The thesis focuses on various techniques to find an alternate approximation\nmethod that could be universally used for a wide range of CFD problems but with\nlow computational cost and low runtime. Various techniques have been explored\nwithin the field of machine learning to gauge the utility in fulfilling the\ncore ambition. Steady advection diffusion problem has been used as the test\ncase to understand the level of complexity up to which a method can provide\nsolution. Ultimately, the focus stays over physics informed machine learning\ntechniques where solving differential equations is possible without any\ntraining with computed data. The prevalent methods by I.E. Lagaris et.al. and\nM. Raissi et.al are explored thoroughly. The prevalent methods cannot solve\nadvection dominant problems. A physics informed method, called as Distributed\nPhysics Informed Neural Network (DPINN), is proposed to solve advection\ndominant problems. It increases the lexibility and capability of older methods\nby splitting the domain and introducing other physics-based constraints as mean\nsquared loss terms. Various experiments are done to explore the end to end\npossibilities with the method. Parametric study is also done to understand the\nbehavior of the method to different tunable parameters. The method is tested\nover steady advection-diffusion problems and unsteady square pulse problems.\nVery accurate results are recorded. Extreme learning machine (ELM) is a very\nfast neural network algorithm at the cost of tunable parameters. The ELM based\nvariant of the proposed model is tested over the advection-diffusion problem.\nELM makes the complex optimization simpler and Since the method is\nnon-iterative, the solution is recorded in a single shot. The ELM based variant\nseems to work better than the simple DPINN method. Simultaneously scope for\nvarious development in future are hinted throughout the thesis.\n