CN118136139B - Solid-liquid phase balance production management method for doxifluridine based on predictive model - Google Patents
Solid-liquid phase balance production management method for doxifluridine based on predictive model Download PDFInfo
- Publication number
- CN118136139B CN118136139B CN202410533003.5A CN202410533003A CN118136139B CN 118136139 B CN118136139 B CN 118136139B CN 202410533003 A CN202410533003 A CN 202410533003A CN 118136139 B CN118136139 B CN 118136139B
- Authority
- CN
- China
- Prior art keywords
- data
- production process
- quality
- parameter
- parameter data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
-
- C—CHEMISTRY; METALLURGY
- C07—ORGANIC CHEMISTRY
- C07H—SUGARS; DERIVATIVES THEREOF; NUCLEOSIDES; NUCLEOTIDES; NUCLEIC ACIDS
- C07H1/00—Processes for the preparation of sugar derivatives
-
- C—CHEMISTRY; METALLURGY
- C07—ORGANIC CHEMISTRY
- C07H—SUGARS; DERIVATIVES THEREOF; NUCLEOSIDES; NUCLEOTIDES; NUCLEIC ACIDS
- C07H19/00—Compounds containing a hetero ring sharing one ring hetero atom with a saccharide radical; Nucleosides; Mononucleotides; Anhydro-derivatives thereof
- C07H19/02—Compounds containing a hetero ring sharing one ring hetero atom with a saccharide radical; Nucleosides; Mononucleotides; Anhydro-derivatives thereof sharing nitrogen
- C07H19/04—Heterocyclic radicals containing only nitrogen atoms as ring hetero atom
- C07H19/06—Pyrimidine radicals
- C07H19/067—Pyrimidine radicals with ribosyl as the saccharide radical
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/90—Programming languages; Computing architectures; Database systems; Data warehousing
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Organic Chemistry (AREA)
- Computing Systems (AREA)
- Crystallography & Structural Chemistry (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Databases & Information Systems (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Biochemistry (AREA)
- Biotechnology (AREA)
- Genetics & Genomics (AREA)
- Molecular Biology (AREA)
- Manufacturing & Machinery (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Analytical Chemistry (AREA)
Abstract
The invention discloses a solid-liquid phase balance production management method for doxifluridine based on a prediction model, belonging to the technical field of doxifluridine production management; according to the invention, the solid-liquid phase equilibrium production process of the doxifluridine is monitored, and the change condition of Apelblat model reaction parameter data in the production process is obtained in real time; dividing the reaction parameter data into normal data and abnormal data; comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine; according to the evaluation result, the reaction parameter data in the production process is divided into a high-quality data level, a general quality data level and a low-quality data level, and the high-quality data can be directly used for establishing and optimizing a Apelblat model so as to guide the adjustment of the real-time production process and ensure the stable improvement of the product quality and the production efficiency.
Description
Technical Field
The invention relates to the technical field of production management of deoxyfluorouridine, in particular to a solid-liquid phase balance production management method for deoxyfluorouridine based on a predictive model.
Background
The solid-liquid phase equilibrium production management for the doxifluridine based on the prediction model refers to predicting the law of the solubility of substances along with the change of temperature and components by using a Apelblat model in the solid-liquid phase equilibrium production process of the doxifluridine so as to guide the control and optimization of the production process. By the model, the conditions of the solid-liquid phase reaction can be adjusted to maximize the yield and quality of the product and ensure the stability and economy of the production process.
In actual operation, the production management based on the prediction model includes real-time monitoring and adjustment of various parameters in the production process to ensure that the reaction conditions are in an optimal state. Meanwhile, the quality of raw materials, the running state of equipment and other factors are required to be monitored, and the possible problems in the production process are found and solved in time, so that the production of the doxifluridine can be carried out in time, efficiently and stably.
However, when the prediction model is used for carrying out solid-liquid phase balance production management on the doxifluridine, the prediction result of the Apelblat model is highly dependent on the used parameter data, if the used data quality is uneven, the accuracy of model prediction can be reduced, the adjustment and optimization in the production process can not be accurately guided, and the production efficiency and the product quality are affected.
Disclosure of Invention
The invention aims to provide a solid-liquid phase balance production management method for doxifluridine based on a prediction model, which can guide the adjustment of a real-time production process and ensure the stable improvement of product quality and production efficiency.
In order to achieve the above object, the present invention provides the following technical solutions: a solid-liquid phase balance production management method for doxifluridine based on a prediction model comprises the following steps:
s1: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
S2: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
S3: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
S4: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
S5: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
The invention is mainly suitable for the solid-liquid phase balance production process of the doxifluridine, and can also be suitable for the production of other products, and the parameters in the invention can be properly adjusted.
In S2, the pre-processed reaction parameter data is matched with the history data by the manhattan distance, and the reaction parameter data is divided into normal data and abnormal data according to the matching result, which specifically includes:
acquiring preprocessed reaction parameter data, and calculating Manhattan distance between each preprocessed real-time data point and each data point in the historical data set;
The feature vector of the acquired real-time data point is marked as x, the feature vector of the historical data point is marked as y, and the specific calculation expression is as follows: ; where d (x, y) is the Manhattan distance between the real-time data point and the historical data point, n is the dimension of the feature vector, x i is the ith real-time data point, and y i is the ith historical data point;
setting a similarity threshold according to the distribution condition and the actual demand of the historical data;
Comparing the calculated Manhattan distance with a set similarity threshold for each real-time data point; if d (x, y) is less than the similarity threshold, then classifying the data point as normal data; if d (x, y) is greater than the similarity threshold, the data point is classified as anomalous.
In the preferred scheme of the invention, in S3, according to different stability degrees of the quality of the doxifluridine product in the production process, the reaction parameter precision abnormality indexes under corresponding conditions are respectively obtained, and the reaction parameter precision abnormality indexes are obtained by the following steps:
Modeling the relation between the reaction parameter a and the product quality B as a linear relation, modeling the parameters of a model, which is θ= (θ0, θ1,) and θn, where θ0 is the intercept, θ1,) and θn is the slope, and for each parameter θi, i=1, -), n, obtaining its normal distribution, i.e. Wherein mu i andIs the mean and variance of the parameters;
updating posterior distribution of parameters according to the observation data by using a Bayes theorem, calculating the width of confidence intervals of the parameters, namely calculating the abnormal index of accuracy of the reaction parameters, and for a linear regression model, assuming that error items ϵ of likelihood functions accord with normal distribution, wherein a specific calculation expression is as follows: ; nk r is the abnormal index of accuracy of the reaction parameters, B x is the model predictive value, and the specific calculation expression is: b x =θ0+θ1+θ2+a2+ & gt θn An; The variance of the error is the parameter is the error; n is the number of data points, B i is the i-th product quality value, A1, A2, an are the labels of the reaction parameters, respectively.
In S3, the change amplitude of each parameter data in the production process is continuously monitored, the mutual interference degree between different parameter data is judged, the parameter amplitude interference index between each parameter is obtained, and the parameter amplitude interference index is obtained by the following steps:
Marking the number of parameters in the production process as m, carrying out standardized treatment on each parameter in the production process, and specifically calculating the expression as follows: ; wherein Z ij represents normalized parameter data, X ij represents the value of the ith parameter at the jth sample point, X i is the mean value of the ith parameter, and k i is the standard deviation of the ith parameter;
a covariance matrix between the normalized parameter data is calculated, ; Where Cov is the covariance matrix,Representing the transpose of Z ij, n being the number of matrices; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues and corresponding eigenvectors, wherein cov=λ=v; wherein v is a eigenvector, representing the direction of the principal component, λ is an eigenvalue, representing the variance of the principal component;
and linearly combining the standardized parameter data with the selected main component to obtain the parameter loading, wherein the specific calculation expression is as follows: ; wherein W ij is the parameter loading amount, which represents the contribution degree of the parameter to the principal component, v ij represents the eigenvector of the ith parameter to the jth principal component, and the parameter amplitude interference index is calculated by the specific calculation expression: ; where EC g is the parameter amplitude interference index.
As a preferred scheme of the invention, in S4, the accuracy of the reaction parameter data and the consistency among the parameters in the production process are comprehensively analyzed, and the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is evaluated, specifically:
And carrying out normalization treatment on the response parameter precision abnormality index and the parameter amplitude interference index, and calculating a quality evaluation coefficient of response parameter data in the solid-liquid phase equilibrium production process of the doxifluridine according to the response parameter precision abnormality index and the parameter amplitude interference index after the normalization treatment.
As a preferred embodiment of the present invention, in S5, according to the evaluation result, the reaction parameter data in the production process is divided into a high quality data level, a general quality data level and a low quality data level, specifically:
Comparing the quality evaluation coefficient of the obtained reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with a gradient standard threshold, wherein the gradient standard threshold comprises a first standard threshold and a second standard threshold, the first standard threshold is smaller than the second standard threshold, and comparing the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with the first standard threshold and the second standard threshold respectively;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than a second standard threshold value, marking the reaction parameter data as high-quality data, and generating no abnormal signal at the moment;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than or equal to a first standard threshold value and smaller than or equal to a second standard threshold value, marking the reaction parameter data as general quality data, and generating a second-level abnormal signal at the moment;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is smaller than a first standard threshold value, marking the reaction parameter data as low-quality data, and generating a first-level abnormal signal at the moment.
The invention also provides a solid-liquid phase balance production management system for the doxifluridine based on the prediction model, which is used for realizing the management method, and comprises a data preprocessing module, a data matching module, a quality evaluation module, a comprehensive analysis module and a grading processing module;
And a data preprocessing module: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
And a data matching module: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
The quality evaluation module: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
And the comprehensive analysis module is used for: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
And the grading processing module is used for: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
In the technical scheme, the invention has the technical effects and advantages that:
1. The invention divides high, medium (general) and low quality data levels by comprehensively monitoring, preprocessing, matching and evaluating the reaction parameter data in the production process. The comprehensive analysis can effectively improve the quality and reliability of the data, and ensure that the used data has good support for the accuracy and predictability of Apelblat models.
2. By evaluating and dividing the data quality, the invention can process and utilize the response parameter data with different quality levels in a targeted manner. The high-quality data can be directly used for establishing and optimizing Apelblat models to guide the adjustment of the production process, so that the stability of the production process and the product quality are improved; for general and low-quality data, further cleaning, correction or elimination can be performed according to specific conditions so as to improve the quality of the data, thereby reducing the production efficiency reduction and the product quality fluctuation caused by the quality problem of the data.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a block diagram of a system according to the present invention.
Detailed Description
The technical solutions of the embodiments of the present invention are clearly and completely described below with reference to the accompanying drawings.
Example 1 as shown in fig. 1, the solid-liquid phase equilibrium production management method for doxifluridine based on the prediction model in this example includes the following steps:
s1: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
S2: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
S3: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
S4: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
S5: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
The reaction parameter data refers to various parameters related to the reaction, including but not limited to reaction temperature, pressure, material concentration, reaction rate, etc., recorded and monitored during the process of equilibrium production of the solid-liquid phase of the doxifluridine. These data reflect the state and performance changes of the reaction system during production.
In the S1, monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data, specifically;
in the solid-liquid phase equilibrium production management of doxifluridine based on a predictive model, the monitored reaction parameter data is preprocessed to improve the quality and usability of the data for subsequent data analysis and processing. The method specifically comprises the following steps:
the data is flushed to remove erroneous, outliers, or unreasonable data points. This involves detecting and deleting abnormal data introduced by sampling device faults, sensor drift, artifacts, or other noise sources.
Data smoothing, smoothing data to reduce fluctuations and noise in the data. Common smoothing techniques include moving average, weighted moving average, exponential smoothing, etc., which can make the data more stable and easier to identify trends and patterns.
Interpolation of data refers to estimating the value of missing data using existing data. During production, data loss due to sensor failure or communication problems may occur. Through interpolation technology, possible values of missing data can be deduced according to the trend and mode of the existing data, so that the missing part of the data is filled.
Data normalization converts data of different dimensions or units into a unified standard form for comparison and analysis. Normalization can eliminate dimensional effects between data so that different data are comparable.
The abnormal value processing may be performed according to the actual situation with respect to the detected abnormal value, for example, the abnormal value is removed, replaced with an adjacent value, or corrected.
Data sampling and resampling, where the sampling frequency is non-uniform or synchronization with other data is required, may be performed such that the data has a consistent time interval or is synchronized with other data.
S2: and constructing a data matching model, matching the preprocessed reaction parameter data with the historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result.
Wherein, the data matching model includes: the feature extractor is responsible for extracting useful features or feature vectors from the reaction parameter data, which features can reflect important information of the data. Features may include statistical properties of the data (e.g., mean, variance, kurtosis, skewness, etc.), frequency domain features, time domain features, etc.
And the similarity measure is used for measuring the similarity degree between the current real-time data and the historical data. Common similarity measurement methods include euclidean distance, manhattan distance, cosine similarity, and the like. These metrics can represent differences or similarities between the data.
And a historical data storage which stores reaction parameter data in the historical production process. The historical data may be stored in the form of a database, a data warehouse, or a simple data file.
And the anomaly detector is responsible for judging whether the current data is anomalous according to the similarity measurement result. If the similarity between the current data and the historical data is lower than a certain threshold value, the current data can be judged to be abnormal data. The anomaly detector may employ a threshold-based approach, a statistical-based approach, or a machine-learning based approach.
And a feedback regulator, in some cases, if the current data is judged to be abnormal data, the feedback regulator can take some automatic measures to regulate the production process so as to gradually restore the data to a normal state.
The preprocessed reaction parameter data is matched with the historical data through the Manhattan distance, and then the reaction parameter data is divided into normal data and abnormal data according to the matching result, wherein the reaction parameter data comprises the following specific steps: and acquiring preprocessed reaction parameter data, and calculating Manhattan distance between each preprocessed real-time data point and each data point in the historical data set. The feature vector of the acquired real-time data point is marked as x, the feature vector of the historical data point is marked as y, and the specific calculation expression is as follows: ; where d (x, y) is the Manhattan distance between the real-time data point and the historical data point, n is the dimension of the feature vector, x i is the ith real-time data point, and y i is the ith historical data point.
And setting a similarity threshold according to the distribution condition and the actual demand of the historical data, and judging the similarity degree between the data.
For each real-time data point, the calculated Manhattan distance is compared to a set similarity threshold. If d (x, y) is less than the similarity threshold, then classifying the data point as normal data; if d (x, y) is greater than the similarity threshold, the data point is classified as anomalous.
S3: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process.
According to different stability degrees of the deoxyfluorouridine product quality in the production process, respectively obtaining the response parameter precision abnormality indexes under corresponding conditions, wherein the response parameter precision abnormality indexes are obtained by the following steps:
and establishing a linear regression model, and modeling the relation between the reaction parameter A and the product quality B as a linear relation. Modeling parameters of the model, the parameters of the model being θ= (θ0, θ1,..and θn), where θ0 is the intercept, θ1,..and θn is the slope. For each parameter θi, its normal distribution is obtained, i.e Wherein mu i andIs the mean and variance of the parameters.
According to the observation data, updating posterior distribution of parameters by using Bayes theorem, calculating the width of confidence interval of the parameters, namely the range of the posterior distribution of the parameters under a certain confidence level, namely calculating the abnormal index of accuracy of the reaction parameters, wherein for a linear regression model, a likelihood function usually assumes that an error term ϵ accords with normal distribution, and a specific calculation expression is as follows: ; nk r is the abnormal index of accuracy of the reaction parameters, B x is the model predictive value, and the specific calculation expression is: b x =θ0+θ1+θ2+a2+ & gt θn An; The variance of the error is the parameter is the error; n is the number of data points, B i is the i-th product quality value, A1, A2, an are the labels of the reaction parameters, respectively.
When the accuracy abnormality index of the reaction parameter is larger, the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is poorer, and the abnormality or uncertainty is larger.
A high anomaly index indicates that the variability of the response parameter data is large and deviates far from the historical data or expected values, and thus the accuracy of the data may be compromised. This may be due to experimental error, equipment failure or improper operation.
An abnormal exponential increase in the reaction parameters may reflect instability of the production process, i.e. a large fluctuation or variation in the reaction conditions. This may lead to fluctuations in product quality or deviations from expectations, increasing the risk of the production process.
An increase in the index of reaction parameter accuracy anomaly may reduce the controllability and adjustability of the production process, making it more difficult to adjust the reaction conditions effectively in time to maintain the product quality. This may require more frequent monitoring and adjustment and more complex control strategies to be adopted.
Analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among the parameters in the production process, wherein the method specifically comprises the following steps:
the variation amplitude of the reaction parameter data is analyzed to know the fluctuation degree of each parameter in the production process. By monitoring and analyzing the variation amplitude of the parameters, fluctuation or abnormal conditions in the production process can be identified, and measures can be taken in time to adjust or correct so as to keep the stability of the production process.
The production process involves a number of parameters which may influence or depend on each other. Assessing consistency between parameters means analyzing whether there is a correlation or consistency between the parameters. If there is a correspondence between the parameters, i.e. their trend of variation or pattern of fluctuation is similar, it is stated that there is a certain correlation between these parameters in the production process. Conversely, if the variations between the parameters are inconsistent or contradictory, further adjustments to the production process or control parameters may be required to improve the stability and consistency of the production process.
Continuously monitoring the variation amplitude of each parameter data in the production process, judging the mutual interference degree between different parameter data, and acquiring the parameter amplitude interference index between each parameter, wherein the acquisition method of the parameter amplitude interference index comprises the following steps:
Marking the number of parameters in the production process as m, carrying out standardized treatment on each parameter in the production process, and specifically calculating the expression as follows: ; where Z ij represents normalized parameter data, X ij represents the value of the ith parameter at the jth sample point, X i is the mean value of the ith parameter, and k i is the standard deviation of the ith parameter.
A covariance matrix between the normalized parameter data is calculated,; Wherein Cov is a covariance matrix reflecting the linear relationship between parameters,Representing the transpose of Z ij, n being the number of matrices and n being the number of matrices; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues and corresponding eigenvectors, wherein cov=λ=v; where v is a eigenvector, representing the direction of the principal component, and λ is an eigenvalue, representing the variance of the principal component. The first few feature vectors having the largest feature value are used as the principal components, and for example, the first 3 feature vectors having the largest feature value may be used as the principal components.
And linearly combining the standardized parameter data with the selected main component to obtain the parameter loading, wherein the specific calculation expression is as follows: ; wherein W ij is the parameter loading amount, which represents the contribution degree of the parameter to the principal component, v ij represents the eigenvector of the ith parameter to the jth principal component, and the parameter amplitude interference index is calculated by the specific calculation expression: ; where EC g is the parameter amplitude interference index.
The larger the parameter amplitude disturbance index, the greater the degree of contribution of the parameter to the principal component, that is to say the more pronounced the effect of the parameter on the overall production process. In the management of the equilibrium production of doxifluridine in the solid-liquid phase, this may indicate a greater fluctuation or variation of this parameter during the reaction, with a higher degree of interference with other parameters. In other words, the quality of the reaction parameter data may be subject to large instabilities and fluctuations.
S4: and comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine.
And carrying out normalization treatment on the response parameter precision abnormality index and the parameter amplitude interference index, and calculating a quality evaluation coefficient of response parameter data in the solid-liquid phase equilibrium production process of the doxifluridine according to the response parameter precision abnormality index and the parameter amplitude interference index after the normalization treatment.
For example, the following formula can be used to calculate the quality evaluation coefficient of the reaction parameter data in the solid-liquid equilibrium production process of doxifluridine, and the specific calculation expression is: Wherein rh k is a quality evaluation coefficient, nk r is a reaction parameter accuracy anomaly index, EC g is a parameter amplitude interference index, a 1、a2 is a reaction parameter accuracy anomaly index, a proportionality coefficient of the parameter amplitude interference index, and a 2>a1 >0.
According to the calculation expression, the reaction parameter precision abnormality index, the parameter amplitude interference index and the quality evaluation coefficient are in inverse proportion, and the quality evaluation coefficient is reduced along with the increase of the reaction parameter precision abnormality index and the parameter amplitude interference index, which indicates that the quality of reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is reduced.
S5: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
Comparing the quality evaluation coefficient of the obtained reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with a gradient standard threshold, wherein the gradient standard threshold in the embodiment comprises a first standard threshold and a second standard threshold, the first standard threshold is smaller than the second standard threshold, and comparing the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with the first standard threshold and the second standard threshold respectively;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than a second standard threshold value, the higher the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is, the stronger the stability of a Apelblat model is, the reaction parameter data is marked as high-quality data, and no abnormal signal is generated at the moment;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than or equal to a first standard threshold value and smaller than or equal to a second standard threshold value, the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is general, the stability of a Apelblat model is general, the reaction parameter data is marked as general quality data, and a secondary abnormal signal is generated at the moment;
If the quality evaluation coefficient of the reaction parameter data in the process of the solid-liquid phase equilibrium production of the doxifluridine is smaller than a first standard threshold value, the lower the quality of the reaction parameter data in the process of the solid-liquid phase equilibrium production of the doxifluridine is, the lower the stability of a Apelblat model is, the reaction parameter data is marked as low-quality data, and a first-level abnormal signal is generated at the moment.
Wherein, the high quality data can be directly used for the establishment and optimization of Apelblat models to guide the adjustment of the real-time production process. General quality data can be used to aid in analysis and decision making and should be carefully used. Low quality data requires further cleaning, correction or rejection to improve data quality.
It should be noted that, the importance degree of the first-level early warning signal is higher than that of the second-level early warning signal, and a person skilled in the relevant art can judge the quality of the reaction parameter data according to the level of the early warning signal.
In the embodiment, the change condition of Apelblat model reaction parameter data in the production process is obtained in real time by monitoring the solid-liquid phase equilibrium production process of the doxifluridine; dividing the reaction parameter data into normal data and abnormal data; judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process; comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine; according to the evaluation result, the reaction parameter data in the production process is divided into a high-quality data level, a general quality data level and a low-quality data level, and the high-quality data can be directly used for establishing and optimizing the Apelblat model so as to guide the adjustment of the real-time production process. General quality data can be used to aid in analysis and decision making and should be carefully used. Low quality data requires further cleaning, correction or rejection to improve data quality. And the production process is optimized and regulated, so that the stable improvement of the product quality and the production efficiency is ensured.
Embodiment 2, as shown in fig. 2, the present embodiment provides a solid-liquid phase balance production management system for doxifluridine based on a prediction model, which is used for implementing the management method of embodiment 1, and includes a data preprocessing module, a data matching module, a quality evaluation module, a comprehensive analysis module and a hierarchical processing module;
And a data preprocessing module: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
And a data matching module: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
The quality evaluation module: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
And the comprehensive analysis module is used for: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
And the grading processing module is used for: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
The above formulas are all formulas with dimensions removed and numerical values calculated, the formulas are formulas with a large amount of data collected for software simulation to obtain the latest real situation, and preset parameters in the formulas are set by those skilled in the art according to the actual situation. The analysis and calculation referred to in the present invention can be performed by an electronic device including a memory, a processor, and a computer program stored on the memory and executable on the processor, and the algorithm described above can be implemented by the processor executing the program.
Claims (2)
1. A solid-liquid phase balance production management method for deoxyfluorouridine based on a prediction model is characterized by comprising the following steps of: comprises the following steps of;
s1: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
S2: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
acquiring preprocessed reaction parameter data, and calculating Manhattan distance between each preprocessed real-time data point and each data point in the historical data set;
The feature vector of the acquired real-time data point is marked as x, the feature vector of the historical data point is marked as y, and the specific calculation expression is as follows: d (x, y) = Σ i n ＝1|xi-yi |; where d (x, y) is the Manhattan distance between the real-time data point and the historical data point, n is the dimension of the feature vector, x i is the ith real-time data point, and y i is the ith historical data point;
setting a similarity threshold according to the distribution condition and the actual demand of the historical data;
Comparing the calculated Manhattan distance with a set similarity threshold for each real-time data point; if d (x, y) is less than the similarity threshold, then classifying the data point as normal data; if d (x, y) is greater than the similarity threshold, then classifying the data point as anomalous;
S3: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
According to different stability degrees of the deoxyfluorouridine product quality in the production process, respectively obtaining the response parameter precision abnormality indexes under corresponding conditions, wherein the response parameter precision abnormality indexes are obtained by the following steps:
Modeling the relation between the reaction parameter A and the product quality B as a linear relation, modeling the parameters of a model, wherein the parameters of the model are theta= (theta 0, theta 1,) and theta N, theta 0 is an intercept, theta 1, -), theta N is a slope, and obtaining normal distribution of each parameter theta i, namely theta i-N (mu i,σi 2), wherein mu i and sigma i 2 are the mean value and variance of the parameters;
Updating posterior distribution of parameters according to the observation data by using Bayes theorem, calculating the width of confidence intervals of the parameters, namely calculating the abnormal index of accuracy of the reaction parameters, and for a linear regression model, assuming that error items epsilon of likelihood functions accord with normal distribution, wherein a specific calculation expression is as follows: nk r is the abnormal index of accuracy of the reaction parameters, B x is the model predictive value, and the specific calculation expression is: b x＝θ0+θ1*A1+θ2*A2+...+θn*An;σi 2 is the variance of the error, N is the number of data points, B i is the i-th product quality value, A1, A2, an are the labels of the reaction parameters, respectively;
Continuously monitoring the variation amplitude of each parameter data in the production process, judging the mutual interference degree among different parameter data, and acquiring the parameter amplitude interference index among each parameter, wherein the acquiring method of the parameter amplitude interference index comprises the following steps:
Marking the number of parameters in the production process as m, carrying out standardized treatment on each parameter in the production process, and specifically calculating the expression as follows: Wherein Z ij represents normalized parameter data, X ij represents the value of the ith parameter at the jth sample point, X i is the mean value of the ith parameter, and k i is the standard deviation of the ith parameter;
a covariance matrix between the normalized parameter data is calculated, Wherein Cov is covariance matrix, Z ij T is transpose of Z ij, and n is matrix number; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues and corresponding eigenvectors, wherein cov=λ=v; wherein v is a eigenvector, representing the direction of the principal component, λ is an eigenvalue, representing the variance of the principal component;
and linearly combining the standardized parameter data with the selected main component to obtain the parameter loading, wherein the specific calculation expression is as follows: w ij＝vij*Zij; wherein W ij is the parameter loading amount, which represents the contribution degree of the parameter to the principal component, v ij represents the eigenvector of the ith parameter to the jth principal component, and the parameter amplitude interference index is calculated by the specific calculation expression: EC g＝∑Wij*Xij; wherein EC g is the parameter amplitude interference index;
S4: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
Normalizing the response parameter precision abnormality index and the parameter amplitude interference index, and calculating a quality evaluation coefficient of response parameter data in the solid-liquid phase balance production process of the doxifluridine according to the response parameter precision abnormality index and the parameter amplitude interference index after normalization;
s5: dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels;
Comparing the quality evaluation coefficient of the obtained reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with a gradient standard threshold, wherein the gradient standard threshold comprises a first standard threshold and a second standard threshold, the first standard threshold is smaller than the second standard threshold, and comparing the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine with the first standard threshold and the second standard threshold respectively;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than a second standard threshold value, marking the reaction parameter data as high-quality data, and generating no abnormal signal at the moment;
If the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is larger than or equal to a first standard threshold value and smaller than or equal to a second standard threshold value, marking the reaction parameter data as general quality data, and generating a second-level abnormal signal at the moment;
if the quality evaluation coefficient of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine is smaller than a first standard threshold value, marking the reaction parameter data as low-quality data, and generating a first-level abnormal signal at the moment.
2. The solid-liquid phase balance production management system for the doxifluridine based on the prediction model is used for realizing the solid-liquid phase balance production management method for the doxifluridine based on the prediction model, and is characterized in that: the system comprises a data preprocessing module, a data matching module, a quality evaluation module, a comprehensive analysis module and a grading processing module;
And a data preprocessing module: monitoring a solid-liquid phase equilibrium production process of the doxifluridine, acquiring the change condition of Apelblat model reaction parameter data in the production process in real time, and preprocessing the monitored reaction parameter data;
And a data matching module: constructing a data matching model, matching the preprocessed reaction parameter data with historical data, and dividing the reaction parameter data into normal data and abnormal data according to a matching result;
The quality evaluation module: when the reaction parameter data are normal data, judging the stability degree of the quality of the doxifluridine product in the production process, evaluating the accuracy degree of the reaction parameter data at the moment, analyzing the variation amplitude of the reaction parameter data, and evaluating the consistency among all the parameters in the production process;
And the comprehensive analysis module is used for: comprehensively analyzing the accuracy of the reaction parameter data and the consistency among the parameters in the production process, and evaluating the quality of the reaction parameter data in the solid-liquid phase equilibrium production process of the doxifluridine;
And the grading processing module is used for: and dividing the reaction parameter data in the production process into a high-quality data level, a general quality data level and a low-quality data level according to the evaluation result, and correspondingly processing and utilizing the reaction parameter data with different quality levels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410533003.5A CN118136139B (en) | 2024-04-30 | 2024-04-30 | Solid-liquid phase balance production management method for doxifluridine based on predictive model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410533003.5A CN118136139B (en) | 2024-04-30 | 2024-04-30 | Solid-liquid phase balance production management method for doxifluridine based on predictive model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118136139A CN118136139A (en) | 2024-06-04 |
| CN118136139B true CN118136139B (en) | 2024-07-19 |
Family
ID=91230385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410533003.5A Active CN118136139B (en) | 2024-04-30 | 2024-04-30 | Solid-liquid phase balance production management method for doxifluridine based on predictive model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118136139B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115860579A (en) * | 2023-02-27 | 2023-03-28 | 山东金利康面粉有限公司 | Production quality monitoring system for flour processing |
| CN117933808A (en) * | 2024-01-25 | 2024-04-26 | 广州博依特智能信息科技有限公司 | Production quality assessment method and system based on production process monitoring |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113255795B (en) * | 2021-06-02 | 2021-10-26 | 杭州安脉盛智能技术有限公司 | Equipment state monitoring method based on multi-index cluster analysis |
-
2024
- 2024-04-30 CN CN202410533003.5A patent/CN118136139B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115860579A (en) * | 2023-02-27 | 2023-03-28 | 山东金利康面粉有限公司 | Production quality monitoring system for flour processing |
| CN117933808A (en) * | 2024-01-25 | 2024-04-26 | 广州博依特智能信息科技有限公司 | Production quality assessment method and system based on production process monitoring |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118136139A (en) | 2024-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bakdi et al. | A new adaptive PCA based thresholding scheme for fault detection in complex systems | |
| US11264121B2 (en) | Real-time industrial plant production prediction and operation optimization | |
| CN105700518B (en) | A kind of industrial process method for diagnosing faults | |
| Zhao et al. | A sparse dissimilarity analysis algorithm for incipient fault isolation with no priori fault information | |
| Jiang et al. | Weighted kernel principal component analysis based on probability density estimation and moving window and its application in nonlinear chemical process monitoring | |
| CN120063575B (en) | High-precision adaptive pressure transmitter calibration method and system | |
| CN107831736B (en) | Model-plant mismatch detection using model parameter data clustering for paper machines or other systems | |
| CN119106259B (en) | Water and fertilizer consumption data collection and analysis system and method for melon and fruit planting | |
| Zhao et al. | Fine-scale modeling and monitoring of wide-range nonstationary batch processes with dynamic analytics | |
| CN120196048A (en) | Automatic detection and compensation method for abrasion of numerical control machine tool based on artificial intelligence | |
| CN118249810A (en) | Method and system for testing multi-channel AD/DA chips | |
| Tan et al. | Monitoring statistics and tuning of kernel principal component analysis with radial basis function kernels | |
| CN120178786A (en) | Real-time error control system for CNC machining of hardware parts | |
| CN118482817B (en) | Digital textile equipment management method and system based on data analysis | |
| CN120122053A (en) | A method, device, terminal equipment and computer-readable storage medium for detecting electric energy meter faults | |
| CN113505862B (en) | Hybrid fault detection method and device | |
| Casini et al. | Continuous Record Asymptotics for Change‐Point Models | |
| CN120739687A (en) | High-reliability multi-sensor fusion pump station monitoring system and intelligent early warning control method | |
| CN116401570B (en) | Intelligent processing system for printing quality monitoring big data | |
| CN118136139B (en) | Solid-liquid phase balance production management method for doxifluridine based on predictive model | |
| US11378063B2 (en) | System and method for detecting turbine underperformance and operation anomaly | |
| Zhao et al. | Sparse dissimilarity analysis based on distribution dissimilarity decomposition for online diagnosis of incipient faults | |
| CN120252854B (en) | A method, system, and storage medium for monitoring high-pressure gas relays. | |
| CN120663186A (en) | Piezoelectric quartz wafer grinding control method and system based on grinding frequency monitoring | |
| Tao et al. | A novel approach for moving window size selection utilizing recursive PCA |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |