by Quantum computing and machine learning are emerging as powerful tools in various fields of study, and researchers from Shanghai Jiao Tong University have recently investigated their potential application in fluid dynamics. This research aims to improve the accuracy and efficiency of solving complex problems related to the flow of air over airfoils, such as airplane wings, and the occurrence of flow separation.
Traditionally, engineers have relied on classical computing methods to analyze airflow and detect flow separation in order to prevent aircraft stalls. However, the team of researchers at Shanghai Jiao Tong University, led by Xi-Jun Yuan and Zi-Qiao Chen, explored the use of quantum computing along with machine learning as a more accurate alternative.
The researchers utilized a quantum support vector machine, a type of supervised machine learning algorithm, instead of a classical support vector machine. The results demonstrated a significant improvement in accuracy. The classification accuracy for detecting flow separation increased from 81.8% to 90.9%, while the accuracy for classifying the angle of attack increased from 67.0% to 79.0%.
These findings suggest that quantum computing methods have the potential to outperform classical computing methods in fluid dynamics problems, particularly when dealing with large datasets. Apart from aircraft design, quantum support vector machines could also find applications in underwater navigation and target tracking.
The research team conducted two classification tasks to evaluate the effectiveness of quantum computing in fluid dynamics. The first task involved binary classification, determining whether flow separation had occurred or not. A small dataset consisting of pressure sensor data from an airfoil in a wind tunnel was used for this task. The dataset consisted of 45 points, with 27 cases of no flow separation and 18 cases of flow separation. The data was divided into training and testing sets with 34 and 11 points, respectively.
The second, more complex task aimed to classify the angle of attack of the airfoil after flow separation into four classes. The problem was divided into four separate binary classification problems. The dataset for this task was created through simulations and consisted of 63 points. The training and testing sets were divided into 43 and 20 points, respectively. The training and testing process was repeated 10 times with different combinations of data, and the average accuracy was calculated.
The researchers chose a quantum-annealing-based supervised machine learning algorithm called a support vector machine for classification. They used the D-Wave Advantage 4.1 system, a physical quantum computing device, for quantum annealing. Quantum annealing-based support vector machines have demonstrated superior performance compared to classical counterparts. Classical methods are structurally simple and robust but often suffer from high storage and computation costs, limiting scalability.
Quantum annealing is an optimization process that leverages quantum fluctuations to search for a global minimum among a set of solutions. Unlike other optimization algorithms that can get stuck at local minima, quantum annealing generates multiple good candidates for the global minimum, resulting in more accurate and reliable results.
In conclusion, the study conducted by researchers at Shanghai Jiao Tong University highlights the potential of quantum computing and machine learning in improving the accuracy and efficiency of solving complex fluid dynamics problems. The results indicate that quantum support vector machines have the ability to surpass classical methods in terms of accuracy and scalability. Future applications may extend beyond aircraft design to fields such as underwater navigation and target tracking.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it
