Process Instrument Early Alert Using Linear and Cyclic Prediction

Solution Overview

Problem

Existing process instruments rely on standard warning and alarm mechanisms that only trigger after a measurement value exceeds a preset threshold, lacking the capability for early prediction and alerting of potential process variable deviations.

Innovation Solution

A method and apparatus for predicting future process variable measurement values using a linear prediction model and accounting for data periodicity, allowing for early alert signals to be sent if predicted values exceed a preset warning threshold.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If a standard warning threshold mechanism is used, then the system structure is simple and easy to operate, but the warning is issued only after the measurement value has already exceeded the threshold, lacking early prediction capability

Engineering Contradiction:

Improvewarning timeVSAvoidprediction model complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by performing prediction calculations before the measurement value actually exceeds the warning threshold. The linear prediction model uses recent historical data to estimate future values, and the periodicity analysis predicts values based on cyclical patterns, allowing the system to issue warnings in advance of threshold violations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent uses partial action by selectively applying different prediction methods based on data characteristics. The system determines whether to use linear prediction, periodicity-based prediction, or both methods depending on the specific measurement data, rather than always applying complex prediction algorithms to all data points.

Inventive Principle:

Principle #16Partial or excessive action

2Reliability

If no prediction mechanism is used, then the device complexity is low, but the reliability of process monitoring is reduced due to inability to predict future deviations

Engineering Contradiction:

Improvefault prediction accuracyVSAvoidprediction system complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements feedback by continuously comparing predicted values with actual measurement values. The system uses recent historical data to build prediction models, issues warnings when predictions indicate threshold violations, and continuously updates its predictions based on new measurement data, creating a closed-loop feedback system that improves reliability.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent applies preliminary action by performing prediction calculations before the measurement value actually exceeds the warning threshold. The linear prediction model uses recent historical data to estimate future values, and the periodicity analysis predicts values based on cyclical patterns, allowing the system to issue warnings in advance of threshold violations.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If dual prediction methods (linear and periodicity-based) are used, then the measurement precision of future values is improved, but the computational complexity increases

Engineering Contradiction:

Improveprediction precisionVSAvoidcalculation complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent uses partial action by selectively applying different prediction methods based on data characteristics. The system determines whether to use linear prediction, periodicity-based prediction, or both methods depending on the specific measurement data, rather than always applying complex prediction algorithms to all data points.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent applies segmentation by dividing the prediction process into distinct modules: linear prediction module, periodicity analysis module, and value combination module. Each module handles a specific aspect of the prediction, allowing the system to achieve high precision through multiple specialized methods while maintaining manageable complexity through modular architecture.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250155877A1Early Alert Method and Apparatus, Process Instrument Comprising Early Alert Apparatus, and Computing Device

Publication Date: 2025.05.15 SIEMENS AG

AI summary

An example includes: receiving a process variable; predicting a estimation value by linear fitting on a number of recent values, comparing that with a warning threshold, and if greater, setting the estimation value as a first value; comparing the current value with a value at the same position in the previous cycle to determine a deviation value, adding the deviation value to the previous peak value or trough value to serve as a maximum estimation value, comparing the maximum value with the preset warning threshold, and if greater, setting the maximum value as a second estimation value; computing a estimation value on the basis of the first and the second estimation; comparing the process variable estimation value with the preset warning threshold; and issuing an early alert signal if the estimation value is greater than the threshold.