Fermentation Soft Sensor Using Knowledge Reuse for Stage-Specific Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fermentation processes face challenges in accurately predicting cell concentration due to the complexity of the process and the lack of effective online measurement tools, leading to weak generalization of soft sensor models and high costs associated with establishing separate models for different stages.
Innovation Solution
A knowledge reuse-based method and system that constructs a universal cell concentration soft sensor model for a fermentation process divided into stages, using a parameter estimation and gain matrix to design an online soft sensor for subsequent stages, thereby improving prediction accuracy by leveraging dynamic characteristics and detection delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single soft sensor model is used for the entire fermentation process, then the device complexity is reduced, but the measurement precision and reliability deteriorate due to reduced model generalization
Solution Approach 1:
The fermentation process is divided into multiple stages (lag phase, exponential growth phase, stationary phase, decline phase) based on dynamic characteristics. A separate soft sensor model is established for each stage, allowing the system to adapt to the specific characteristics of each phase while maintaining overall system manageability through modular model structure.
Solution Approach 2:
The patent implements dynamic model selection by identifying the current fermentation stage and switching between corresponding soft sensor models. The system dynamically adjusts which model is active based on real-time process conditions, ensuring optimal prediction accuracy for each stage while maintaining a unified overall framework.
2Measurement precision
If separate soft sensor models are established for each fermentation stage, then the measurement precision improves, but the device complexity and establishment costs increase
Solution Approach 1:
A unified soft sensor framework is established that can serve multiple fermentation stages. The same basic model structure and methodology are applied across all stages, with parameters adapted to each stage's characteristics. This universal approach reduces the overall complexity compared to developing entirely separate models for each stage.
Solution Approach 2:
The patent adjusts model parameters according to the specific characteristics of each fermentation stage while maintaining a consistent model structure. By changing parameters rather than entire model architectures, the system achieves stage-specific accuracy while minimizing the complexity increase associated with multiple models.
3Measurement precision
If separate soft sensor models are established for each fermentation stage, then the measurement precision improves, but the loss of time and resources increases due to high establishment costs
Solution Approach 1:
The patent performs preliminary identification of fermentation stages and prepares corresponding models in advance. By pre-establishing the framework and having stage-specific models ready, the system reduces the time required for model establishment and deployment during actual fermentation processes.
Solution Approach 2:
Rather than developing entirely new models for each stage, the system reuses the base model structure and adjusts parameters for each fermentation stage. This parameter adaptation approach significantly reduces the time and resources required compared to complete model re-establishment, while still achieving stage-specific prediction accuracy.
Data Source
AI summary
The present invention provides a knowledge reuse-based method and system for predicting a cell concentration in a fermentation process. The method includes: constructing a cell concentration soft sensor universal model in a fermentation process; acquiring and preprocessing process data of a fermentation stage A; determining a cell concentration soft sensor model of the fermentation stage A; designing a cell concentration online soft sensor of a fermentation stage B; and predicting a cell concentration of the fermentation stage B according to the cell concentration online soft sensor of the fermentation stage B. The present invention resolves the problems of weak generalization of a cell concentration soft sensor model and high costs of establishing models for fermentation stages separately, thereby improving the prediction accuracy of a cell concentration soft sensor.