Closed-loop controller using statistical distribution optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern closed-loop control systems face challenges with measurement noise and uncertainties, leading to extreme sensitivities and reduced operational envelopes, which limit system performance and require oversized systems to mitigate uncertainties.
Innovation Solution
A closed-loop controller that utilizes a control effectiveness function with a statistical distribution (μ,σ2) to generate manipulated variables, allowing for expanded operation by minimizing errors through a probability density function around system state and output parameters, rather than relying on specific values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If inviolable thresholds are designed well within the system's available operating envelope to reduce uncertainty impacts, then system reliability is improved, but the operational envelope size is reduced and system performance is degraded
Solution Approach 1:
The patent transforms the control approach by changing from deterministic parameter values to statistical distribution parameters (mean and variance). The control allocator minimizes expected quadratic error based on distributed variables representing uncertainty, allowing the system to operate near boundaries while maintaining reliability through statistical rather than conservative threshold design.
2Reliability
If oversized systems are designed to mitigate uncertainties, then system reliability is improved, but device complexity and system size increase
Solution Approach 1:
The patent replaces physical oversizing (mechanical solution) with a computational control strategy. Instead of designing larger systems with margins to handle uncertainty, the invention uses real-time statistical optimization in the control allocator to adaptively manage uncertainty, achieving the same reliability effect without increasing physical system size.
3Measurement precision
If measurement noise and uncertainties are treated as limitations on system fidelity, then measurement precision is maintained, but system performance and responsiveness are degraded
Solution Approach 1:
The patent converts the harmful effect of measurement noise and uncertainties into a beneficial feature by explicitly modeling them as distributed variables with known statistical properties. The control allocator uses these distributions to compute expected errors and minimize them proactively, transforming uncertainty from a performance-degrading factor into a manageable parameter that enables expanded operation.
Data Source
AI summary
The disclosure provides a closed-loop controller for a controlled system comprising a comparison element generating an error e, a compensator generating a control uN value based on the error e, and a control allocator determining a manipulated parameter uM value based on the control uN. The control allocator typically utilizes a control effectiveness function and determines uM value by selecting one or more specific system x0 signals from the system state xi or system input yj values or system parameters values pk reported, defining a plurality of distributed xD around each specific system x0 signal, and minimizing an error function E(zi), where the error function E(zi) is based on errors which arise from use of the plurality of distributed xD in the control effectiveness function rather than one or more specific system x0 signals.