ≤ B ≤ 4.25 and 2 ≤ D ≤ 2.24}. If the bottleneck parameters in the first optimization unit operation and the second optimization unit operation are the same, then the intersection of the first optimization parameter range and the second optimization parameter range is taken as the final optimization parameter range, and the first and second optimization unit operations are optimized according to the final optimization parameter range. Figure 11 yes Figure 9 and Figure 10 A schematic diagram illustrating the final optimized parameter range. (See diagram below.) Figure 11 As shown, the final optimized parameter range is denoted as Gfinal, and the value of Gfinal is the intersection of the first optimized parameter range G1 and the second optimized parameter range G2. The value of Gfinal is {0≤A≤30 and 2≤B≤2.85 and 2≤D≤2.24}.

[0139] In some embodiments, when the first optimization parameter range G1 and the second optimization parameter range G2 do not intersect, it indicates that the first optimization unit operation S1 and the second optimization unit operation S2 cannot be optimized simultaneously. The first optimization parameter range G1 can be used as the final optimization parameter range Gfinal, prioritizing the optimization of the first optimization unit operation S1. Alternatively, conflicting key process parameters can be removed from feature X, and then the first optimization unit operation S1 and the second optimization unit operation S2 can be optimized simultaneously using other key process parameters.

[0140] The process and production line matching optimization method in this embodiment improves the matching degree between the process and the production line by restricting key process parameters and optimizing the mismatch without changing the existing production line layout. By using a decision tree approach to locate the process bottleneck parameters causing the production line mismatch, adjusting or restricting the values ​​of these bottleneck parameters can effectively solve the production line mismatch problem.

[0141] In some embodiments, it is possible to quickly determine whether a project using the platform process can be adapted to the target production line based on the key process bottleneck parameters and the range of mismatch parameters of the process bottleneck parameters on the production line. Alternatively, if the target production line is determined, restrictions and guidance can be imposed on the key process parameters experimentally selected for the project developed using the platform process to ensure that they can fully match the production line capabilities. Figure 12 This is a schematic diagram illustrating the production line capacity according to an embodiment of the present invention. Figure 12 Due to differences in production line design, the critical bottleneck parameters and the range of mismatched bottleneck parameters for the same process may differ on different production lines. Process 1 has a critical bottleneck parameter and a range of mismatched bottleneck parameters 121 on production line P1, and a range of mismatched bottleneck parameters 122 on production line P2. When a project plans to use Process 1, a rapid matching can be performed on production lines P1 and P2 solely by comparing the bottleneck process parameters, selecting the production line whose bottleneck process parameter value range meets the project requirements. This application can perform critical process parameter analysis on each existing production line for each fixed process, and the analysis results can be visualized as production line capabilities, thereby assisting in the rapid matching of processes or guiding process development.

[0142] Figure 13 This is a system block diagram of a process and production line matching calculation device 1300 (hereinafter referred to as calculation device 1300) according to an embodiment of the present invention. (See reference) Figure 13As shown, the computing device 1300 may include an internal communication bus 1301, a processor 1302, a read-only memory (ROM) 1303, a random access memory (RAM) 1304, and a communication port 1305. When applied to a personal computer, the computing device 1300 may also include a hard disk 1306. The internal communication bus 1301 enables data communication between components of the computing device 1300. The processor 1302 can make judgments and issue prompts. In some embodiments, the processor 1302 may consist of one or more processors. The communication port 1305 enables data communication between the computing device 1300 and external devices. In some embodiments, the computing device 1300 can send and receive information and data from a network through the communication port 1305. The computing device 1300 may also include different forms of program storage units and data storage units, such as the hard disk 1306, the read-only memory (ROM) 1303, and the random access memory (RAM) 1304, capable of storing various data files used for computer processing and / or communication, as well as possible program instructions executed by the processor 1302. The processor executes these instructions to implement the main part of the method. The results processed by the processor are transmitted to the user equipment via the communication port and displayed on the user interface. The above-described operation method can be implemented as a computer program, stored in the hard disk 1306, and loaded into the processor 1302 for execution to implement the process and production line matching calculation and optimization method of this application.

[0143] The present invention also includes a computer-readable medium storing computer program code that, when executed by a processor, implements the aforementioned process and production line matching calculation method and optimization method.

[0144] When the methods for calculating and optimizing the matching degree of processes and production lines are implemented as computer programs, they can also be stored as articles of manufacture in computer-readable storage media. For example, computer-readable storage media may include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic stripes), optical discs (e.g., compact discs (CDs), digital multifunction discs (DVDs)), smart cards, and flash memory devices (e.g., electrically erasable programmable read-only memory (EPROM), cards, sticks, key drives). Furthermore, the various storage media described herein can represent one or more devices and / or other machine-readable media used for storing information. The term "machine-readable medium" may include, but is not limited to, wireless channels and various other media (and / or storage media) capable of storing, containing, and / or carrying code and / or instructions and / or data.

[0145] The present invention also provides a process and production line matching degree optimization device. The structure of the process and production line matching degree optimization device can be referred to the process and production line matching degree calculation device 1300, and will not be described in detail here.

[0146] Some aspects of this application can be executed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The aforementioned hardware or software may be referred to as a "data block," "module," "engine," "unit," "component," or "system." The processor may be one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DAPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, or combinations thereof. Furthermore, aspects of this application may manifest as computer products residing in one or more computer-readable media, including computer-readable program code. For example, computer-readable media may include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic tapes, etc.), optical discs (e.g., compressed CDs, digital multifunction DVDs, etc.), smart cards, and flash memory devices (e.g., cards, sticks, key drives, etc.).

[0147] A computer-readable medium may contain a propagated data signal containing computer program code, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and so on, or suitable combinations thereof. A computer-readable medium can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program code located on the computer-readable medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, radio frequency signals, or similar media, or any combination of the above media.

[0148] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this disclosure method does not imply that the subject matter of the application requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of the single embodiments disclosed above.

[0149] The basic concepts have been described above. Obviously, for those skilled in the art, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

[0150] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of the application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0151] Although this application has been described with reference to specific embodiments, those skilled in the art should recognize that the above embodiments are only used to illustrate this application, and various equivalent changes or substitutions can be made without departing from the spirit of this application. Therefore, any changes or modifications to the above embodiments within the essential spirit of this application will fall within the scope of the claims of this application.

Claims

1. A method of calculating matching degree of a process and a production line, characterized by, include: Obtain the key process parameters and probability distribution of each of the M unit operations that make up the process flow, where M is an integer greater than 1. Establish a process flow model, which includes M unit operation models; Based on Monte Carlo simulation, the key process parameters of each unit operation and their corresponding probability distribution are input into the process flow model to obtain output information, which includes output load distribution, output volume distribution and output product retention time distribution. The overall matching degree between the process and the production line is calculated based on the output information and the production line's limiting parameters; wherein, the Monte Carlo simulation-based input of the key process parameters of each unit operation and their corresponding probability distributions into the process flow model includes: Step a: Based on Monte Carlo simulation, input the key process parameters of each unit operation and the probability distribution of the key process parameters into the corresponding unit operation model, and each unit operation model outputs a set of output information; Step b: Repeat step a N times, with each unit operation model outputting N sets of output information, and the process flow model outputting N... M groups of output information, where N is a positive integer greater than 1; The step of calculating the overall matching degree between the process and the production line based on the output information and the production line's limiting parameters includes: The matching degree between the process flow and the production line in multiple dimensions is calculated based on the output information and the production line's limiting parameters; The overall matching degree between the process and the production line is calculated based on the matching degree of the multiple dimensions.

2. The method as described in claim 1, characterized in that, Key process parameters for each unit operation are obtained by fitting the distribution of expert knowledge or historical data.

3. The method of claim 2, wherein, The steps for obtaining key process parameters for each unit operation based on historical data distribution fitting include: Determine whether the project to be produced has complete process flow information. If so, obtain the key process parameters of each unit operation from the historical process flow information of the project. If not, obtain the key process parameters of each unit operation from the historical process flow information of projects similar to the project.

4. The method of claim 1, wherein, The steps to obtain the probability distribution of each key process parameter include matching the most suitable distribution type among a variety of mathematical distributions based on maximum likelihood estimation.

5. The method of claim 1, wherein, The matching degree across multiple dimensions includes load capacity matching degree, product collection volume matching degree, and product placement time matching degree.

6. The method of claim 1, wherein, The overall matching degree between the process and the production line is calculated using the following formula: where P is the total match degree, P a is the match degree for each dimension, and m is the total number of matching dimensions.

7. The method of claim 5, wherein, The production line's limiting parameters include the maximum load capacity. The step of calculating the load capacity matching degree based on the output information and the production line's limiting parameters includes: Based on the output load distribution of the N sets of output information for each unit operation, N predicted loads are obtained; Compare the N predicted loads with the maximum load, and count the number of loads that do not exceed the maximum load. Divide the number of units that do not exceed the maximum load by N to obtain the sub-load matching degree of each unit operation; The load matching degree is obtained by multiplying the sub-load matching degrees of M unit operations.

8. The method of claim 5, wherein, The limiting parameters of the production line include the maximum collection volume. The step of calculating the product collection volume matching degree based on the output information and the limiting parameters of the production line includes: Based on the output volume distribution of the N sets of output information for each unit operation, N predicted product volumes are obtained; Compare the N predicted product volumes with the maximum collection volume, and count the number of products that do not exceed the maximum collection volume; Divide the number of units that do not exceed the maximum collection volume by N to obtain the sub-product collection volume matching degree for each unit operation; The product collection volume matching degree is obtained by multiplying the sub-product collection volume matching degrees of M unit operations.

9. The method of claim 5, wherein, The production line's limiting parameters include the maximum placement time. The step of calculating the product placement time matching degree based on the output information and the production line's limiting parameters includes: Calculate N predicted product placement times based on the product retention time distribution of the N sets of output information for each unit operation; Compare the predicted placement time of N products with the maximum placement time, and count the number of products that do not exceed the maximum placement time; Divide the number of products that have not exceeded the maximum placement time by N to obtain the sub-product placement time matching degree for each unit operation; The product placement time matching degree is obtained by multiplying the sub-product placement time matching degrees of M unit operations.

10. The method of claim 9, wherein, The predicted product placement time includes process consumption time, settling time, and rotation time.

11. The method as described in claim 10, characterized in that, The predicted product placement time is calculated using the following formula: in, The predicted product placement time is denoted as 'a', where 'a' is the step in which the product begins to enter the collection container, and 'n' is the step in which product placement ends. It is the time consumed in the j-th step. It is the settling time for the j-th step. It is the rotation time of the j-th step. This is the predicted product volume for the j-th step. It is the flow rate in the j-th step. It is an estimation factor for the rotation time of the j-th step.

12. The method of claim 1, wherein, Also includes: Determine whether the total matching degree meets the requirements. If yes, select the production line for project production. If no, switch to another production line to calculate the matching degree or optimize the total matching degree of the production line.

13. A method of optimizing matching of a process with a production line, characterized by, include: The matching degree of each unit operation in each dimension is calculated by the method described in any one of claims 5-11, and the mismatch degree of each dimension is calculated based on the matching degree of each dimension. Sort all mismatches and select the unit operation with the highest mismatch as the first optimization unit operation. Locate the process bottleneck parameter in the first optimization unit operation and the first mismatch parameter range of the process bottleneck parameter; The first optimized parameter range of the process bottleneck parameter is calculated based on the first mismatch parameter range and the allowable range of the process bottleneck parameter, and the first optimization unit operation is optimized based on the first optimized parameter range.

14. The method of claim 13, wherein, The steps of locating the process bottleneck parameter and the first mismatch parameter range of the process bottleneck parameter in the first optimization unit operation include: Construct the optimal binary decision tree and visualize it; The process bottleneck parameters and the range of the first mismatch parameters in the first optimization unit operation are located based on the selection operation of the optimal binary decision tree.

15. The method of claim 14, wherein, The steps to construct an optimal binary decision tree include: Establish a dataset consisting of features and targets, where the features include key process parameters for each unit operation, and the targets have values ​​of 0 or 1; A binary decision tree is constructed based on the classification tree in the CART algorithm, and the binary decision tree with the highest accuracy is selected as the optimal binary decision tree.

16. The method of claim 15, wherein, The value of the target is determined through the following steps: The features are input into the process flow model based on Monte Carlo simulation to calculate the mismatch degree of the first optimization unit operation; Determine whether the mismatch degree of the first optimization unit operation is greater than a first threshold. If it is, set the value of the target to 1; otherwise, set the value of the target to 0.

17. The method of claim 14, wherein, The steps of locating the process bottleneck parameter and the first mismatch parameter range of the process bottleneck parameter in the first optimization unit operation according to the optimal binary decision tree include: locating the branch from the node at the bottom of the optimal binary decision tree containing the most mismatch data to the root node, and taking all the optimal splitting features and optimal splitting points on the branch as the process bottleneck parameter and the first mismatch parameter range, respectively.

18. The method of claim 17, wherein, The optimal segmentation feature and the optimal segmentation point are determined using the Gini index.

19. The method as described in claim 15, characterized in that, Before constructing a binary decision tree based on the classification tree in the CART algorithm, the features are further filtered.

20. The method of claim 19, wherein, The steps for filtering the features include filtering based on process principles, filtering based on correlation, and manual filtering based on specific rules.

21. The method of claim 20, wherein, Relevance-based filtering involves calculating the correlation between each feature and the target, and deleting features whose correlation is below a second threshold.

22. The method of claim 13, wherein, The steps for optimizing the operation of the first optimization unit based on the first optimization parameter range include: The value of the process bottleneck parameter is optimized according to the first optimization parameter range, and the optimized value of the process bottleneck parameter is input into the unit operation model to calculate the optimization mismatch degree of the first optimization unit operation. Determine whether the difference between the highest mismatch degree and the optimized mismatch degree is greater than or equal to a third threshold. If so, the first optimization unit operation completes the optimization.

23. The method of claim 13, wherein, Also includes: Use the unit operation corresponding to one of the other mismatch degrees as the second optimization unit operation; Locate the process bottleneck parameter and the second mismatch parameter range of the process bottleneck parameter in the second optimization unit operation; The second optimized parameter range is calculated based on the second mismatch parameter range and the allowable range of the process bottleneck parameter; If the process bottleneck parameter in the first optimization unit operation is the same as the process bottleneck parameter in the second optimization unit operation, then the intersection of the first optimization parameter range and the second optimization parameter range is taken as the final optimization parameter range, and the first optimization unit operation and the second optimization unit operation are optimized according to the final optimization parameter range.

24. A device for calculating the matching degree between a process and a production line, comprising: Memory is used to store instructions that can be executed by the processor; A processor for executing the instructions to implement the method as described in any one of claims 1-12.

25. A device for optimizing the matching degree of a process and a production line, comprising: Memory is used to store instructions that can be executed by the processor; A processor for executing the instructions to implement the method as described in any one of claims 13-23.

26. A computer-readable medium storing computer program code that, when executed by a processor, implements the method as claimed in any one of claims 1-23.

Images