Following the data collection process, the pilot and main gas fuel valve models are developed using the same approach but a distinct set of variables. The first step in the development process is the variable selection which assisted by the expert in the field and tabulated in Table \ref{320525}. The inputs are the control signals to the valves for both pilot and main fuel valves, and the outputs are the signals of the actual position of the valve. The training inputs and outputs behaviour are discussed in Section 4. In the second step, the whole training dataset of the selected signals is transferred to the system identification in MATLAB. The training data is simulated to determine the transfer function block as pre-defined in Section \ref{270763}. For this part, the models are trained using the First Order and Second Order transfer function. The output objectives for the system identification are G 1(s) for pilot fuel valve model, G 2(s ) for the main fuel valve models, valve positioner time constant, b pilot for pilot and b main for main gas fuel valve model.