Intelligent production monitoring method and system for injection

By dynamically adjusting local clustering differences and anomaly indices, the shortcomings of fixed thresholds and traditional density clustering algorithms in the production process of injectable solutions are addressed, enabling high-precision monitoring of the injectable solution production process and ensuring drug quality and safety.

CN122087489BActive Publication Date: 2026-07-14RUICHENG LVMAN BIOLOGICAL PHARM CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
RUICHENG LVMAN BIOLOGICAL PHARM CO LTD
Filing Date
2026-04-27
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

In the production process of injectable solutions, the fixed threshold method cannot adapt to the dynamic changes of multiple process stages, resulting in false alarms or missed alarms. Traditional density clustering algorithms ignore the continuity and contextual relationship of data points in the time dimension, making it difficult to distinguish between normal stage transitions and real process anomalies, thus affecting the accuracy of monitoring.

Method used

A local clustering difference calculation method is adopted. By dynamically adjusting the number of temporally adjacent data points and the merging distance threshold, stage boundary points are identified. Anomaly index is used to determine the abnormal state of the injection preparation tank. The merging strategy of clusters is dynamically adjusted by utilizing the neighborhood and temporal relationship of multi-dimensional data points.

Benefits of technology

It significantly reduced the false alarm rate, improved the monitoring accuracy in the production process of injectable solutions, and was able to accurately identify stage boundaries and anomalies, ensuring drug quality and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122087489B_ABST
    Figure CN122087489B_ABST
Patent Text Reader

Abstract

The present application belongs to the technical field of intelligent monitoring, and particularly relates to an intelligent production monitoring method and system for injection liquid, which comprises the following steps: collecting multi-dimensional data of an injection liquid preparation tank at each moment as data points; for any data point, determining a local clustering degree according to the change of its neighborhood data points in each dimension; calculating the difference of the local clustering degrees of its time-sequential neighboring data points as a local clustering degree difference; in response to the local clustering degree difference being greater than a difference threshold, iteratively obtaining the number of final time-sequential neighboring data points and determining a merging distance threshold; dynamically merging clustering clusters based on the merging distance threshold, obtaining stable clustering clusters and calculating an anomaly index, and determining the state abnormal moment of the injection liquid preparation tank. The present application solves the problem of misjudgment of the stage boundary as an anomaly in the traditional method, and significantly improves the monitoring accuracy.
Need to check novelty before this filing date? Find Prior Art