To enable the de-carbonisation of the energy sector, solar-driven renewable energytechnologies hold a key spot in the overall technological portfolio. Solar thermal,which uses solar heat to generate heating for domestic and industrial purposes has adecade-long history of implementation. A parabolic trough collector (PTC) is onetechnology that can efficiently generate heat for industries. Expanding the PTCperformance to system performance is critical to quantify the effective savings tothe customer. However, analyzing the system performance is more challenging dueto the effect of additional parameters such as heating load variation, hourly variationin irradiation. Thus, system performance prediction requires detailed information onthe shifting heat demand in line with the targeted share of renewable heat and istypically done using dynamic simulations which may require high computationaltime.This thesis aims to use an optimal machine learning (ML) algorithm to predict theperformance of a PTC-based system for an industrial setting. The work is intendedto compare the techno-economic feasibility results obtained from the existingsimulation tool and compare it with results obtained after using the ML algorithmsfor the boundary conditions of the system to represent those in industries, for Indianclimatic conditions.An artificial neural network (ANN) model is developed to achieve the aims of thisthesis. An industrial PTC collector is used as a reference, and simulations areperformed using a system simulator provided by the manufacturer. The simulationdata is structured to import into the model and is then used for testing, training, andvalidation. The losses for the model were 9.902 and 27.362 for MAE and RMSErespectively. The model shows a good accuracy in predicting the systemperformance and has a significant potential to reduce the time needed for currentmethods of simulations.