Determination of firing parameters of a missilelaunched from cold launch platform by artificialneural network

Within the scope of this thesis, it is aimed to determine the Cold Launch method and its parameters, which is one of the launching systems used for firing missile and ammunition systems from land or over the sea into the air. Hot Launch systems are used classically and mostly for launching missiles and ammunitions from the ground. In the hot launching, the propulsion system of the missile is fired in the canister, allowing it to separate from the launch system and aim at the target. However, this situation causes losses in thrust; contains risks in terms of personnel and platform security. In addition, it requires that the missile and launcher system have a durable and functional design against the pressure and temperature loads that will arise from firing in the canister. As a result of this situation, cost and weight burdens arise. In order to prevent the above-mentioned risks, the Cold Launching Method is recently used in military products of various countries. However, it is difficult to calculate and predict parameters of the cold launching with the classical numerical calculation methods due to the complex dynamics of the event and interactions of many different parameters. Hence, within the scope of the thesis, numerical analysis tools such as MSC Adams and Easy5, and Matlab/Simulink are used simultaneously with the tests to analyze and predict the firing parameters of the cold launch process. Various artificial neural network models are also developed by taking the missile launch angle and length, the ambient temperature and pyrotechnical pressure as input parameters and the missile acceleration as the output. The artificial neural network model with four hidden layers is found to be the most suitable one based on training verification and testing.