Sensor Signal Analysis for Transient Error Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control systems in industrial settings, such as power plants, often fail to detect transient errors in sensor signals due to their sampling rates, leading to delayed responses or unnecessary corrective actions, and may introduce aliasing or noise, which can be harmful.
Innovation Solution
A sensor analysis device is coupled between sensors and control systems to monitor sensor signals at an increased sample rate, detect transient errors, and modify the output signal to suppress or mask these errors, preventing their transmission to the control system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the control system uses a pre-defined sample rate (e.g., every 40ms), then the control system operates with standardized timing, but transient errors occurring between samples remain undetected
Solution Approach 1:
The patent applies preliminary action by implementing a buffer that continuously stores incoming sensor samples at the original sample rate before they are processed by the control system. This buffer allows the system to retrospectively analyze any time window, enabling detection of transient errors that occurred between original sample points without requiring the control system to operate at a higher continuous sample rate.
Solution Approach 2:
The patent introduces another dimension by adding a time-windowing capability that allows analysis of historical data at different resolution levels. Instead of simply increasing the sampling rate in the time domain, the system creates a multi-resolution analysis capability where data can be examined at both the original coarse resolution and a finer interpolated resolution within specific time windows, effectively adding a dimensional aspect to error detection.
2Object-affected harmful factors
If low-pass filtering is applied to mitigate transient detection issues, then noise is reduced, but the response is delayed or errors go undetected
Solution Approach 1:
The patent applies segmentation by dividing the signal processing into distinct functional stages: (1) continuous buffering of raw samples at original resolution, (2) selective extraction of time windows based on control system needs, (3) optional interpolation for enhanced analysis, and (4) controlled filtering only when and where needed. This segmentation allows different processing strategies to be applied to different portions of the data stream, maintaining transient detection capability while providing noise reduction where beneficial.
3Reliability
If the sampling frequency is increased to detect transient errors, then transient detection capability is improved, but aliasing occurs when transients occur periodically at near-integer multiples of the sampling frequency
Solution Approach 1:
The patent uses interpolation as an intermediary technique between the original sampled data and the desired high-resolution analysis. By applying interpolation algorithms to the buffered samples, the system generates intermediate data points that estimate what the signal values would have been between actual sample points. This intermediary approach enables transient detection at effective rates much higher than the original sampling rate without actually acquiring data at those rates, thereby avoiding aliasing issues.
4Measurement precision
If transient errors are detected or noisy signals generate erroneous samples, then error detection sensitivity is improved, but the control system triggers unnecessary corrective responses
Solution Approach 1:
The patent applies dynamics by implementing adaptive thresholding and configurable time-window analysis. Instead of using fixed thresholds for error detection, the system allows thresholds to be dynamically adjusted based on historical data, signal characteristics, and operational context. The time-windowing capability also allows the system to evaluate whether detected anomalies persist over meaningful time periods before triggering alerts, filtering out transient noise that would otherwise cause false alarms while maintaining sensitivity to genuine errors.
Data Source
AI summary
A sensor analysis system for monitoring operation of a turbine engine of a power plant, the sensor analysis system including a turbine control system configured to monitor a plurality of sensors, a sensor device configured to monitor operation of the turbine engine and transmit a first analog signal, and a sensor analysis device coupled to the sensor device. The sensor analysis device is configured to receive the first analog signal, capture a first plurality of samples from the first analog signal at a first rate defined by a device sampling period, determine a second rate defined by the device sampling period, capture a second plurality of samples from the first analog signal at the second rate, and detect one or more samples of the second plurality of samples that includes a transient signal above or below a threshold value.