Abstract:
Noise generated by mechanical systems on a warship can pose serious concerns for the ship and crew members. Our project "Identification and Reduction of Noise by Mechanical Systems" focuses on analysis and mitigation of any problem that can appear in mechanical systems. Excessive noise in warships can be a serious issue, it can mean that warship stealth is compro mised, which can increase detectability through sonar by enemies ship, which usually results in high costs. In this study, we have employed advanced Machine learning techniques and signal processing techniques to process noise in warships. Our system helps naval engineers improve warship noise control through data driven, AI based solutions. Our solution is implemented as a web application that allows crew members and engineers to detect unusual noises in real time using the ship’s internal noise. Unlike traditional systems that require complex hardware and fixed installations, this approach is lightweight, portable, and accessible on everyday devices. We used CRNN, which is a Convolutional Recurrent Neural Networks to identify and classify noises along with lightweight Convolutional Encoder Decoder network for noise reduction. By identifying these noises, we can improve situational awareness and also maintenance. In addition to detection, the system keeps the logs of all the identified noise so that the user can study all the noise patterns across time and detect common issues. The features of noise detection, classification and logs storage are important in providing improved decision making to the naval engineers and also minimizing the operational risks and maintenance expenses. With focus on portability as much as it can work on web in addition to Artificial intelligence, the system shows how AI and digital signal processing can increase the reliability and efficiency of naval operations. This work can be extended in future to incorporate shipboard equipment, increase the scope of the sound classification database and be implemented in other applications like industrial equipment and submarines.
