Abstract
This article describes a solution to create a hardware-in-the-loop (HIL) simulation system of civil aircraft thrust reverser with Modelica-based simulation platform --MWorks in Windows system. The HIL system uses simulation platform "MWorks" to model and simulate the thrust reverser hydraulic system, and takes hardware --PLC's output signals as the inputs of the simulation. Modeling module, communication module, solving module, animation module and HIL control module are included in the simulation platform, whose key technology and implementation details are specified. The HIL system has been successfully applied to the simulation of ARJ21 aircraft thrust reverser hydraulic system. It can simulate the hydraulic system in normal status, fault status as well as other working conditions to verify control logic and evaluate key performance of the system, thereby helping to reduce the cost of experiments and to optimize the design of the system.
Introduction
Testing aircraft thrust reversers—the systems that help planes slow down after landing—is a high-stakes endeavor. Traditional verification requires physical ground or flight experiments, which are expensive, potentially dangerous, and limited by weather or environmental conditions.
While Hardware-in-the-Loop (HIL) simulation, where real control hardware interacts with a virtual model, is the industry standard for solving this, "true" real-time HIL systems are notoriously expensive and complex. This research explores whether a high-fidelity, multi-domain simulation could be achieved using a standard Windows-based PC and the Modelica language, providing a more accessible way to verify control logic and system performance for the ARJ21 regional jet.
System Architecture and Methods
The researchers built their system around MWorks, a simulation platform based on the Modelica language. Modelica is ideal for this application because it is object-oriented and handles "heterogeneous" systems, allowing for the unified modeling of mechanical, electronic, and hydraulic components.
The architecture integrates several key modules:
- Modeling Module: Utilizes the custom Hydrau_Comac Library, featuring both detailed models for offline analysis and simplified versions optimized for real-time simulation.
- HIL Control and Solving Modules: Manage the simulation execution and mathematical solving.
- Communication Module: Facilitates data exchange with a Siemens S7-300 PLC via an RS232 serial port at 57.6kbps.
- Animation Module: Provides visual feedback of the system's state.
Hybrid Real-Time Strategy
Since Windows is not a "hard" real-time operating system, the team implemented a hybrid strategy using high-resolution multimedia timers (timeSetEvent) and QueryPerformanceCounter to synchronize simulation logic with real-world time. The solver was designed to detect changes in PLC signals as "events," triggering immediate re-calculation to maintain synchronicity.
Implementation and Visualization
To ensure the simulation remains responsive, the team optimized the animation module. Instead of heavy real-time 3D rendering, they pre-rendered animation frames and mapped them to a single variable (actuator deployment length). This allows for smooth visual feedback without significant CPU overhead.
Key Results and Validation
The system successfully simulated the ARJ21 thrust reverser with high accuracy, proving that a standard Windows PC is sufficient for hydraulic systems with a 50Hz requirement (20ms cycle).
- Timing Accuracy: The simulation achieved a stable 20ms sampling period with a timing error of less than 1ms.
- Predictive Accuracy:
- Deployment Time: The simulation predicted 1.04s compared to the experimental 1.08s (a 3.7% error).
- Stowing Time: The simulation predicted 2.66s compared to the experimental 2.68s (a 0.7% error).
The simulated pressure curves closely matched experimental benchmarks from the ARJ21 project, confirming the model's physical fidelity.
Conclusion and Future Work
This HIL system demonstrates a cost-effective alternative to traditional aerospace testing environments. By using Modelica and MWorks on Windows, engineers can verify control logic and evaluate system performance under various conditions, including faults, without the risks of physical testing.
While successful for this hydraulic application, Windows-based HIL has limits and is not suitable for microsecond-level control loops. Future work will focus on enhancing real-time capabilities and developing the ability to generate target code for dedicated high-performance real-time hardware.
Summary updated