Method for determining oscillations occurring in a measuring signal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting oscillations in control variables of PID controllers, such as those used in heating, ventilation, and air conditioning systems, face challenges in accurately identifying oscillations in complex systems with changing dynamics, as they often lose information about the timing of oscillations and struggle to provide robust detection across different sensor types and systems.
Innovation Solution
The method employs the rainflow counting method to identify and analyze closed loops in measuring signals, allowing for efficient detection of oscillations regardless of sensor type, by counting half loops and forming closed loops, which are then used to determine oscillation amplitude, period, and timing, enabling robust and time-saving analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Fast Fourier Transform with a broad window is used to analyze oscillations, then frequency resolution is improved, but time resolution deteriorates making it difficult to detect changing frequencies
Solution Approach 1:
The patent segments the measuring signal into multiple partial signals using a sliding window approach, where each partial signal is analyzed separately for closed loops. This allows the system to achieve both good frequency resolution (by analyzing each segment with sufficient window length) and good time resolution (by moving the window through the signal to capture changing frequencies over time). The segmentation enables localized frequency analysis without losing temporal information about when oscillations occur.
2Measurement precision
If Fast Fourier Transform is used to separate control variable into oscillation components, then oscillation analysis is enabled, but information about precise timing of oscillation start and end is lost
Solution Approach 1:
The patent applies preliminary action by first identifying extreme values and determining closed loops in the time domain before performing any frequency analysis. The closed loop detection method identifies the precise start and end times of oscillations by detecting when the signal crosses reference lines and forms complete cycles. This preliminary time-domain analysis preserves timing information that would otherwise be lost in pure frequency-domain approaches like FFT.
3Reliability
If PID controller parameters are set using Ziegler and Nichols heuristic methods, then disturbance behavior is improved, but management behavior becomes oscillating
Solution Approach 1:
The patent implements feedback by continuously monitoring the control variable for closed loops and using this information to adaptively adjust PID controller parameters. The system detects oscillations in real-time by identifying closed loops in the measuring signal, then feeds this information back to modify the controller parameters to eliminate the oscillations. This closed-loop feedback mechanism allows the system to maintain good disturbance behavior while achieving stable management behavior through continuous adaptation.
4Stability of the object's composition
If controller parameters are adaptively set based on oscillation detection, then management behavior stability is improved, but detection accuracy becomes decisive for regulating behavior
Solution Approach 1:
The patent replaces complex frequency-domain mechanical analysis (FFT) with a simpler time-domain geometric approach for detecting closed loops. Instead of transforming the signal to frequency domain and analyzing spectral components, the method directly identifies closed loops by detecting extreme values, drawing reference lines, and counting signal crossings. This substitution maintains high detection accuracy while simplifying the computational mechanism, making the system more robust and easier to implement.
Data Source
AI summary
A method for determining oscillations occurring in a measuring signal. The method includes the steps of receiving a measuring signal, determining the extreme values of the received measuring signal, and ascertaining closed loops of the measuring signal, by a) identifying a closed loop in the measuring signal (a closed loop being formed by two half loops having identical oscillation width and opposite direction, b) storing the identified closed loop, c) removing the identified closed loop from the measuring signal, and d) repeating steps a) through c) until all closed loops have been ascertained.