Filling Process Control Using Historical Flow Rate Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing filling processes face inaccuracies due to measurement noise and variances in flow behavior, leading to incomplete correction of filling quantities, especially when dealing with liquids that have high particle content or fluctuating flow rates.
Innovation Solution
A method that corrects current flow rate measurements using historical data from previous filling processes, integrating the current and earlier measurement values to reduce noise and improve accuracy, allowing for precise control of the filling quantity by estimating the filled volume through a time series of flow rate measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If flow rate measurement is used to determine filling quantity, then filling accuracy can be improved, but measurement noise and errors in flow rate determination lead to inaccurately determined filled quantity
Solution Approach 1:
The patent implements feedback by using a control module that continuously monitors the actual filling quantity via flow rate measurement and compares it with the target filling quantity. Based on this comparison, the control module adjusts the filling valve in real-time to compensate for deviations, ensuring accurate filling despite measurement noise and flow rate variations.
Solution Approach 2:
The patent applies preliminary action by pre-calculating the expected filling curve (target course) before the filling process begins. This target course serves as a reference for the control module to compare against actual measurements and make corrective adjustments during the filling process, preventing accumulation of errors.
2Measurement precision
If valve closing time is corrected during filling based on detected flow rate changes, then filling accuracy can be improved, but measurement errors and noise of the flow rate remain uncompensated and fully affect the closing time
Solution Approach 1:
The control module continuously monitors flow rate during filling and provides real-time feedback to adjust valve closing time. When measurement noise or errors are detected, the system compensates by comparing actual flow rate data against the pre-calculated target course and adjusting the closing time accordingly, preventing measurement errors from fully affecting the final filling accuracy.
Solution Approach 2:
The patent pre-calculates the target filling curve before the filling process begins, establishing an expected flow rate profile. This preliminary action provides a reference framework that allows the control system to distinguish between normal flow variations and actual measurement errors, enabling more precise correction of valve closing time.
3Measurement precision
If filling time is adjusted from total flow rate of multiple fillings to average out measurement errors, then individual measurement inaccuracies are reduced, but actual variance of individual filling is also averaged out
Solution Approach 1:
Instead of averaging multiple filling measurements, the patent implements real-time feedback during each individual filling process. The control module monitors flow rate continuously and adjusts the filling process dynamically, ensuring that each filling operation achieves the target quantity independently without being affected by variations in other fillings.
Solution Approach 2:
The patent pre-calculates a target filling curve for each individual filling operation based on expected flow rate characteristics. This preliminary action enables precise control of each filling independently, maintaining the actual variance of individual filling quantities rather than averaging them out.
4Extent of automation
If a neural network is used to determine flow rate and adjust control data, then filling control can be improved, but measurement errors for the flow rate still lead to inaccuracies in the filled quantity
Solution Approach 1:
The patent implements feedback by continuously comparing actual flow rate measurements against the neural network's predictions and adjusting the filling valve in real-time. This feedback mechanism compensates for measurement errors by dynamically correcting the filling process based on the deviation between expected and actual flow rates.
Solution Approach 2:
The neural network is trained beforehand to establish the relationship between valve position, flow rate, and filling quantity. This preliminary training action enables the system to predict optimal control data that compensates for expected measurement errors and flow rate variations during actual filling operations.
Data Source
AI summary
A method for controlling a filling process, wherein a predetermined filling quantity of a medium is filled into a container, the flow rate of the medium flowing into the container is measured as a time series of measured values for the instantaneous flow rate and a filling quantity already filled is estimated from the time series, wherein at least one current measured value of the time series is corrected on the basis of at least one earlier measured value of an earlier time series of measured values of the flow rate of an earlier filling process.