How is AI Optimizing Helicopter Engine Efficiency?

By Muhammad OsamaReviewed by Susha Cheriyedath, M.Sc.Nov 25 2024

A recent study published in Energies introduced a novel method for optimizing the energy characteristics of helicopter turboshaft engines (TEs). The research leverages neural networks to regulate free turbine rotor speed and fuel consumption, offering improved efficiency, reduced fuel usage, and enhanced adaptability across diverse flight conditions.

The goal was to enhance engine performance, reliability, and adaptability across different flight modes by developing a mathematical model that establishes a relationship between rotor speed and engine output power. The researchers designed a fuel controller using a neuro-fuzzy network to analyze input data and make real-time adjustments for improved operational efficiency.

Helicopter Turboshaft Engine Control

Helicopter TEs are crucial in ensuring efficient rotor operation and providing the necessary lift and maneuverability for flight. The automatic control system (ACS) of these engines focuses on maintaining stable main rotor speed, which heavily depends on precise control of power characteristics, particularly fuel consumption and free turbine rotor speed.

Traditional methods, such as proportional-integral-derivative (PID) controllers, face significant challenges in handling rapidly changing external loads and complex dynamic conditions. While adaptive control systems and neural network-based models have shown potential to improve the prediction and adjustment of engine parameters, they often encounter limitations, including slow data processing, high computational demands, and the necessity for large, high-quality training datasets. Additionally, existing research frequently lacks dynamic regulation using predictive neural network models, particularly for real-time fuel optimization and adaptation to sudden external changes.

A Novel Neuro-Fuzzy Control Approach for Helicopter TEs

This study introduced an innovative method for managing helicopter TE power characteristics by regulating free turbine rotor speed and fuel consumption. The researchers developed a comprehensive mathematical model that interconnected key parameters, including main rotor speed, free turbine rotor speed, gas-generator rotor speed, and fuel consumption. This model effectively captured the dynamic relationships among these factors, enabling precise control over engine power.

At the core of the study was a neuro-fuzzy network used to design a fuel consumption controller. This advanced network processed input data such as the desired and current rotor speeds, their derivatives, and feedback from the gas-generator rotor speed derivative. By utilizing fuzzy logic, the system converted precise numerical inputs into fuzzy values, enabling decision-making based on established fuzzy rules.

To further enhance the controller's adaptability, the researchers incorporated a long-short-term memory (LSTM) recurrent structure within the network. This deep neural network component allowed the system to learn from historical data and adjust controller settings in real-time, ensuring optimal performance under varying conditions. The training and testing datasets were carefully prepared using time sampling and adaptive quantization techniques, which improved signal quality and maintained data consistency.

Performance Evaluation and Key Findings

The study compared the performance of the developed controller with that of a traditional linear PID controller using flight data from a Mi-8MTV helicopter equipped with a TV3-117 turboshaft engine. The outcomes showed that the neuro-fuzzy controller significantly enhanced control performance, reducing the transient fuel consumption process time by 8.92 %. Specifically, it improved accuracy by 18.28 % and increased the F1 score, which measures the harmonic mean of precision and recall, by 21.32 %. The overshoot remained below 1.302 %, and the control accuracy of engine power dynamics reached 99.3 %.

These results highlighted substantial improvements in transient processes, power distribution, accuracy metrics, and loss metrics, underscoring the superiority of the neuro-fuzzy controller. Furthermore, cluster analysis of the training and testing datasets confirmed that the data used for model training and validation were homogeneous and representative.

The authors emphasized the importance of adaptive control in addressing the challenges posed by sudden external load changes and dynamic operating conditions. By incorporating feedback mechanisms and employing advanced data processing techniques, the neuro-fuzzy network demonstrated superior responsiveness compared to conventional PID controllers. This adaptability is important for ensuring the reliability and safety of helicopter operations, particularly in unpredictable environments.

Applications

This research has the potential to make a big impact on the aerospace industry, especially in improving the performance and reliability of helicopter engines. By using neuro-fuzzy networks in engine control systems, the study offers practical benefits like better fuel efficiency, lower operating costs, and enhanced flight safety. The method could help develop smarter, more efficient control systems for helicopter turboshaft engines, making them more adaptable to different flight conditions.

One of the standout advantages is the system’s ability to optimize fuel use in real-time and adjust quickly to sudden changes in external loads or operating conditions. These improvements in accuracy and faster response times not only boost engine performance but also lead to significant fuel savings, making operations more cost-effective and sustainable. This approach could also be adapted for other aircraft engines, opening the door to better engine management across the aviation industry.

Conclusion and Future Directions

This study demonstrated the effectiveness of innovative methodologies for controlling the energy characteristics of helicopter turboshaft engines by regulating free turbine rotor speed and fuel consumption. The approach shows great promise for transforming the aerospace industry by enhancing the reliability, efficiency, and safety of helicopter operations.

However, the authors noted some challenges. The system’s reliance on high-quality, extensive datasets for training neural networks could limit its performance in extreme or unforeseen conditions. Additionally, the algorithm’s complexity might make it difficult to interpret the control processes, posing challenges in understanding cause-and-effect relationships.

Future research should focus on validating the method in real-world applications to confirm its practicality and effectiveness under operational conditions. Refining modern control strategies and improving comparison techniques could further enhance the accuracy and reliability of engine performance monitoring. Investigating the adaptive component’s ability to reduce oscillations and improve stability under varying loads will also be key to optimizing control performance. These advancements will help ensure the method’s readiness for broader adoption in the aviation industry.

Journal Reference

Vladov, S.; & et al. Method for Helicopter Turboshaft Engines Controlling Energy Characteristics Through Regulating Free Turbine Rotor Speed and Fuel Consumption Based on Neural Networks. Energies 2024, 17, 5755. DOI: 10.3390/en17225755, https://www.mdpi.com/1996-1073/17/22/5755

Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.