Veterinary drug production quality control method

Abstract

The invention discloses a veterinary drug production quality control method, and belongs to the technical field of veterinary drug production, and the method comprises the steps: obtaining a veterinary drug production material quality characteristic value, a production condition characteristic value, production equipment state data and production operator behavior data; generating an original production quality comprehensive risk index according to the production material quality characteristic value, the production condition characteristic value and the production equipment state data; generating an operator efficiency regulation factor according to the production operator behavior data; generating a final production quality risk index according to the original comprehensive risk index and the operator efficiency regulation factor; and according to the final production quality risk index, the production of veterinary drugs is controlled. By integrating production materials, production conditions, equipment states and personnel behavior data, establishing the dynamic risk assessment model and generating the regulatory factors, real-time quality control based on multi-dimensional data is finally realized, the risk assessment comprehensiveness of veterinary drug production quality is improved, and dynamic quality control is realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of veterinary drug production, and particularly relates to a veterinary drug production quality control method. BACKGROUND

[0002] With the development of large-scale livestock farming and the increasing requirements for animal disease prevention and control, veterinary drugs, as an important material to protect animal health, directly relate to the safety of livestock products and public health safety. The current veterinary drug production process is becoming increasingly complex, involving raw material inspection, environmental control, equipment operation and personnel operation, and the traditional quality control method relying on manual sampling has been difficult to meet the fine supervision needs of modern veterinary drug production.

[0003] In the prior art, veterinary drug production quality control mainly adopts methods such as periodic sampling detection, production record auditing and static standard comparison. These methods include detecting the physical and chemical indexes of raw materials through laboratory, monitoring the production conditions such as temperature and humidity using environmental sensors, and maintaining the state through equipment operation log records. Some systems also establish a basic quality traceability system to track product flow through bar code or RFID technology. In terms of personnel management, most enterprises still adopt a regular training and examination system, and lack quantitative evaluation of the daily behavior efficiency of operating personnel.

[0004] However, the prior art has many limitations: first, various detection data are independent of each other, lacking fusion analysis of multi-dimensional data such as raw material quality, production conditions and equipment state, making it difficult to form a comprehensive risk assessment; second, the existing methods mostly ignore the dynamic influence of personnel operation on quality, and cannot quantitatively evaluate the behavior efficiency of operating personnel; third, the traditional risk prediction model is relatively simple, and a dynamic risk calculation mechanism considering personnel adjustment factors has not been established, resulting in insufficient accuracy and timeliness of quality early warning. SUMMARY

[0005] In view of the deficiencies of the prior art, the application provides a veterinary drug production quality control method, which solves the above problems.

[0006] To achieve the above purpose, the application is implemented by the following technical scheme: a veterinary drug production quality control method, comprising: obtaining production material quality characteristic values, production condition characteristic values, production equipment state data and production operator behavior data of the veterinary drug; establishing an original risk prediction model according to the production material quality characteristic values, the production condition characteristic values and the production equipment state data, and generating an original production quality comprehensive risk index; According to the production operator behavior data, an operator performance analysis model is established, and an operator performance adjustment factor is generated; wherein the production operator behavior data includes an operation personnel inspection task on-time completion rate, an operation personnel abnormal record reporting rate of production equipment and an operation personnel operation error rate; According to the original comprehensive risk index and the operator performance adjustment factor, a final production quality risk index is generated; According to the final production quality risk index, the production of the veterinary drug is controlled.

[0007] On the basis of the above technical scheme, the present application further provides the following optional technical schemes: Further technical scheme: the generation mode of the original production quality comprehensive risk index specifically includes: According to the production material quality characteristic value, a production material quality risk index is generated; According to the production condition characteristic value, a production condition risk index is generated; According to the production equipment state data, a production equipment risk index is generated; wherein the production equipment state data includes a historical key process compliance rate of the production equipment and a recent failure rate of the production equipment; According to the production material quality risk index, the production condition risk index and the production equipment risk index, an original risk prediction model is established, and an original production quality comprehensive risk index is generated.

[0008] Further technical scheme: the generation mode of the production material quality risk index specifically includes: Through the formula:

[0009] The production material quality risk index is generated ; In the formula, represents the i-th normalized production material quality characteristic value, represents the weight coefficient of the i-th material quality characteristic, and p represents the number of production material quality characteristics.

[0010] Further technical scheme: the generation mode of the production condition risk index specifically includes: Through the formula:

[0011] The production condition risk index is generated ; In the formula, represents the j-th production condition characteristic value, represents the weight coefficient of the j-th production condition characteristic, and q represents the number of production condition characteristics.

[0012] Further technical solutions: the generation method of the production equipment risk index specifically includes: Through the formula:

[0013] generate a production equipment risk index ; In the formula, represents the historical key process compliance rate of the production equipment, represents the minimum value of the standard key process compliance rate, represents the recent failure rate of the production equipment, represents the maximum failure rate of the standard production equipment, represents the weight coefficient of the key process compliance rate.

[0014] Further technical solutions: the expression of the original risk prediction model is specifically:

[0015] In the expression, R represents the original production quality comprehensive risk index, represents the production material quality risk index, represents the production condition risk index, represents the production equipment risk index, , , are weight coefficients, and .

[0016] Further technical solutions: the generation method of the operator efficiency adjustment factor specifically includes: According to the operator behavior data, an operator efficiency score is generated; According to the operator efficiency score, an operator efficiency analysis model is established to generate an operator efficiency adjustment factor.

[0017] Further technical solutions: the generation method of the operator efficiency score specifically includes: Through the formula:

[0018] generate an operator efficiency score ; In the formula, represents the normalized value of the historical inspection task on-time completion rate of the operator, represents a normalized value of the historical reporting rate of the operator to the abnormal record of the production equipment, represents a normalized value of the historical operation failure rate of the operator, 、 、 are weight coefficients, and .

[0019] A further technical solution: the expression of the operator performance analysis model is specifically:

[0020] In the formula, represents an operator performance adjustment factor, represents an operator performance score, represents a benchmark value of the operator performance score, and k represents an adjustment intensity coefficient.

[0021] A further technical solution: the generation mode of the final production quality risk index is specifically: The final production quality risk index is generated through the formula:

[0022] The final production quality risk index ; In the formula, represents an original production quality comprehensive risk index, represents an operator performance adjustment factor.

[0023] The present application provides a veterinary drug production quality control method, which has the following beneficial effects compared with the prior art: The present application integrates production materials, production conditions, equipment states and personnel behavior data, establishes a dynamic risk assessment model and generates an adjustment factor, and finally realizes real-time quality control based on multi-dimensional data, improves the comprehensiveness of risk assessment of veterinary drug production quality, and realizes dynamic quality control. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 A flowchart of a veterinary drug production quality control method provided by the present application. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0026] The specific implementation of the present application will be described in detail below in combination with specific examples.

[0027] Referring to Figure 1 A veterinary drug production quality control method is provided for an embodiment of the present application, comprising the following steps: Step S10: Obtain the production material quality characteristic value, production condition characteristic value, production equipment state data and production operator behavior data of the veterinary drug; It should be explained that the characteristic value is obtained by normalizing the data; in addition, the characteristic value is in a direct proportional relationship with its corresponding effect, for example, the larger the production material quality characteristic value, the better the production material quality; Step S20: Establish an original risk prediction model according to the production material quality characteristic value, production condition characteristic value and production equipment state data, and generate an original production quality comprehensive risk index; Step S30: Establish an operator performance analysis model according to the production operator behavior data, and generate an operator performance adjustment factor; wherein the production operator behavior data includes the operator inspection task completion rate, the operator abnormal record reporting rate of the production equipment and the operator operation error rate; Step S40: Generate a final production quality risk index according to the original comprehensive risk index and the operator performance adjustment factor; Step S50: Control the production of the veterinary drug according to the final production quality risk index; The production material quality characteristic value refers to the normalized raw material quality quantitative index, which can be specifically realized by standardizing the calculation of physicochemical detection data to eliminate the dimensional differences of different detection items. The production condition characteristic value refers to the standardized value of the environmental parameter, which can be specifically obtained by linear transformation of the temperature and humidity sensor data to quantify the production environment state; The production equipment state characteristic value refers to the normalized value of the equipment operation index, which can be specifically obtained by extracting key parameters from device log data to evaluate the equipment health; The operator behavior data refers to the digitized information of personnel operation records, which can be specifically structured by inspection records and abnormal reporting logs to quantify personnel operation performance; The original risk prediction model refers to a weighted calculation model of multi-dimensional risk index, which can be specifically fused with material, environmental and equipment risk sub-items by linear weighting algorithm to generate a basic risk assessment value; The performance adjustment factor refers to a dynamic coefficient reflecting the influence of personnel operation level on risk, which can be specifically calculated by the difference between the personnel performance score and the reference value to correct the basic risk value.

[0028] Specifically, the production material quality characteristic value is weighted to generate a material risk index, the production condition characteristic value is converted into an environmental risk index, and the equipment state characteristic value is calculated to obtain an equipment risk index, and the three are weighted and summed to form an original risk index. The operator behavior data is processed through an efficiency score model to generate an adjustment factor, which is multiplied by the original risk index to limit the upper limit value and generate a final risk index. When the final risk index exceeds a set threshold, a warning mechanism is triggered and the production parameters are adjusted.

[0029] Compared with the prior art, the method breaks through the single-dimensional data evaluation mode, realizes the fusion analysis of material, environment and equipment data, and constructs a multi-dimensional risk calculation framework. Unlike traditional static models, a dynamic personnel efficiency adjustment mechanism is introduced, so that the risk assessment result reflects the change of operation level in real time. Compared with the warning system relying only on equipment logs, the application quantifies the influence of human factors through personnel behavior data analysis, and improves the sensitivity of risk warning.

[0030] Through the above technical solutions, the application effectively solves the evaluation deviation problem caused by isolated analysis of multi-source data, realizes dynamic tracking and accurate early warning of production risk. The risk value is dynamically corrected by the personnel efficiency factor, which reduces the probability of quality accidents caused by human operation errors. The final risk index provides a quantitative basis for production control, supports timely implementation of risk intervention measures, such as automatically starting a backup stirring device when the risk of the mixing process increases.

[0031] Preferably, the application further proposes that the generation mode of the original production quality comprehensive risk index specifically comprises: Step S21: generating a production material quality risk index according to the production material quality characteristic value; Step S22: generating a production condition risk index according to the production condition characteristic value; Step S23: generating a production equipment risk index according to the production equipment state data; wherein the production equipment state data comprises a historical key process compliance rate of the production equipment and a recent failure rate of the production equipment; Step S24: establishing an original risk prediction model according to the production material quality risk index, the production condition risk index and the production equipment risk index, and generating an original production quality comprehensive risk index; The production material quality risk index refers to a quantitative index obtained by weighted calculation of the normalized material quality characteristic value, which can be realized by using weighted average method combined with a preset weight coefficient, and is used to reflect the contribution degree of raw material quality to the overall risk. The production condition risk index refers to an evaluation value generated by linear combination based on the environmental parameter characteristic value, which can be realized by using the combination of normalized data and dynamic weight distribution, and is used to quantify the influence of production environment fluctuation on quality stability. The production equipment risk index refers to a composite index generated by comprehensively integrating historical operation data and real-time state parameters of the equipment, and can be specifically realized by weighted calculation of the key process compliance rate and the recent failure rate, and is used to represent the potential threat of equipment reliability to the production quality. The original risk prediction model refers to a mathematical model that linearly superimposes the three types of risk indexes, and can specifically allocate the influence proportion of different risk dimensions by using preset weight coefficients, and is used to realize the collaborative analysis and comprehensive risk assessment of multi-source data.

[0032] Specifically, the production material quality risk index is calculated by summing the product of the normalized characteristic value and the weight coefficient, which can quantify the contribution difference of different material quality parameters to the overall risk. The production condition risk index is dynamically evaluated by using similar methods for environmental parameters such as temperature, humidity, and cleanliness, to capture the impact trend of production condition fluctuations on quality. The production equipment risk index is calculated by combining the historical compliance rate and the recent failure rate, taking into account the long-term performance and short-term operating state of the equipment. After generating the three types of risk indexes, they are linearly superimposed by using preset weight coefficients to form an original production quality comprehensive risk index that comprehensively reflects the multi-dimensional risk of materials, environment, and equipment. The step-by-step generation of the risk indexes of each dimension can avoid data coupling interference, and the model fusion can enhance the comprehensiveness of risk assessment.

[0033] Compared with the prior art, the traditional method only analyzes single-dimensional data in isolation, such as separately detecting the physical and chemical indicators of raw materials or monitoring the operating state of the equipment, and lacks a collaborative evaluation mechanism for multi-dimensional data. The present application calculates the material, condition, and equipment risk indexes in layers, and establishes a weighted fusion model, breaking through the limitations of single-dimensional analysis and enabling the identification of composite risks resulting from the interaction of multiple factors. The prior art does not consider the correlation between the historical performance and the recent state of the equipment, and the present application realizes multi-time dimension evaluation of equipment risk by combining the key process compliance rate and the failure rate.

[0034] Through the above technical solutions, the present application can effectively integrate the three types of core data of production material quality, environmental conditions, and equipment state, and establish a multi-dimensional risk assessment system. By calculating the risk indexes in different dimensions and fusing them into a comprehensive index, the superposition effect of risk sources such as raw material quality fluctuations, environmental parameter abnormalities, and equipment performance degradation can be accurately identified, avoiding potential risk omissions caused by traditional single-dimensional analysis. The weighted fusion model can adjust the weight distribution of different risk dimensions according to actual production needs, enhancing the adaptability and accuracy of risk assessment and providing comprehensive and reliable data support for the quality control of veterinary drugs.

[0035] Preferably, the present application further proposes that the generation method of the production material quality risk index specifically includes: Through the formula:

[0036] Generating production material quality risk index ; In the formula, represents the i-th normalized production material quality characteristic value, represents the weight coefficient of the i-th material quality characteristic, and p represents the number of production material quality characteristics; wherein the production material quality characteristic value refers to a value mapped to the interval [0, 1] by linear transformation of original data of different dimensions, which can be realized by range method or standard deviation method, for eliminating the influence of dimension difference of different detection indexes on the evaluation result; the weight coefficient refers to a parameter reflecting the contribution degree of each material quality characteristic to the overall risk, which can be determined by analytic hierarchy process or entropy weight method, for reflecting the importance difference of different quality characteristics; The number p of production material quality characteristics refers to the total number of raw material detection indexes participating in the evaluation, which can be the number of detection items such as raw material effective component content, impurity content, and microbial index.

[0037] Specifically, by normalizing a plurality of production material quality characteristic values, different detection indexes have comparability. For example, the raw material effective component content can be expressed in percentage, while the microbial index can be expressed in colony number, and after normalization, they are converted into values in the range of 0-1. Then the weighted sum of the normalized characteristic values is calculated according to the preset weight coefficients, wherein the sum of the weight coefficients is 1, to ensure that the contribution proportion of each characteristic to the risk is reasonable. By weighted sum and then average processing, the calculation result fluctuation caused by the difference in the number of characteristics is avoided. Finally, the risk index is generated by subtracting the average value from 1. When the material quality characteristic value is larger, the weighted average value is larger, and the risk index is smaller, i.e. the larger the characteristic value is, the better the material quality is, and the smaller the risk index is.

[0038] Compared with the prior art, the traditional method usually detects material physicochemical indexes separately without multi-dimensional integration, such as only judging whether a single index meets the standard without comprehensive evaluation. The present application solves the data island problem by establishing a mathematical formula to convert discrete detection indexes into a unified risk index. The prior art uses a simple pass / fail binary judgment method, which cannot quantify the risk degree, while the present application realizes dynamic quantitative evaluation of the risk degree through weight coefficient and normalization processing. For example, when the effective component content of a batch of raw materials meets the standard but is close to the lower limit, the traditional method still determines it as qualified, while the present application can accurately reflect its potential risk through normalization value reduction and weight calculation.

[0039] Through the technical scheme, the application realizes dynamic integration and quantitative evaluation of multi-dimensional production material quality characteristics, and solves the problem that the traditional method cannot comprehensively judge material quality risks. The dimension difference is eliminated through normalization processing, and the comparability of different detection indexes is ensured. The feature importance difference is reflected through the weight coefficient distribution, and subjective judgment deviation is avoided. The discrete data is converted into a unified risk index through a mathematical formula, and accurate quantitative basis is provided for subsequent quality control. For example, when the microbial index of a batch of raw materials is abnormal but the effective component meets the standard, the system can automatically identify the risk level through weighted calculation, which significantly improves the evaluation efficiency and accuracy compared with the artificial sampling inspection method.

[0040] Preferably, the application further proposes that the generation mode of the production condition risk index specifically includes: Through the formula:

[0041] generate a production condition risk index ; In the formula, represents the jth production condition characteristic value, represents the weight coefficient of the jth production condition characteristic, and q represents the number of production condition characteristics; wherein the production condition characteristic value refers to the normalized environment monitoring parameter, which can specifically adopt the workshop environment data collected by a temperature and humidity sensor in real time, and is converted into a value in the interval [0, 1] through the range method, so as to eliminate the influence of different dimensions on the evaluation result; the weight coefficient refers to a proportional parameter reflecting the influence degree of each production condition on the quality risk, which can specifically adopt the analytic hierarchy process to determine the priority of different environmental factors, for example, setting the temperature weight as 0.4, the humidity weight as 0.3, and the air cleanliness weight as 0.3, and reflecting the role of key control points through weighted calculation; Normalization processing refers to a method of converting original monitoring data into comparable values, which can specifically adopt the maximum and minimum value standardization formula to ensure that the characteristic values of different dimensions have the same measurement scale.

[0042] Specifically, first, the environmental monitoring data of the production workshop are collected, such as temperature, humidity, air cleanliness and the like, and original values of q dimensions are obtained. The dimension data are converted into characteristic values in the interval [0, 1] through a normalization formula, wherein the greater the value is, the more the production condition meets the standard requirement. Then, the q characteristic values are weighted and summed according to preset weight coefficients, and the closer the result is to 1, the better the overall production condition is. Finally, the result is converted into a production condition risk index in the manner of 1-, and when the weighted average value reaches 1, the risk index decreases to 0, and when the result completely deviates from the standard, the risk index increases to 1, thereby realizing linear mapping of the production condition and the risk degree.

[0043] Compared with the prior art, the traditional method only judges whether a single environmental parameter is qualified through threshold alarm, while the present application can identify a complex environmental anomaly by weighting and calculating multiple environmental factors such as temperature and humidity according to the influence degree through multi-dimensional data fusion. For example, when the temperature is normal but the humidity and cleanliness are simultaneously over-standard, the traditional method cannot evaluate the overall risk, while the present application can obtain a risk index higher than the threshold value through weighted calculation, thereby realizing dynamic evaluation of the overall production condition.

[0044] Through the above technical solution, the present application can accurately quantify the comprehensive influence degree of different production conditions on the quality of veterinary drugs. For example, in a high-temperature and high-humidity environment in summer, even if a single temperature parameter is not over-standard, but the humidity is continuously high, the risk index after weighted calculation will exceed the threshold value, thereby triggering quality control measures. The present application solves the problem of one-sidedness in the environmental risk judgment of the traditional method, and dynamically adjusts the importance of each production condition through weight coefficients, such as increasing the weight configuration of temperature in the freeze-drying process stage, thereby realizing accurate risk evaluation for different production links.

[0045] Preferably, the present application further proposes that the generation mode of the production equipment risk index comprises: The production equipment risk index is generated through the formula:

[0046] The production equipment risk index is generated through the formula: In the formula, represents the historical key process compliance rate of the production equipment, represents the minimum value of the standard key process compliance rate, represents the recent failure rate of the production equipment, represents the maximum failure rate of the standard production equipment, represents the weight coefficient of the key process compliance rate; wherein the historical key process compliance rate is calculated through the formula: ​is the proportion of the device completing the key process requirements within a certain period of time, which can be calculated by comparing the device sensor data with the process standard, and is used to reflect the stability of the long-term operation of the device; standard minimum key process compliance rate is the minimum qualified threshold specified by industry standards or enterprise internal standards, which may be 95%, for example, to determine whether the device performance deviates from the benchmark; recent failure rate is the frequency of device failure within a predetermined period, for example, by counting the number of failures in the last 30 days and the running time, to capture the immediate state change of the device; standard maximum failure rate is the highest failure threshold allowed by the device, for example, no more than 0.5% per month, as a basis for determining whether the device is in an abnormal state; weighting coefficient is used to adjust the proportion of historical compliance rate and recent failure rate in risk calculation, for example, taking 0.6 means paying more attention to the long-term performance of the device.

[0047] Specifically, when the historical key process compliance rate of the device is lower than the standard value, a positive value will be generated, and the max(0,·) function is used to filter the situation where the compliance rate is higher than the standard value, and only the device with performance decline is accumulated. At the same time, when the recent failure rate exceeds the standard maximum failure rate, the term will amplify the impact of failure anomaly. Through the and the weighted combination of 1- , not only the reference value of the historical operation data of the device is retained, but also the early warning effect of the recent failure state is strengthened. For example, for the sterilization kettle device, if its historical compliance rate is 92% (lower than the standard key process compliance rate minimum value 95%), the recent failure rate is 0.6% (exceeds the standard production device maximum failure rate 0.5%), =0.7, the risk calculation will focus on the long-term performance degradation problem, while taking into account the current failure frequency state.

[0048] Compared with the prior art, the traditional method usually only uses device failure rate or single-dimensional running data for evaluation, which cannot distinguish the influence difference between historical performance degradation and immediate failure state. The prior art does not set a double judgment mechanism for compliance rate deviation and failure rate exceeding the standard, which may misjudge the small fluctuation of the device under normal working condition as a risk, while the real risk signal is covered by noise. In addition, the existing evaluation model lacks a weight adjustment mechanism, which is difficult to adapt to the sensitivity difference of different device types to historical data and real-time state.

[0049] By the technical scheme, dynamic quantitative evaluation of equipment risk is realized, and influences of long-term performance degradation and short-term fault anomaly of equipment on production quality are effectively distinguished. The double threshold judgment mechanism avoids interference of normal working condition data, and ensures that the risk index only reflects a real abnormal state. The introduction of the weight coefficient enables the evaluation model to flexibly adapt to running characteristics of different equipment, for example, historical compliance rate monitoring is emphasized for precise filling equipment, and recent fault rate monitoring is emphasized for high-pressure reaction kettle. The scheme solves technical defects of single evaluation dimension of equipment state and split of historical data and real-time state in traditional methods, and provides accurate quantitative basis for risk early warning of veterinary drug production equipment.

[0050] Preferably, the application further provides an expression of the original risk prediction model, which is specifically:

[0051] In the expression, R represents an original production quality comprehensive risk index, represents a production material quality risk index, represents a production condition risk index, represents a production equipment risk index, , , are weight coefficients, and ; wherein the weight coefficients , , are parameters for dynamically adjusting contribution proportions of the three types of risk indexes of material quality, production condition and equipment state in comprehensive evaluation, and can be determined by an expert experience method or an analytic hierarchy process, for example, different weights are allocated according to importance differences of production stages; the production material quality risk index is a quantitative index reflecting deviation of raw material quality from a standard, and can be realized by weighted deviation calculation of normalized characteristic values, for example, when the material quality characteristic value is lower than a preset threshold value, the risk index is increased; the production condition risk index is an evaluation result representing deviation of environmental parameters from a target range, and can be realized by weighted calculation of normalized production condition characteristic values, for example, when temperature and humidity fluctuation exceeds an allowed range, the risk index is increased; the production equipment risk index is a composite index of historical compliance rate and recent fault rate of equipment, and can be realized by weighted combination of key process compliance rate and fault rate, for example, when the recent fault rate of equipment exceeds a standard value, the risk index is significantly increased.

[0052] Specifically, the three types of risk indexes are fused by a linear weighting model, and the weight coefficients are dynamically allocated according to actual production needs. For example, in a production batch with large fluctuations in raw material quality, The value can be set to a higher value to enhance the impact of material quality risk; in the scenario of equipment aging, The value can be increased to reflect the importance of the equipment state. Each risk index is normalized to ensure dimensional consistency and comparability. By constraining the sum of the weight coefficients to be 1, the influence of a single factor on the comprehensive evaluation result is avoided, while the adjustment flexibility is retained.

[0053] Compared with the prior art, the traditional method usually analyzes material, condition or equipment data separately, lacking a multi-dimensional collaborative evaluation mechanism. For example, the prior art can only determine whether a single indicator exceeds the threshold by threshold judgment, but cannot quantify the superposition effect of different risk sources. The present application converts isolated risk indicators into a unified comprehensive index by establishing a weighted fusion model, and allows dynamic adjustment of the weight according to the production scenario, thereby more comprehensively reflecting potential risks.

[0054] Through the above technical solution, the present application realizes multi-dimensional data fusion analysis of material quality, production conditions and equipment state, solving the one-sidedness problem of evaluation caused by data isolation in traditional methods. Through the dynamic weight distribution mechanism, the risk characteristics of different production scenarios can be adapted, improving the accuracy and comprehensiveness of risk warning, and providing a reliable foundation for subsequent dynamic risk calculation combined with personnel performance adjustment factors.

[0055] Preferably, the present application further proposes that the generation method of the operator performance adjustment factor specifically comprises: Step S31: generating an operator performance score according to operator behavior data; Step S32: establishing an operator performance analysis model according to the operator performance score to generate an operator performance adjustment factor; The operator behavior data refers to quantifiable behavior indicators directly related to the production process, which can be realized by using the normalized values of historical inspection task completion rate, abnormal reporting rate and operation error rate to objectively reflect the execution efficiency and standardization of the operator; The operator performance score refers to a weighted comprehensive evaluation of multi-dimensional behavior data, which can be realized by using a linear weighting model to calculate the total score of the inspection completion rate, abnormal reporting rate and operation error rate according to the preset weight coefficients, for converting discrete behavior indicators into a unified performance measure; The operator performance analysis model refers to a mathematical relationship for converting the performance score into a risk adjustment parameter, which can be realized by using a linear function of benchmark value comparison and intensity coefficient adjustment to establish a dynamic correlation mechanism between personnel performance and quality risk.

[0056] Specifically, first, the normalized values of the on-time completion rate, the abnormal report rate and the error rate of the operator are calculated by collecting the completion of the inspection task, the frequency of abnormal event reporting and the number of operation errors of the operator in the historical production cycle. Then, the three indicators are linearly combined according to the preset weight to generate a score value reflecting the comprehensive performance. Further, the score is compared with the preset reference value, and the performance adjustment factor is calculated by adjusting the intensity coefficient, so that when the actual performance of the operator is lower than the reference, the adjustment factor will amplify the original risk index, and vice versa. Thus, the dynamic conversion of personnel behavior data to risk parameters is realized, and the operator level can affect the quality risk assessment results in real time.

[0057] Compared with the prior art, the traditional method only qualitatively assesses the ability of the personnel by periodic training examination, lacks continuous quantitative monitoring of daily operation behavior, and cannot capture the influence of dynamic fluctuations of personnel performance on quality risk. The present application converts key factors such as operator operation specification, response timeliness, etc. into calculable adjustment parameters by constructing a scoring system based on multi-dimensional behavior indicators, so that the quality risk assessment system can reflect the changes in operator operation state in real time, and make up for the shortcomings of the traditional static personnel management mode.

[0058] Through the above technical solution, the present application realizes dynamic quantitative evaluation of the daily behavior performance of the operator on the quality risk, solves the problem that the personnel factor is difficult to be integrated into the automatic quality control system in the prior art, and makes the production quality risk index able to be adjusted in real time according to the operator level, significantly improving the accuracy and timeliness of risk early warning.

[0059] Preferably, the present application further proposes that the generation method of the operator performance score specifically includes: The operator performance score is generated by the formula:

[0060] The operator performance score is generated by the formula: ; In the formula, represents the normalized value of the on-time completion rate of the historical inspection task of the operator, represents the normalized value of the historical reporting rate of the operator to the abnormal record of the production equipment, represents the normalized value of the historical operation error rate of the operator, , , are weight coefficients, and ; Among them, the on-time completion rate of the historical inspection task of the operator is The ratio of the operation personnel to complete the equipment inspection task on time within a preset time period, which can be realized by matching the completion time and the planned time of the inspection task recorded by the statistical system, and is used to reflect the reliability of the operation personnel in performing the task. Historical reporting rate of operation personnel on production equipment abnormality records The ratio of the operation personnel to actively report abnormal events of the production equipment in the production process, which can be realized by analyzing the ratio of the reporting source of abnormal events in the production log to the total number of abnormal events, and is used to represent the initiative of the operation personnel in finding quality risks. Historical operation failure rate of operation personnel The frequency of the operation personnel to make wrong operations when performing the key process, which can be realized by comparing the deviation times of the standard operation process and the actual operation record, and is used to evaluate the operation standardization. Weight coefficient , , The contribution ratio of each behavior index in the comprehensive score, which can be dynamically adjusted by expert scoring method or analytic hierarchy process to adapt to the requirement difference of personnel ability in different production stages.

[0061] Specifically, the operation personnel performance scoring model converts discrete operation records into continuous quantitative scores by fusing three types of behavior data, including inspection completion rate, abnormality reporting rate and operation failure rate. Normalization processing makes the behavior data of different dimensions comparable, for example, the inspection completion rate can be converted into a value in the interval [0, 1] by dividing the total amount of preset tasks. In the weighted linear combination method, the dynamic configuration of the weight coefficient allows the focus of each index to be adjusted according to the production scene, for example, the weight of the abnormality reporting rate can be increased in the equipment debugging stage, and the operation standardization can be focused on in the stable production stage. The scoring model integrates multi-dimensional behavior characteristics into a single performance index through mathematical operation, providing a data basis for the dynamic correction of the subsequent risk index.

[0062] Compared with the prior art, the existing veterinary drug production control system usually only records the basic attendance and training results of the operation personnel, and lacks quantitative analysis of dynamic behavior data such as daily inspection, abnormality reporting and operation standardization. The traditional method relies on artificial subjective evaluation or static examination index, and cannot reflect the actual influence of personnel operation on quality risk in real time. The present application establishes a mathematical correlation between personnel performance and quality risk by constructing a scoring model based on multi-dimensional behavior data, solving the problem that personnel factors are difficult to quantify and dynamically evaluate in the prior art.

[0063] By the technical scheme, the continuous quantitative evaluation of the daily behavior efficiency of the operator is realized, so that the operator operation factor can dynamically affect the calculation of the quality risk index. The data dimension difference is eliminated through the normalization processing, and the adaptability of the scoring model to different production environments is enhanced by combining the adjustable weight coefficient. The scheme provides data support for accurately identifying the quality hidden danger caused by non-standard operation of the operator, thereby improving the timeliness of the veterinary drug production quality risk early warning and the pertinence of the control measures.

[0064] Preferably, the application further provides an expression of the operator efficiency analysis model, which is specifically:

[0065] In the formula, represents an operator efficiency adjustment factor, represents an operator efficiency score, represents a benchmark value of the operator efficiency score, and k represents an adjustment intensity coefficient. The operator efficiency score is a comprehensive score calculated based on normalized values of the historical inspection task completion rate, the abnormal report rate, and the operation error rate, and can be specifically realized by using a weighted summation method to reflect the influence degree of each behavior index on the efficiency. The benchmark value of the operator efficiency score is a reference threshold for determining whether the operator efficiency meets the standard, and can be specifically realized by using an industry standard value or a statistical average value of historical efficiency data of the enterprise to quantitatively evaluate the difference between the operator efficiency level and the expected standard. The adjustment intensity coefficient k is a parameter for controlling the correction amplitude of the risk index caused by the efficiency difference, and can be specifically realized by using a preset fixed value or a variable adjusted dynamically according to the risk level to balance the sensitivity and stability of the risk prediction.

[0066] Specifically, the calculation process of the operator efficiency adjustment factor is as follows: when the operator efficiency score is lower than the benchmark value of the operator efficiency score , is a positive value, the operator efficiency adjustment factor is greater than 1, so that the final production quality risk index changes with the operator efficiency adjustment factor ; when the operator efficiency score is higher than the benchmark value of the operator efficiency score , is a negative value, will be equal to 1, so that the final production quality risk index does not change with the operator efficiency adjustment factor The change is generated. The adjustment intensity coefficient k controls the amplification multiple of the difference value. The model converts the personnel performance difference into a dynamic correction amount of the risk index through mathematical mapping, realizes real-time association of risk prediction and personnel behavior state.

[0067] Compared with the prior art, the traditional method only judges the compliance of personnel operation through a fixed threshold, and does not establish a dynamic association mechanism of the performance score and the risk index. The present application introduces an operation personnel performance adjustment factor , which quantifies the degree of deviation of personnel performance from the reference value as a correction coefficient of the risk index, overcoming the technical defect that the influence of personnel factors in the static model is difficult to dynamically feedback.

[0068] Through the above technical scheme, the present application realizes the technical effect of dynamically adjusting the production quality risk index according to the actual behavior performance of the operation personnel. When the operation personnel inspection completion rate decreases or the operation error rate increases, the system automatically raises the risk warning level. This dynamic correction mechanism effectively solves the problem that the personnel behavior factor is difficult to quantify in the traditional quality control and risk assessment, and improves the accuracy and timeliness of risk warning.

[0069] Preferably, the present application further proposes that the generation mode of the final production quality risk index is specifically: The final production quality risk index R is generated through the formula:

[0070] The final production quality risk index R is generated through the formula: In the formula, R represents the original production quality comprehensive risk index, represents the operation personnel performance adjustment factor; The original production quality comprehensive risk index R is an initial risk assessment value calculated based on the production material quality, production conditions and equipment state characteristic values, and can be realized by using a weighted summation model. This index reflects the influence degree of objective environmental factors on the quality of veterinary drugs. The operation personnel performance adjustment factor is a dynamic adjustment coefficient generated based on personnel behavior data, which can be realized by calculating the difference between the performance score and the reference value, and is used to quantify the correction effect of personnel operation on quality risk. In order to limit the product result within the range of 0 to 1, a numerical truncation algorithm can be used to realize it, which can prevent the risk index from exceeding the reasonable threshold due to too low personnel performance.

[0071] Specifically, when the operation personnel performance score is higher than the reference value, the adjustment factor θ is less than 1, at this time the original risk index R is reduced, and the final risk index​ decrease; when the performance score is lower than the benchmark value, greater than 1, the original risk index is amplified, increase. This dynamic adjustment mechanism realizes the quantitative compensation of human factors by coupling the personnel performance with the objective risk index. Constraint condition Ensure that even if the personnel performance is abnormally low, the risk index will not exceed the preset upper limit, and maintain the stability of the risk warning system.

[0072] Compared with the prior art, the traditional method only relies on a fixed threshold to judge the risk level, and fails to establish a dynamic correlation between personnel performance and objective risk, resulting in a lag in risk assessment. The present application introduces an adjustment factor and a constraint mechanism, which not only realizes real-time feedback of personnel operation performance, but also builds an adaptive adjustment system for risk index, so that quality warning can reflect the changes in personnel operation state in real time.

[0073] Through the above technical solution, the present application solves the technical defect that the traditional model cannot dynamically compensate for the influence of personnel operation, and realizes real-time correction of the risk index. Specifically, when the operator delays or makes mistakes, the adjustment factor automatically increases the risk index, and the time point of triggering the warning can be advanced to before the quality anomaly occurs; when the personnel operation is efficient, the system can appropriately reduce the risk level to avoid production interruption caused by excessive warning. This dynamic calculation mechanism enables the quality control system to more accurately reflect the actual risk state of the production site.

[0074] Preferably, the present application also proposes a veterinary drug production quality control system for executing the above-mentioned veterinary drug production quality control method, specifically comprising: A data acquisition unit is used to acquire the production material quality characteristic value, production condition characteristic value, production equipment state data and production operator behavior data of the veterinary drug; the characteristic value is the better the larger, for example, the larger the production material quality characteristic value, the better the production material quality; the characteristic value is obtained by normalizing the data; A comprehensive analysis unit is used to establish an original risk prediction model according to the production material quality characteristic value, production condition characteristic value and production equipment state data, and generate an original production quality comprehensive risk index; A behavior analysis unit is used to establish an operator performance analysis model according to the production operator behavior data, and generate an operator performance adjustment factor; A final risk analysis unit is used to generate a final production quality risk index according to the original comprehensive risk index and the operator performance adjustment factor; A control unit is used to control the production of the veterinary drug according to the final production quality risk index.

[0075] Preferably, the application further provides that the comprehensive analysis unit specifically comprises: a production material quality analysis module, configured to generate a production material quality risk index according to production material quality characteristic values; a production condition analysis module, configured to generate a production condition risk index according to production condition characteristic values; a production equipment analysis module, configured to generate a production equipment risk index according to production equipment state data, wherein the production equipment state data comprises historical key process compliance rates of the production equipment and recent failure rates of the production equipment; a comprehensive risk index generation module, configured to establish an original risk prediction model according to the production material quality risk index, the production condition risk index and the production equipment risk index, and generate an original production quality comprehensive risk index.

[0076] Preferably, the application further provides that the behavior analysis unit specifically comprises: an operator performance score generation module, configured to generate an operator performance score according to operator behavior data; an adjustment factor generation module, configured to establish an operator performance analysis model according to the operator performance score, and generate an operator performance adjustment factor.

[0077] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, alternatives and variations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for quality control of veterinary drug production, characterized in that, The method specifically comprises the following steps: Obtaining production material quality characteristic values, production condition characteristic values, production equipment state data and production operator behavior data of the veterinary drug; According to the production material quality characteristic values, the production condition characteristic values and the production equipment state data, an original risk prediction model is established to generate an original production quality comprehensive risk index; According to the production operator behavior data, an operator performance analysis model is established to generate an operator performance adjustment factor; wherein the production operator behavior data comprises an operator inspection task completion rate, an operator abnormal record reporting rate of the production equipment and an operator operation error rate; According to the original comprehensive risk index and the operator performance adjustment factor, a final production quality risk index is generated; According to the final production quality risk index, the production of the veterinary drug is controlled.

2. The method of production quality control of veterinary drugs according to claim 1, characterized by, The generation mode of the original production quality comprehensive risk index specifically comprises: According to the production material quality characteristic values, a production material quality risk index is generated; According to the production condition characteristic values, a production condition risk index is generated; According to the production equipment state data, a production equipment risk index is generated; wherein the production equipment state data comprises a historical key process compliance rate of the production equipment and a recent failure rate of the production equipment; According to the production material quality risk index, the production condition risk index and the production equipment risk index, the original risk prediction model is established to generate the original production quality comprehensive risk index.

3. The method of veterinary pharmaceutical production quality control according to claim 2, c h a r a c t e r i z e d b y, The generation mode of the production material quality risk index specifically comprises: Through the formula: Generating a production material quality risk index ; In the formula, represents the i-th normalized production material quality characteristic value, represents the weight coefficient of the i-th material quality characteristic, and p represents the number of production material quality characteristics.

4. The method of production quality control of veterinary drugs according to claim 2, characterized by, The generation mode of the production condition risk index specifically comprises: Through the formula: Generating a production condition risk index ; In the formula, represents the jth production condition characteristic value, represents the weight coefficient of the jth production condition characteristic, and q represents the number of production condition characteristics.

5. The method for controlling production quality of veterinary drugs according to claim 2, wherein, The generation mode of the production equipment risk index specifically comprises: Through the formula: Generating a production facility risk index ; In the formula, represents the historical critical process compliance rate of the production equipment, represents the minimum value of the standard critical process compliance rate, represents the recent failure rate of the production equipment, represents the maximum failure rate of the standard production equipment, represents the weight coefficient of the critical process compliance rate.

6. The method for controlling production quality of veterinary drugs according to claim 2, wherein, The expression of the original risk prediction model is specifically: In the expression, R represents the original production quality comprehensive risk index, represents the production material quality risk index, represents the production condition risk index, represents the production equipment risk index, , , are weight coefficients, and .

7. The method of veterinary pharmaceutical production quality control according to claim 1, wherein, The generation mode of the operator performance adjustment factor specifically comprises: According to the operator behavior data, an operator performance score is generated; According to the operator performance score, the operator performance analysis model is established to generate the operator performance adjustment factor.

8. The method of veterinary pharmaceutical production quality control according to claim 7, c h a r a c t e r i z e d b y, The generation mode of the operator performance score specifically comprises: Through the formula: Generating operator performance scores ; In the formula, represents the normalized value of the historical on-time completion rate of the operator's inspection tasks, represents the normalized value of the historical reporting rate of the operator to the abnormal records of the production equipment, represents the normalized value of the historical operation error rate of the operator, are weight coefficients, and .​​ 9. The method of veterinary pharmaceutical production quality control according to claim 7, wherein, The expression of the operator performance analysis model is specifically: In the formula, represents the operator performance adjustment factor, represents the operator performance score, represents the reference value of the operator performance score, and k represents the adjustment intensity coefficient.

10. The method of veterinary pharmaceutical production quality control according to claim 1, wherein, The generation mode of the final production quality risk index is specifically: Through the formula: generating a final production quality risk index ; In the formula, represents the original production quality comprehensive risk index, represents the operator performance adjustment factor.

Images