Search

The utilization of data in physics-informed neural network (PINN) may be considered as a necessity as it allows the simulation of more complex cases with a significantly lower computational cost. However, doing so would also make it prone to any issue with the data quality, including its noise. This study would primarily focus on developing a special loss function in the PINN to allow an effective utilization of noisy data. However, a study regarding the data location and amount was also conducted in order to allow a better data utilization in PINN. This study was conducted on a lid-driven cavity flow at Re = 200, 1000, and 5000 with a dataset of less than 100 velocity data and a maximum noise of 10% of the maximum velocity. The results show that by ensuring the data are distributed in a certain configuration, it has zero noise, and by using as much data as possible, the computational cost of PINN can be significantly reduced compared to without using any data at all. For Re = 200, it is 7.4 faster by using data, and this speedup is potentially higher for higher Re cases. For the noise in particular, it does not only make the PINN more inaccurate but also necessitate the usage of more data as this is the only way to make it more accurate. This issue though is capable to be solved with our new method, which only uses the data as an approximate solution, and the governing equation would figure out the details. This method was also shown to be capable to improve the PINN accuracy with the potential to almost completely eliminating the noise effect. [ABSTRACT FROM AUTHOR]

Although the grid convergence index is a widely used for the estimation of discretization error in computational fluid dynamics, it still has some problems. These problems are mainly rooted in the usage of the order of a convergence variable within the model which is a fundamental variable that the model is built upon. To improve the model, a new perspective must be taken. By analyzing the behavior of the gradient within simulation data, a gradient-based model was created. The performance of this model is tested on its accuracy, precision, and how it will affect a computational time of a simulation. The testing is conducted on a dataset of 36 simulated variables, simulated using the method of manufactured solutions, with an average of 26.5 meshes/case. The result shows the new gradient based method is more accurate and more precise then the grid convergence index(GCI). This allows for the usage of a coarser mesh for its analysis, thus it has the potential to reduce the overall computational by at least by 25% and also makes the discretization error analysis more available for general usage. [ABSTRACT FROM AUTHOR]