Scientific Computing is an exciting and growing area that provides an important link between Computer Science and the Engineering and Physical Sciences. Today, computer graphics and geometric modeling are used routinely in science, engineering, business; and entertainment. In this thesis we develop object-oriented techniques and software for computing and visualizing implicitly defined manifolds ("surfaces") that arise a wide range of applications. The software differs from existing software for computing such manifolds in its software architecture. Furthermore, its algorithms are based on numerical continuation methods, rather than on subdivision techniques, which allows its practical application to the computation of two-dimensional manifolds in high-dimensional Euclidean spaces. The overall software provides a graphical user interface, algorithms for computing two-dimensional manifolds in higher-dimensional spaces, and graphics routines to visualize the manifolds.
Youssef Omran Gdura(6-2001)

—Predicting the groundwater level has recently become very important research topic especially with the rise of population density and consequently increasing the water demand. This paper uses the Random Forest and linear regression neural network models to predict the groundwater level of Wadi-Alshaty district in the South West part of Libya. The results are compared with that obtained using the hydrologic long-term forecasting graphical approach. One of the most important findings of this study is the effectiveness of the neural network models to investigate the fluctuation of the groundwater levels over time (20 years)
Amna Elhawil, Alarabi Naji, Malak Nuesry, Almabruk Sanossi(12-2021)

Abstract The dynamic model of the heat process trainer PT326 is obtained experimentally by using the Ziegler–Nichols approximation and the system identification MATLAB toolbox. The interfacing data acquisition card is developed in two different designs the first design by using analog to digital, digital to analog IC's (ADC and DAC) and the parallel port, and the other design by using microcontroller and serial port. These two designs were tested in the department and give good results. The velocity PID controller is implemented through the developed cards in real time control system. An adapted neural network algorithm using the backpropagation approximation is tested on real time in the heat process trainer PT326. This research gives a good background for any control and computer engineer in the area of data acquisition and real time control system. The simulation results are the same as the online results of the real time Heat process control system.
نوري احمد عقاب (2010)