Industrial Parameter Trend Prediction for Boundary Warnings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting whether an industrial system parameter will go outside a boundary within a future period of time are inaccurate, inefficient, and unreliable, particularly for non-linear changes over time.
Innovation Solution
A computerized method that divides a preceding period into multiple time periods, calculates time-period-specific intermediate values, combines these values to determine a projected slope, and provides warnings when a parameter is likely to exceed a boundary, using linear regression and slope differences to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional monitoring methods are used to track parameter values, then operators can view current parameter states, but they cannot accurately predict future parameter deviations outside boundaries
Solution Approach 1:
The system performs preliminary calculations by dividing the preceding time period into multiple segments and computing intermediate values in advance. This allows the system to project future parameter values and predict boundary violations before they occur, enabling operators to take preventive actions.
Solution Approach 2:
The preceding time period is divided into multiple time periods, and each period is processed to calculate intermediate values. This segmentation allows for more granular analysis of parameter trends and improves the accuracy of future value predictions by capturing non-linear changes over time.
2Productivity
If multiple parameter values are displayed simultaneously on screens, then operators can monitor current system state, but they cannot efficiently identify parameters projected to go outside boundaries
Solution Approach 1:
The system extracts and highlights only the critical information - parameters that are projected to go outside their boundaries within the future time period. Instead of requiring operators to scan all parameter values, the system isolates and presents only those parameters needing attention, significantly reducing operator workload while improving detection efficiency.
Solution Approach 2:
The system acts as an intermediary between the complex process data and the operator by performing the predictive analysis and boundary violation detection automatically. It translates raw parameter values into meaningful predictions about future boundary violations, eliminating the need for operators to manually analyze trends.
3Measurement precision
If simple projection methods are used to predict parameter values, then calculations are fast and simple, but accuracy is poor especially for non-linear changes
Solution Approach 1:
By dividing the preceding time period into multiple smaller time periods and calculating intermediate values for each segment, the system captures non-linear changes more effectively. This segmented approach allows for more accurate trend analysis while keeping individual calculations manageable.
Solution Approach 2:
The system dynamically adjusts its analysis by considering multiple time periods and calculating intermediate values that reflect changing trends. This dynamic approach allows the projection method to adapt to non-linear parameter changes while maintaining computational efficiency through structured calculations.
Data Source
AI summary
Potentially-abnormal behavior of an industrial system is communicated to an operator using a computer including displaying measured parameters, evaluating whether a parameter is projected to go outside of a boundary, and warning the operator accordingly. A preceding period may be divided into multiple time periods, values of the parameter may be recalled at different times during the periods, and the values may be used to calculate whether the parameter is projected to go outside the boundary. In some embodiments, intermediate values (e.g., slopes) may be calculated for the time periods and may be combined into a combined intermediate value or combined slope used to determine whether the particular parameter is projected to go outside of the boundary within a predetermined future period of time. Changes in intermediate values may be calculated, averaged, subtracted, or divided (e.g., by a slope change factor).