Non-Gaussian and persistence measures for control loop quality assessment

Area of Science:

• Control Engineering
• Data Analysis
• Industrial Systems

Background:

• Traditional control performance assessment methods may not capture complex system dynamics.
• Nonlinear time series analysis offers advanced tools for evaluating system behavior.

Purpose of the Study:

• To review and compare alternative nonlinear time series methodologies for control performance assessment.
• To provide a practical rationale for applying these advanced analytical techniques.
• To evaluate the effectiveness of these methods using real-world industrial data.

Main Methods:

• Non-Gaussian statistical analysis utilizing various probabilistic distribution functions.
• Fractal analysis, including the calculation of the Hurst exponent via multiple approaches.
• Rescaled range (R/S) plot analysis and discussion of the crossover point phenomenon.

Main Results:

• Demonstration of non-Gaussian statistics for identifying system deviations.
• Application of Hurst exponent calculations to quantify time series properties.
• Analysis of R/S plots and crossover points for performance insights.

Conclusions:

• Nonlinear time series analysis provides robust alternatives for control performance assessment.
• The evaluated methods offer practical value when applied to industrial data.
• Further research into these techniques can lead to enhanced system monitoring and control.