A multi-dimensional data processing method for a fully digitalized aprothioazole production line

Abstract

The present invention relates to the technical field of propanthioazole production management, and specifically to a multi-dimensional data processing method for a fully digital propanthioazole production line. The present invention first obtains experimental data and trial production data of propanthioazole; further in the experimental data, according to the fluctuation correlation characteristics of the target parameter and the purity of the corresponding intermediate, the experimental correlation coefficient of the target parameter is obtained; further in the trial production data, the trial production records are divided into trial production record groups and the target record group of the target parameter is screened out; further in each trial production record, according to the fluctuation characteristics of the purity of the intermediate, the purity of the intermediate is corrected; further in the target record group, according to the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, the trial production correlation coefficient of the target parameter is obtained; finally, according to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, the optimal experimental range of the target parameter is corrected to obtain the optimal production range.

Description

Technical Field

[0001] The present invention relates to the technical field of propanthioazole production management, and in particular to a multi-dimensional data processing method for a fully digital propanthioazole production line. Background Art

[0002] Propylthiozole is a new type of fungicide developed through a series of processes such as structural modification and activity screening of triazole compounds. It has excellent properties such as reduced toxicity and enhanced systemic conductivity. The researchers continuously adjusted the concentration ratio of the reactants to obtain high-purity intermediates, and carried out trial production on the pharmaceutical production line with the optimal reactant concentration ratio.

[0003] Usually, it is carried out under small-scale conditions in the laboratory, with a relatively small sample size, and the entire experimental process is closer to the ideal state. However, there are scale changes from the laboratory to the trial production process, and many complex factors that need to be considered in large-scale production. For example, when the equipment size increases, factors such as heat transfer and material mixing uniformity will change to a certain extent. At this time, the process parameter range obtained in the laboratory may not be applicable in actual production, making the trial production results less than ideal, which ultimately affects the actual production effect. Summary of the invention

[0004] In order to solve the technical problem that the process parameter range obtained by experiment is not applicable in actual production, the purpose of the present invention is to provide a multi-dimensional data processing method for a fully digitalized propanthiophene production line, and the technical scheme adopted is as follows:

[0005] Obtaining experimental data and trial production data of propanthioazole; the experimental data and trial production data at least include the purity of the intermediate of each process and the data of each process parameter; determining the optimal experimental range of each process parameter of each process according to the experimental data; and conducting trial production after the experiment is completed and according to the optimal experimental range;

[0006] Select any one of the process parameters of any of the steps as a target parameter, and other process parameters in the same step as other parameters; in the experimental data, obtain an experimental correlation coefficient between the target parameter and the corresponding intermediate according to the fluctuation correlation characteristics between the target parameter and the purity of the corresponding intermediate;

[0007] In the trial production data, the trial production records are arbitrarily combined in a preset ratio to form a number of trial production record groups; in each of the trial production record groups, according to the violent fluctuation characteristics of the other parameters, the target record group of the target parameter is screened out; the process of the target parameter is used as the target process; in each of the trial production records, according to the fluctuation characteristics of the purity of the intermediate before the target process, combined with the purity of the intermediate of the target process, the corrected purity of the intermediate of the target process is obtained; in the target record group, according to the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, the trial production correlation coefficient of the target parameter and the corresponding intermediate is obtained;

[0008] According to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, combined with the optimal experimental range of the target parameter, the optimal production range of the target parameter is obtained.

[0009] Furthermore, the method for obtaining the target record group includes:

[0010] In each of the trial production record groups, at least according to the variance of each of the other parameters in all the trial production records, obtain the fluctuation severity coefficient of each of the other parameters; at least according to the mean of the fluctuation severity coefficients of all the other parameters, obtain the disturbance coefficient of each of the trial production record groups;

[0011] A target record group is obtained according to the disturbance coefficient.

[0012] Furthermore, the method for obtaining a target record group according to the disturbance coefficient includes:

[0013] The trial production record group with the smallest disturbance coefficient is selected as the target record group.

[0014] Furthermore, the method for obtaining the corrected purity of the intermediate includes:

[0015] In each of the trial production records, when the target process is the first process, the purity of the intermediate of the target process is directly used as the corrected purity of the intermediate;

[0016] When the target process is not the first process, the purity of the intermediates of the target process and the preceding process sequentially preceding the target process is used to form a purity sequence; in the purity sequence of the target process, the sum of the differences in the purity of the intermediates of all adjacent processes is used as the process conduction coefficient of the target process; based on the purity of the intermediates of the target process and the process conduction coefficient, the corrected purity of the intermediates of the target process corresponding to the trial production record is obtained;

[0017] Among the differences in the purity of the intermediates of the adjacent processes, the purity of the intermediate of the process with the smallest order is the subtrahend; and the process conductance coefficient is negatively correlated with the corrected purity of the intermediate.

[0018] Further, the method for obtaining the corrected purity of the intermediate of the target process corresponding to the trial production record according to the purity of the intermediate of the target process and the process conductance coefficient includes:

[0019] The purity of the intermediate of the target process is taken as a first minuend, the product of the purity of the intermediate and the process conductance coefficient is taken as a second minuend, and the difference between the first minuend and the second minuend is taken as a corrected purity of the intermediate of the target process.

[0020] Furthermore, the method for obtaining the experimental correlation coefficient includes:

[0021] In the experimental data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence corresponding to the purity of the intermediate is used as the experimental correlation coefficient between the target parameter and the corresponding intermediate.

[0022] Furthermore, the method for obtaining the trial production correlation coefficient includes:

[0023] In the target record group of the trial production data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence of the corrected purity of the corresponding intermediate is used as the trial production correlation coefficient between the target parameter and the corresponding intermediate.

[0024] Furthermore, the method for obtaining the optimal production range includes:

[0025] The ratio of the experimental correlation coefficient between the target parameter and the intermediate and the trial production correlation coefficient is used as the production application rate of the target parameter;

[0026] Obtain the median and range radius of the optimal experimental range of the target parameter, and use the product of the range radius and the production application rate as the corrected range radius; obtain the optimal production range of the target parameter based on the corrected range radius and in combination with the median.

[0027] Furthermore, after obtaining the optimal production range, the method further includes:

[0028] A secondary trial production is carried out within the optimal production range of each process parameter in each of the steps, and a polynomial regression model is constructed according to each process parameter and the purity of the intermediate to obtain the optimal process parameter value of each process parameter in each of the steps.

[0029] Furthermore, the process parameters include at least temperature, stirring speed, reaction time and pressure.

[0030] The present invention has the following beneficial effects:

[0031] The present invention first obtains experimental data and trial production data of propanthioazole to provide a data basis for subsequent analysis; further determines the optimal experimental range of each process parameter of each process according to the experimental data, provides a design basis for trial production, avoids purposeless waste of resources, and improves the efficiency of determining the production parameters suitable for actual production of propanthioazole; further, in the experimental data, according to the fluctuation correlation characteristics of the target parameter and the purity of the corresponding intermediate, obtains the experimental correlation coefficient of the target parameter and the corresponding intermediate, and characterizes the experimental correlation characteristics of the target parameter and the corresponding intermediate; further, in the trial production data, divides the trial production records into dry trial production record groups and screens out the target record groups of the target parameters, controls the data range of each analysis, avoids the interference of too much irrelevant data that cannot reflect the correlation characteristics of the target parameter and the corresponding intermediate, and avoids the single trial production record from being able to comprehensively reflect the correlation characteristics of the target parameter and the corresponding intermediate. Reflect the influence of the correlation and change trend between the process parameters and the purity of the intermediates; further in each trial production record, according to the fluctuation characteristics of the purity of the intermediates before the target process, combined with the purity of the intermediates of the target process, obtain the corrected purity of the intermediates of the target process, reduce the interference of the process before the target process on the purity of the intermediates of the target process, and provide more accurate data basis for analyzing the production correlation characteristics of the target parameters and the intermediates; further in the target record group, according to the fluctuation correlation characteristics of the target parameters and the corrected purity of the corresponding intermediates, obtain the trial production correlation coefficient of the target parameters and the corresponding intermediates, and characterize the correlation characteristics of the target parameters and the corresponding intermediates in the actual trial production; finally, according to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, combined with the optimal experimental range of the target parameters, obtain the optimal production range of the target parameters. The present invention analyzes the correlation differences between process parameters and intermediates in experiments and trial production, adaptively adjusts to obtain the optimal production range suitable for production, corrects the gap between experiments and production, and provides accurate adjustment space for process parameters in production. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 A flowchart of a multi-dimensional data processing method for a fully digitalized propanthiophene production line provided by one embodiment of the present invention;

[0034] Figure 2 A flowchart of a method for obtaining corrected purity provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a multi-dimensional data processing method for a fully digitalized propanthioazole production line proposed by the present invention, in combination with the accompanying drawings and preferred embodiments, and its specific implementation, structure, characteristics and effects are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] 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.

[0037] The specific scheme of the multi-dimensional data processing method for a fully digitalized propanthiophene production line provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] See also Figure 1 , which shows a flowchart of a multi-dimensional data processing method for a fully digitalized propanthiophene production line provided by an embodiment of the present invention, specifically comprising:

[0039] Step S1: Obtain experimental data and trial production data of propanthioazole; both the experimental data and the trial production data include at least the purity of the intermediate of each process and the data of each process parameter; determine the optimal experimental range of each process parameter of each process according to the experimental data; and the trial production is carried out after the experiment is completed and according to the optimal experimental range.

[0040] In the embodiment of the present invention, the experimental data and trial production data of propanethiolide are first obtained to provide a basis for data analysis; since the process parameter setting of each process step will affect the purity of the intermediate, and ultimately affect the intermediate of the next process, and continue to accumulate until it affects the final product, the experimental data and trial production data at least include the purity of the intermediate of each process step and the data of each process parameter.

[0041] In one embodiment of the present invention, the process parameters include at least temperature, stirring speed, reaction time and pressure; the process of producing propanthioazole includes:

[0042] 1. Replacement reaction: CPCB-F, ethanol or methanol, and hydrazine hydrate are added to the first reactor, the temperature is raised under nitrogen protection, the ethanol is distilled and then the temperature is lowered, toluene is added, stirred and allowed to stand, and the aqueous phase is transferred to the second reactor.

[0043] 2. Extraction and preservation: Add toluene to the second reactor for extraction, transfer the aqueous phase to the hydrazine hydrate recovery tank, and transfer the organic phase to the third reactor and store it under nitrogen as CPCB-275 toluene solution.

[0044] 3. Addition reaction: Cool the CPCB-275 toluene solution in the third reaction kettle, add formaldehyde dropwise, and stir while keeping warm.

[0045] 4. Post-treatment: After the addition reaction is completed, add water, stir, centrifuge, rinse with toluene and water, and dry at a certain temperature to obtain CPCB-346 dry product.

[0046] 5. Oxidation reaction: Add methanol and CPCB-346 dry product to the fourth reactor, cool, and add in batches , replace nitrogen, introduce oxygen and maintain pressure, control the temperature to react until there is no pressure change, and recover methanol by reduced pressure distillation.

[0047] 6. Extraction and separation: Add ethyl acetate and water to the fourth reactor, stir and let stand, then transfer the aqueous phase to the fifth reactor, then add ethyl acetate to the fifth reactor for extraction, transfer the aqueous phase to the wastewater station, and transfer the organic phase to the sixth reactor.

[0048] 7. Extraction and distillation: add water to the sixth reactor, stir and let stand, transfer the aqueous phase to the iron salt recovery reactor, recover the ethyl acetate by vacuum distillation, add toluene to the iron salt recovery reactor to evaporate, then add toluene to heat and dissolve, add n-heptane, cool and centrifuge, rinse with a toluene n-heptane mixed solution.

[0049] 8. Drying to obtain the finished product: Drying the rinsed solid phase to obtain propanthioazole.

[0050] Before trial production, the optimal experimental range of each process parameter of each process is determined based on experimental data; trial production is carried out after the experiment and based on the optimal experimental range to avoid purposeless waste of resources and improve the efficiency of determining the production parameters of propidazole suitable for actual production.

[0051] It should be noted that the intermediate corresponding to each process is the substance obtained after each process is completed. The intermediate obtained after each process may be a variety of required substances. In this case, the purity of the intermediate is the average value of the purity of all required substances.

[0052] It should be noted that the data collection method for various process parameters is already an existing technology; the process for producing propanthioazole is already an existing technology. By controlling the variable method, a single process parameter is changed, and other process parameters are fixed. According to the purity of the intermediate in the experiment and the purity requirement for this intermediate in the experimental manual, the optimal experimental range of the single process parameter can be determined; within the optimal experimental range obtained after the end of the experiment, propanthioazole trial production is carried out to obtain trial production data, which are all technical means well known to those skilled in the art and will not be repeated here.

[0053] Step S2: Select any process parameter of any process as the target parameter, and other process parameters in the same process as other parameters; in the experimental data, obtain the experimental correlation coefficient of the target parameter and the corresponding intermediate according to the fluctuation correlation characteristics of the target parameter and the purity of the corresponding intermediate.

[0054] In order to analyze the impact of process parameters on experiments and production one by one, and thus analyze the scenario error in the process of converting experiments to trial production, the applicable scope of process parameters is adjusted. First, any process parameter of any process is selected as the target parameter, and other process parameters in the same process are selected as other parameters.

[0055] Taking into account the fluctuation correlation characteristics between a certain process parameter and the purity of the intermediate, it reflects the influence of the target parameter on the purity of the intermediate during the experiment. Therefore, in the experimental data, the experimental correlation coefficient of the target parameter and the corresponding intermediate is obtained according to the fluctuation correlation characteristics of the target parameter and the purity of the corresponding intermediate, and the experimental correlation characteristics between the target parameter and the corresponding intermediate are characterized. This is for the subsequent combination of the correlation characteristics in the actual trial production, the analysis of the impact of the mode change, and preparation for corrections to adapt to production.

[0056] Preferably, in one embodiment of the present invention, considering that the more the Pearson coefficients of different data sequences deviate from 0, the stronger the fluctuation correlation of the different data sequences is, so in the experimental data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence of the purity of the corresponding intermediate is used as the experimental correlation coefficient between the target parameter and the corresponding intermediate.

[0057] It should be noted that in other embodiments of the present invention, the implementer can also obtain the absolute value of the cosine similarity between the sequence of the target parameter and the sequence of the purity of the corresponding intermediate, obtain the experimental correlation coefficient, and express the fluctuation correlation of different data sequences through cosine similarity. Both the Pearson correlation coefficient and the cosine similarity have been implemented in technology and will not be described in detail here.

[0058] Step S3: In the trial production data, the trial production records are arbitrarily combined in a preset ratio to form several trial production record groups; in each trial production record group, according to the violent fluctuation characteristics of other parameters, the target record group of the target parameter is screened out; the process of the target parameter is taken as the target process; in each trial production record, according to the fluctuation characteristics of the purity of the intermediate before the target process, combined with the purity of the intermediate of the target process, the corrected purity of the intermediate of the target process is obtained; in the target record group, according to the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, the trial production correlation coefficient of the target parameter and the corresponding intermediate is obtained.

[0059] After analyzing the experimental data, it is necessary to analyze the trial production data to obtain the trial production correlation coefficients between the target parameters and the corresponding intermediates, so as to compare them with the experimental correlation coefficients and analyze the impact of the production mode change.

[0060] Taking into account that in order to test the most suitable process parameters for actual production, trial production usually traverses the optimal experimental range of each target parameter, and obtains a large number of trial production records. Not all production records can reflect the correlation characteristics between the target parameters and the corresponding intermediates; at the same time, a single trial production record only represents the results under a specific combination of process parameters, and cannot fully reflect the correlation and changing trend between the process parameters and the purity of the intermediates. Therefore, in the trial production data, the trial production records are arbitrarily combined in a preset proportion to form several trial production record groups to control the data range of each analysis. This is suitable for parallel computing to improve computing efficiency while improving the accuracy of the trial production correlation coefficient.

[0061] Preferably, in one embodiment of the present invention, the preset ratio is , assuming that there are N trial production records, and considering that N may not be an integer multiple of 3, there are A trial production record group, where C represents the number of combinations in the permutation and combination, Indicates rounding down.

[0062] Taking into account that different trial production record groups have different reflection effects on the correlation between target parameters and corresponding intermediates, it is necessary to screen out the target record group first; considering that the drastic fluctuation characteristics of other parameters represent the degree of interference of other parameters, in each trial production record group, the target record group of the target parameter is screened out according to the drastic fluctuation characteristics of other parameters, so as to better analyze the correlation between the target parameter and the corresponding intermediate and improve the accuracy of the trial production correlation coefficient.

[0063] Preferably, in one embodiment of the present invention, considering that different process parameters may have mutual influences, in order to accurately evaluate the influence of a specific process parameter on the purity of the intermediate, other related parameters must be controlled to avoid interference caused by these interactions, so the more violent the fluctuation of other parameters, the stronger the interference; considering that the variance of data is usually used to represent the violent fluctuation characteristics of a data set;

[0064] Based on this, in each trial production record group, at least according to the variance of each other parameter in all trial production records, obtain the fluctuation intensity coefficient of each other parameter; at least according to the mean of the fluctuation intensity coefficients of all other parameters, obtain the disturbance coefficient of each trial production record group;

[0065] Get the target record group according to the disturbance coefficient.

[0066] As an example, the calculation formula of the perturbation coefficient includes:

[0067] ;

[0068] Wherein, m represents the serial number of the trial production record group; represents the disturbance coefficient of the mth trial production record group; I represents the number of types of all process parameters; I-1 represents the number of other parameters except the target parameters; i represents the serial number of other parameters; represents the linear normalization function; It represents the data set of the i-th other parameters in the m-th trial production record group; represents the variance of the data set of the i-th other parameter in the m-th trial production record group; It represents the fluctuation intensity coefficient of the i-th other parameter in the m-th trial production record group.

[0069] In the calculation formula of the disturbance coefficient, the variance of other parameters is normalized and used as the fluctuation coefficient, and the fluctuation characteristics of the data are represented by the variance; the average value of the fluctuation coefficients of all other parameters is used to represent the overall fluctuation characteristics of other parameters. The larger the average value, the more violent the fluctuation of other parameters, the easier it is to affect the purity of the intermediate, the greater the interference to the correlation between the analysis target parameter and the corresponding intermediate, and the larger the disturbance coefficient.

[0070] In other embodiments of the present invention, considering that the mean absolute deviation also reflects the characteristics of violent fluctuations in data, the violent fluctuation coefficient can also be obtained by combining the mean absolute deviation and the variance, such as or , It represents the average absolute deviation of the ith other parameter in the mth trial production record group. Similarly, the range of the data reflects the fluctuation range of the data, and can also represent the characteristics of severe fluctuations in the data. It can also be combined with the range and variance of other data, or with the range, variance and average absolute deviation, and fused by a positive correlation method of addition, weighted summation or multiplication to obtain the severe fluctuation coefficient. These are all existing technologies and will not be described in detail.

[0071] In other embodiments of the present invention, implementers may also combine the mean, mode and median of the volatility coefficient to jointly represent the overall volatility characteristics of all other parameters, such as taking the weighted sum of the mean, mode and median of the volatility coefficient with weighted weights of 0.6, 0.2 and 0.2 as the disturbance coefficient of the corresponding production record group.

[0072] Preferably, in one embodiment of the present invention, since the smaller the disturbance coefficient, the more relatively stable the other parameters in the trial production record group are, and the more conducive to analyzing the correlation between the target parameters and the corresponding intermediates, the trial production record group with the smallest disturbance coefficient is selected as the target record group.

[0073] Considering that in the manufacturing production process, the corresponding intermediates generated after each process will enter the next process with their own purity characteristics, and gradually accumulate the impact on the purity of the final product. The purity of the intermediates in each process will be affected by the previous processes. For example, when the purity of the intermediates in the first process is low, the purity of the intermediates in the subsequent process will usually be lower. However, the low purity is not caused by the design of the current process. Therefore, the purity of the intermediates in the target process also needs to be corrected according to the cumulative impact characteristics of the purity.

[0074] The process of the target parameters is taken as the target process; the fluctuation of the purity of the intermediates before the target process is taken into account, which reflects the cumulative impact of purity. Therefore, in each trial production record, the corrected purity of the intermediates of the target process is obtained based on the fluctuation characteristics of the purity of the intermediates before the target process and the purity of the intermediates of the target process, so as to provide more accurate data basis for analyzing the production correlation characteristics of the target parameters and the intermediates.

[0075] Preferably, in one embodiment of the present invention, the method for obtaining the corrected purity of the intermediate comprises:

[0076] See also Figure 2 , which shows a flow chart of a method for obtaining a corrected purity provided by an embodiment of the present invention, specifically comprising:

[0077] Step S301: In each trial production record, when the target process is the first process, the purity of the intermediate of the target process is directly used as the corrected purity of the intermediate.

[0078] Considering that the intermediates generated after the first process will not be affected by other processes, in each trial production record, when the target process is the first process, the purity of the intermediates of the target process is directly used as the corrected purity of the intermediates.

[0079] Step S302: When the target process is not the first process, the purity of the intermediates of the target process and the preceding process in sequence before the target process is used to form a purity sequence; in the purity sequence of the target process, the sum of the differences in the purity of the intermediates of all adjacent processes is used as the process transfer coefficient of the target process; based on the purity of the intermediates of the target process and the process transfer coefficient, the corrected purity of the intermediates of the target process corresponding to the pilot production record is obtained.

[0080] Considering that the purity of the product of the last process, i.e. the final process, will also be affected by the previous processes, when the target process is the last process, the purity of the intermediates also needs to be corrected. Therefore, when the target process is not the first process, the purity of the intermediates of the target process and the preceding process before the target process constitutes a purity sequence.

[0081] Taking into account the difference in purity of intermediates between adjacent processes, it reflects the cumulative influence of the purity of adjacent processes. The sum of the purity differences of all adjacent processes in the purity sequence reflects the overall cumulative influence in the purity sequence. Therefore, in the purity sequence of the target process, the sum of the purity differences of intermediates of all adjacent processes is used as the process transfer coefficient of the target process.

[0082] And because we are analyzing the impact of the previous process on the next process, we need to analyze the change in the purity of the next process relative to the purity of the previous process. Therefore, in the difference in the purity of the intermediates of adjacent processes, the purity of the intermediate of the process with the smallest order is the subtrahend.

[0083] The calculation formula of process transfer coefficient includes:

[0084] ;

[0085] Where n represents the serial number of the trial production record; represents the process conduction coefficient of the target process in the nth trial production record; j represents the sequence number of the purity sequence of the target process; J represents the number of the purity sequence of the target process; Indicates the value of the jth purity of the purity sequence of the target process in the nth trial production record; Represents the value of the j+1th purity in the purity sequence of the target process in the nth trial production record.

[0086] In the calculation formula of the process transfer coefficient, the fluctuation characteristics of the purity of the intermediate before the target process are expressed by summing the purity differences of all adjacent processes in the purity sequence. When the process transfer coefficient is positive and the larger it is, the purity of the intermediate of the current target process may be affected by the previous process and is too large, and needs to be reduced and corrected; when the process transfer coefficient is negative and the smaller it is, the purity of the intermediate of the current target process may be affected by the previous process and is too small, and needs to be increased and corrected. Therefore, the process transfer coefficient is negatively correlated with the corrected purity of the intermediate.

[0087] Preferably, in one embodiment of the present invention, according to the purity of the intermediate of the target process and the process conductivity coefficient, the method for obtaining the corrected purity of the intermediate of the target process corresponding to the pilot production record includes:

[0088] The purity of the intermediate of the target process is taken as the first minuend, the product of the purity of the intermediate and the process conductance coefficient is taken as the second minuend, and the difference between the first minuend and the second minuend is taken as the corrected purity of the intermediate of the target process.

[0089] The calculation formula for corrected purity includes:

[0090] ;

[0091] Where n represents the serial number of the trial production record; The value indicating the corrected purity of the intermediate of the target process in the nth trial production record; It represents the process conduction coefficient of the target process in the nth trial production record; A value indicating the purity of the intermediate of the target process in the nth trial production record.

[0092] In the calculation formula of corrected purity, The calculation simplification shows that is the difference between the first and last steps of the purity sequence. Since the purity value is usually a percentage, that is, in the range of 0-1, the difference between the two purities must not exceed 1. When is positive, is corrected by shrinking; when When it is a negative value, The correction is increased to provide an accurate basis for obtaining the target parameters and the trial production correlation characteristics of the corresponding intermediates.

[0093] It should be noted that, in another embodiment of the present invention, the implementer may also construct a purity sequence in which the purity sequence only includes the preceding process of the target process, and only revise the purity of the target process according to the change in purity in the preceding process:

[0094] Considering that the first process has no predecessor process and the second process has only one predecessor process, it is impossible to analyze the fluctuation characteristics, so the first two processes of producing propanthioazole are excluded, and the purity of the intermediates of the first two processes is directly used as the corrected purity; for the target process that is not the first two processes, a purity sequence corresponding to the predecessor process of the target process is constructed, and the purity sequence is analyzed in the same way as step S302 to obtain the corrected purity.

[0095] After correcting the purity of the intermediate, the trial production correlation coefficient of the target parameter and the corresponding intermediate can be obtained in the target record group based on the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, thereby characterizing the correlation characteristics between the target parameter and the corresponding intermediate in the actual trial production and providing a basis for comparing experiments and production changes.

[0096] Preferably, in one embodiment of the present invention, in the target record group of trial production data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence of the corrected purity of the corresponding intermediate is used as the trial production correlation coefficient between the target parameter and the corresponding intermediate.

[0097] Similarly, in other embodiments of the present invention, the implementer may also obtain the experimental correlation coefficient through cosine similarity; however, it should be noted that the experimental correlation coefficient and the trial production correlation coefficient are obtained using the same algorithm, that is, both use the Pearson correlation coefficient or both use cosine similarity.

[0098] Step S4: According to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, combined with the optimal experimental range of the target parameter, the optimal production range of the target parameter is obtained.

[0099] Through step S2 and step S3, the correlation characteristics between the target parameters and the corresponding intermediates are analyzed from the experimental perspective and the trial production perspective respectively, and after obtaining the experimental correlation coefficient and the trial production correlation coefficient that characterize the correlation relationship, the relative changes between the two can be analyzed to obtain the impact of the pattern change from the experiment to the production process, so as to adjust the optimal experimental range, obtain the optimal production range of the target parameters suitable for production, correct the gap between the experiment and production, and provide accurate adjustment space for the process parameters in production, which is conducive to shortening the debugging time and reducing resource waste.

[0100] Preferably, in one embodiment of the present invention, in order to express the relative relationship between the trial production correlation coefficient and the experimental correlation coefficient, the ratio of the experimental correlation coefficient of the target parameter and the corresponding intermediate to the trial production correlation coefficient is used as the production application rate of the target parameter;

[0101] Considering that when the production application rate is less than 1, it means that the trial production correlation coefficient is larger than the experimental correlation coefficient, which reflects that the influence of the target parameters on the product purity in actual production is greater than the laboratory theoretical expectation, and the target parameters need to be more strictly controlled. Therefore, the optimal experimental range needs to be narrowed; on the contrary, when the production application rate is greater than 1, the optimal experimental range needs to be increased. Based on this, the median and range radius of the optimal experimental range of the target parameter are obtained, and the product of the range radius and the production application rate is used as the corrected range radius; according to the corrected range radius and combined with the median, the optimal production range of the target parameter is obtained.

[0102] The calculation formula for the correction range radius includes:

[0103] ;

[0104] in, Indicates the correction range radius of the target parameter; Indicates the range radius of the target parameter; represents the experimental correlation coefficient between the target parameter and the corresponding intermediate; It represents the correlation coefficient between the target parameter and the pilot production of the corresponding intermediate; Indicates the production application rate of the target parameter.

[0105] In the calculation formula of the corrected range radius, the range radius is corrected by determining the median value and range radius of the optimal experimental range, and the optimal production range is obtained by combining the corrected range radius and the median value of the optimal experimental range; the median value of the optimal experimental range is also the center point of the optimal experimental range; for example, if the optimal experimental range is [30,50], the median value is 40, and the range radius is 10, if the corresponding If it is 0.5, the correction range radius is 5 and the optimal production range is [35,45].

[0106] It should be noted that, in one embodiment of the present invention, after obtaining the optimal production range, the following steps are further included:

[0107] A secondary trial production is carried out within the optimal production range of each process parameter in each process step. A polynomial regression model is constructed based on each process parameter and the purity of the intermediate to obtain the optimal process parameter value of each process parameter in each process step. Finally, the optimal process parameter values ​​of all process parameters in all processes are applied to actual production.

[0108] It should be noted that the polynomial regression model is already an existing technology and will not be described in detail here.

[0109] In summary, the present invention aims at the technical problem that the range of process parameters obtained by experiment is not applicable in actual production, and proposes a multi-dimensional data processing method for a fully digitalized propanthioazole production line. The present invention first obtains the experimental data and trial production data of propanthioazole; further in the experimental data, according to the fluctuation correlation characteristics of the target parameter and the purity of the corresponding intermediate, the experimental correlation coefficient of the target parameter is obtained; further in the trial production data, the trial production records are divided into trial production record groups and the target record group of the target parameter is screened out; further in each trial production record, according to the fluctuation characteristics of the purity of the intermediate, the purity of the intermediate is corrected; further in the target record group, according to the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, the trial production correlation coefficient of the target parameter is obtained; finally, according to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, the optimal experimental range of the target parameter is corrected to obtain the optimal production range.

[0110] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0111] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A multi-dimensional data processing method for a fully digitalized propanthiophene production line, characterized in that: The method comprises: Obtaining experimental data and trial production data of propanthioazole; the experimental data and trial production data at least include the purity of the intermediate of each process and the data of each process parameter; determining the optimal experimental range of each process parameter of each process according to the experimental data; and conducting trial production after the experiment is completed and according to the optimal experimental range; Select any one of the process parameters of any of the steps as a target parameter, and other process parameters in the same step as other parameters; in the experimental data, obtain an experimental correlation coefficient between the target parameter and the corresponding intermediate according to the fluctuation correlation characteristics between the target parameter and the purity of the corresponding intermediate; In the trial production data, the trial production records are arbitrarily combined in a preset ratio to form a number of trial production record groups; in each of the trial production record groups, according to the violent fluctuation characteristics of the other parameters, the target record group of the target parameter is screened out; the process of the target parameter is used as the target process; in each of the trial production records, according to the fluctuation characteristics of the purity of the intermediate before the target process, combined with the purity of the intermediate of the target process, the corrected purity of the intermediate of the target process is obtained; in the target record group, according to the fluctuation correlation characteristics of the target parameter and the corrected purity of the corresponding intermediate, the trial production correlation coefficient of the target parameter and the corresponding intermediate is obtained; According to the relative characteristics of the experimental correlation coefficient and the trial production correlation coefficient, combined with the optimal experimental range of the target parameter, the optimal production range of the target parameter is obtained; The method for obtaining the experimental correlation coefficient includes: In the experimental data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence corresponding to the purity of the intermediate is used as the experimental correlation coefficient between the target parameter and the corresponding intermediate; The method for obtaining the trial production correlation coefficient includes: In the target record group of the trial production data, the absolute value of the Pearson correlation coefficient between the sequence of the target parameter and the sequence of the corrected purity of the corresponding intermediate is used as the trial production correlation coefficient between the target parameter and the corresponding intermediate; The method for obtaining the optimal production range includes: The ratio of the experimental correlation coefficient between the target parameter and the intermediate and the trial production correlation coefficient is used as the production application rate of the target parameter; Obtain the median and range radius of the optimal experimental range of the target parameter, and use the product of the range radius and the production application rate as the corrected range radius; obtain the optimal production range of the target parameter based on the corrected range radius and in combination with the median.

2. The multi-dimensional data processing method of a fully digitalized propanthioazole production line according to claim 1, characterized in that: The method for obtaining the target record group includes: In each of the trial production record groups, at least according to the variance of each of the other parameters in all the trial production records, obtain the fluctuation severity coefficient of each of the other parameters; at least according to the mean of the fluctuation severity coefficients of all the other parameters, obtain the disturbance coefficient of each of the trial production record groups; A target record group is obtained according to the disturbance coefficient.

3. The multi-dimensional data processing method of a fully digitalized propanthazole production line according to claim 2 is characterized in that: The method for obtaining a target record group according to the disturbance coefficient includes: The trial production record group with the smallest disturbance coefficient is selected as the target record group.

4. The multi-dimensional data processing method of a fully digitalized propanthioazole production line according to claim 1 is characterized in that: The method for obtaining the corrected purity of the intermediate comprises: In each of the trial production records, when the target process is the first process, the purity of the intermediate of the target process is directly used as the corrected purity of the intermediate; When the target process is not the first process, the purity of the intermediates of the target process and the preceding process sequentially preceding the target process is used to form a purity sequence; in the purity sequence of the target process, the sum of the differences in the purity of the intermediates of all adjacent processes is used as the process conduction coefficient of the target process; based on the purity of the intermediates of the target process and the process conduction coefficient, the corrected purity of the intermediates of the target process corresponding to the trial production record is obtained; Among the differences in the purity of the intermediates of the adjacent processes, the purity of the intermediate of the process with the smallest order is the subtrahend; and the process conductance coefficient is negatively correlated with the corrected purity of the intermediate.

5. The multi-dimensional data processing method of a fully digitalized propanthioazole production line according to claim 4 is characterized in that: The method for obtaining the corrected purity of the intermediate of the target process corresponding to the trial production record according to the purity of the intermediate of the target process and the process conductance coefficient comprises: The purity of the intermediate of the target process is taken as a first minuend, the product of the purity of the intermediate and the process conductance coefficient is taken as a second minuend, and the difference between the first minuend and the second minuend is taken as a corrected purity of the intermediate of the target process.

6. The multi-dimensional data processing method of a fully digitalized propanthioazole production line according to claim 1, characterized in that: After obtaining the optimal production range, the method further includes: A secondary trial production is carried out within the optimal production range of each process parameter in each of the steps, and a polynomial regression model is constructed according to each process parameter and the purity of the intermediate to obtain the optimal process parameter value of each process parameter in each of the steps.

7. The multi-dimensional data processing method for a fully digitalized propanthioazole production line according to claim 1, characterized in that: The process parameters include at least temperature, stirring speed, reaction time and pressure.

Images