Process Prediction Error Correction for Safer Plant Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction models for monitoring and controlling plants often suffer from accuracy errors, leading to potentially unsafe outcomes when used directly for monitoring or control, and methods to mitigate these errors are limited, especially when constructing multiple accurate models is challenging.
Innovation Solution
A prediction device and method that collects process data, constructs prediction models with error calculation models, and corrects prediction values based on these errors to ensure safer, more efficient outputs, including state monitoring and operation amount determination for abnormal conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a prediction model is used to monitor or control a plant, then monitoring and control functions are achieved, but prediction errors may lead to unsafe or undesired results
Solution Approach 1:
The patent applies preliminary action by constructing an error calculation model in advance that predicts the prediction error before the main prediction is used for monitoring or control. This pre-calculated error information is then used to correct the prediction value, ensuring safety considerations are built into the prediction process before deployment.
Solution Approach 2:
The error calculation model serves as an intermediary between the prediction model and the final monitoring/control decisions. It calculates the prediction error as a separate computational layer that mediates the relationship between raw prediction values and corrected safe values, allowing error compensation without modifying the original prediction model.
2Measurement precision
If multiple prediction models are constructed with weighted integration to reduce errors, then prediction accuracy improves, but device complexity increases
Solution Approach 1:
The patent extracts the error calculation function as a separate, independent model from the prediction model itself. Instead of integrating multiple prediction models, it takes out the error prediction capability as a distinct error calculation model that works alongside a single prediction model, simplifying the overall system architecture.
Solution Approach 2:
The patent changes the parameter being modeled from the target variable itself to the prediction error of the target variable. By constructing an error calculation model that predicts the error rather than the target directly, it enables error compensation with a single auxiliary model instead of requiring multiple prediction models.
3Ease of operation
If prediction values are used directly for monitoring without error correction, then simplicity is maintained, but safety and reliability deteriorate
Solution Approach 1:
The system performs preliminary error calculation and correction before the prediction value is used for monitoring decisions. The error calculation model pre-computes the expected error, and this correction is applied in advance to the prediction value, ensuring safety is built into the monitoring process without adding operational complexity.
Data Source
AI summary
Provided is a prediction device that outputs a prediction value of process data in consideration of a prediction error of a prediction model. A prediction device includes a data collection unit that collects process data of a device; a prediction model construction unit that constructs a prediction model having a predetermined input variable of first process data as an input value and having a predetermined output variable as an output value, and an error calculation model which calculates a prediction error of the prediction model, based on the first process data collected by the data collection unit; and a prediction unit that outputs a prediction value which is corrected based on a prediction value of the output variable for second process data and a prediction error for the prediction value of the output variable, the prediction value being predicted based on the input variable of the second process data collected by the data collection unit, the prediction model, and the error calculation model.