Multi-Stage Production Control for Intermediate Product Fluctuations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In production systems with multiple continuous processes, fluctuations in the state of intermediate products due to device capacity declines, such as catalyst deterioration or fouling, make it difficult to control the yield and quality of the final product.
Innovation Solution
A production system comprising first and second devices, sensors, and a control device that uses machine learning models to estimate and adjust condition lists based on device and product information, allowing for the control of the state of processed products across multiple processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a production system operates multiple continuous processes, then productivity is improved, but device capacity decline causes fluctuations in intermediate product state making final product control difficult
Solution Approach 1:
The system implements a feedback mechanism where the state of the first processed product (detected by sensors) is fed back to the control device, which then adjusts the second condition list to compensate for deviations. This closed-loop control ensures that even when device capacity declines, the system can maintain final product quality by dynamically adjusting process parameters based on real-time state information.
Solution Approach 2:
The system performs preliminary estimation of the first processed product state using a first model before the actual second process occurs. Based on this preliminary estimation, the control device proactively modifies the second condition list in advance, preventing potential quality deviations rather than reacting to them after they occur.
2Manufacturing precision
If device capacity declines due to catalyst deterioration or fouling, then manufacturing precision of intermediate product is affected, but the system needs to maintain final product quality
Solution Approach 1:
When device capacity declines, the system changes parameters in the second condition list to compensate. The control device calculates modified values for control parameters (such as temperature, pressure, flow rate) based on the estimated state of the first processed product, thereby maintaining final product quality despite intermediate product variations caused by device degradation.
Solution Approach 2:
The control device acts as an intermediary that receives information about intermediate product state deviations and translates them into appropriate adjustments for the second process. This intermediary function allows the system to decouple the impact of device capacity decline from final product quality, as the control device mediates between the two processes.
3Measurement precision
If the system uses machine learning models to estimate processed product properties, then control accuracy is improved, but device complexity increases
Solution Approach 1:
The system replaces complex physical measurement and control mechanisms with machine learning models. Instead of using sophisticated sensors and control hardware to directly measure and adjust processed product properties, the system uses computational models (first model and second model) that process sensor data and predict optimal control parameters, thereby reducing the need for complex mechanical control systems.
Data Source
AI summary
A production system includes a first device, a second device, a first sensor, a second sensor and a control device. The control device controls the first device based on a first condition list including a plurality of control parameters of the first device, and controls the second device based on a second condition list including a plurality of control parameters of the second device. The control device acquires a first processed product information, acquires a modified value of the second condition list based on a second model that outputs a modified value of the second condition list in response to the first processed product information, acquires the second condition list modified based on the second condition list and the modified value of the second condition list, and controls the second device based on the modified second condition list.