SMT-Based Stability Verification of an Industrial Switched PI Control Systems

Stylianos Basagiannis, Ludovico Battista, Anna Becchi, Alessandro Cimatti, Georgios Giantamidis, Sergio Mover, Alberto Tacchella, Stefano Tonetta, Vassilios A. Tsachouridis

Abstract

The control of complex systems is typically designed describing the physical system with differential equations. The standard approach to their verification employs numerical analysis, which is suitable to prove stability properties, but is susceptible to numerical errors. On the other side, symbolic techniques give precise analysis results but typically do not scale to industrial size problems. In this paper, we consider the control design of an aircraft engine. The engine model is represented by a linear state space model of 18 internal state variables, 4 outputs, and 3 inputs. The control switches between two PI controllers, one for thrust control and another for low-pressure compressor spool speed control, based on the engine state and pilot commands. After reformulating the PI controllers in terms of differential equations, we obtain a hybrid system with 21 state variables and two modes, for which we want to prove with symbolic techniques the robustness of the stable states to perturbation. We achieved the verification with standard methods to synthesize quadratic Lyapunov functions and SMT techniques to synthesize neighborhoods of the stable states for which we have symbolic proof of stability.

Introduction

In the safety-critical domain of aerospace engineering, ensuring the absolute stability of control systems is paramount. Engineers typically design these systems using differential equations and verify them through numerical simulations. While efficient, numerical methods are susceptible to floating-point errors and cannot provide a rigorous mathematical guarantee that a system will remain stable under all possible perturbations.

Symbolic verification, or formal methods, provides the necessary rigor but often suffers from the "state-space explosion" problem, where computational complexity grows exponentially with the number of variables. This research bridges the gap between numerical efficiency and symbolic precision by applying formal verification to an industrial-grade aircraft engine model featuring 21 variables—a significant leap in scale for formal methods.

System Model and Control Architecture

The study utilizes a high-fidelity mathematical model of a dual-spool turbofan engine. The physical plant is characterized by 18 internal state variables representing various engine components and dynamics.

Diagram of the dual-spool turbofan engine components and internal states.

The control system employs a switched Proportional-Integral (PI) architecture. It alternates between two primary control modes: thrust control and low-pressure compressor (LPC) spool speed control. This switching logic, driven by pilot commands and engine state, creates a hybrid dynamical system that must be verified for stability across its operational envelope.

Control system architecture showing the interaction between controllers, switching logic, and the aero engine.

Verification Methodology

The authors developed a multi-stage pipeline to achieve formal verification:

  1. System Reformulation: The switched PI controller and engine model were integrated into an autonomous Piecewise-Affine (PWA) system, treating the controller and engine as a single hybrid entity.
  2. Lyapunov Synthesis: To prove stability, the team synthesized "Lyapunov functions"—mathematical constructs that ensure the system's energy decreases toward a steady state. Six methods were tested, including solving Lyapunov equations and using Linear Matrix Inequalities (LMIs).
  3. SMT-Based Validation: Once candidate functions were generated numerically, Satisfiability Modulo Theories (SMT) solvers (such as Z3, CVC5, and Mathematica) were used to symbolically verify that the functions met strict Lyapunov requirements, eliminating risks from numerical rounding errors.
  4. Robustness Synthesis: The team calculated ellipsoidal "robust regions" in the state space. If the engine state remains within these regions, stabilization is guaranteed without triggering a mode switch, even under minor perturbations.

Key Results and Insights

The research successfully verified the stability of the full 21-variable system for individual control modes. Several critical insights emerged:

  • Numerical Vulnerability: The authors discovered that some numerically synthesized Lyapunov functions failed symbolic checks, highlighting the danger of relying solely on standard optimization toolboxes for safety-critical systems.
  • Methodological Robustness: The LMIα approach, which incorporates a convergence speed parameter, proved to be the most resilient synthesis method, maintaining validity even when numerical data was rounded to four significant figures.
  • Solver Performance: While general SMT solvers are powerful, an ad-hoc implementation of Sylvester’s Method (checking determinants of matrix minors) was found to be the most efficient way to validate positive definiteness in these specific control problems.

Conclusion

This work demonstrates that symbolic verification techniques are maturing to handle industrial-sized problems. While stability for individual modes was successfully proven, verifying transitions between modes remains a challenge due to numerical limitations in current LMI solvers. Future research will focus on extending these methods to prove complex temporal properties, such as ensuring stabilization occurs within specific timeframes without exceeding critical safety limits like temperature.

Summary updated

Abstract

The control of complex systems is typically designed describing the physical system with differential equations. The standard approach to their verification employs numerical analysis, which is suitable to prove stability properties, but is susceptible to numerical errors. On the other side, symbolic techniques give precise analysis results but typically do not scale to industrial size problems. In this paper, we consider the control design of an aircraft engine. The engine model is represented by a linear state space model of 18 internal state variables, 4 outputs, and 3 inputs. The control switches between two PI controllers, one for thrust control and another for low-pressure compressor spool speed control, based on the engine state and pilot commands. After reformulating the PI controllers in terms of differential equations, we obtain a hybrid system with 21 state variables and two modes, for which we want to prove with symbolic techniques the robustness of the stable states to perturbation. We achieved the verification with standard methods to synthesize quadratic Lyapunov functions and SMT techniques to synthesize neighborhoods of the stable states for which we have symbolic proof of stability.
BibTeX
@inproceedings{basagiannis2023smt,
  author = {Stylianos Basagiannis and Ludovico Battista and Anna Becchi and Alessandro Cimatti and Georgios Giantamidis and Sergio Mover and Alberto Tacchella and Stefano Tonetta and Vassilios A. Tsachouridis},
  title = {{SMT-Based Stability Verification of an Industrial Switched PI Control Systems}},
  booktitle = {DSN-W},
  year = {2023},
  url = {https://openreview.net/forum?id=hJo56YK1hp},
}
Figure