sensitivity analysis, reciprocating compressor, gas pulsation, suction manifold
Most simulation models are complex and nonlinear and so global sensitivity analysis is becoming a popular choice to predict the performance characteristics and behavior of the model. Global sensitivity analysis methods are generally variance-based methods that greatly rely on sampling methods and input parameter distribution. These methods don’t rely on linearity or monotonicity of the model and can be applied to a diverse range of problem. Sobol’s method—a variance based global sensitivity analysis method is applied to a nonlinear function to highlight and outline the implementation details of the method. It is used to calculate the sensitivities of the input parameter on the model output. The method is then applied to a reciprocating compressor model to determine the sensitivity of gas pulsation to suction manifold design parameters, namely radius, width and depth. The same method can be readily applied to any compressor simulation model.