Membrane Scouring Airflow Control for Stable Filtration Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing membrane separation systems face challenges in maintaining stable membrane filtration operations and energy conservation due to delays in air diffusion amount control when sudden changes occur in membrane filtration conditions, leading to potential inter-membrane clogging or excessive energy consumption.
Innovation Solution
A membrane scouring airflow control system that utilizes machine learning to infer the state of the separation membrane and adjusts air diffusion levels based on real-time input data, implementing first and second control levels to stabilize operations and minimize delays in air diffusion adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning inference is performed for each cycle of membrane filtration operation, then the air diffusion amount can be closely controlled based on input data such as trans-membrane pressure, but the system cannot respond quickly to sudden changes in operating conditions
Solution Approach 1:
The system performs preliminary machine learning inference during pause periods to predict the membrane state for the upcoming operation period. This advance preparation allows the system to have control parameters ready before sudden changes occur, improving response speed while maintaining control precision through the learned model.
Solution Approach 2:
The control system dynamically adjusts its operation mode between pause periods (learning/inference) and operation periods (execution). During operation periods, the system executes control based on inferred states, while during pause periods, it updates predictions and prepares for upcoming changes, creating a dynamic response mechanism that balances precision and speed.
2Productivity
If the membrane filtration flow rate suddenly increases, then the treatment capacity is improved, but the air diffusion amount becomes insufficient causing inter-membrane clogging
Solution Approach 1:
The system continuously monitors operation data including membrane filtration flow rate and uses machine learning to infer the membrane state. When the flow rate suddenly increases, the feedback mechanism detects this change and adjusts the air diffusion amount accordingly during the next pause period, preventing clogging while maintaining high productivity.
Solution Approach 2:
During pause periods, the system performs preliminary inference about the membrane state based on accumulated operation data. This allows the system to predict and prepare appropriate air diffusion adjustments before the next operation period begins, ensuring that when high flow rates are needed, the membrane is already in an optimal state to handle the increased load without clogging.
3Productivity
If the membrane filtration flow rate suddenly decreases, then the operational load is reduced, but the air diffusion amount becomes excessive making energy saving difficult
Solution Approach 1:
The system uses feedback from operation data to continuously update its understanding of the membrane state. When the filtration flow rate suddenly decreases, the feedback mechanism detects this reduction and adjusts the air diffusion amount downward in the subsequent pause period, reducing energy consumption to match the actual operational needs and preventing excessive air diffusion.
4Reliability
If air diffusion amount is increased to prevent clogging, then membrane operation stability is improved, but energy consumption increases
Solution Approach 1:
The system dynamically changes the air diffusion parameter based on inferred membrane state and actual operating conditions. Rather than maintaining a constant high air diffusion rate, the system adjusts this parameter in response to detected changes in filtration flow rate and membrane state, ensuring adequate clogging prevention while minimizing unnecessary energy consumption during stable operation periods.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
To provide a membrane scouring airflow control system, a membrane scouring airflow control method and a membrane scouring airflow control program that can stabilize the membrane filtration operation and improve energy conservation. A membrane scouring airflow control system 100 includes: an operation data acquiring apparatus 4 that acquires operation data measured during membrane filtration operation performed by a membrane separating apparatus 90; an input data calculating apparatus 5 that derives input data from the operation data; a learning model generating apparatus 1 that generates a learning model by using machine learning; a storage apparatus 3 that stores the learning model; an inference apparatus 2 that infers a state of a separation membrane 93, based on the input data by using the learning model; an input data determining apparatus 6 that determines the input data; and an air diffusion amount controlling apparatus 8 that executes first control of an air diffusing apparatus 95 so that air diffusion is performed at a first level, and executes at the same time second control of the air diffusing apparatus 95 so that the air diffusion is performed at the second level at a time when the input data determined by the input data determining apparatus 6 becomes a threshold or more or a threshold or less.