Material Properties Synthesis and Neural Network Based Property Prediction

Physical Optics Corporation (POC) proposes to develop a novel Neural Network Assisted Material Property Prediction (NNAMPP) software. This Windows-based software engine consists of modules to calculate electrical conductivity, mechanical properties, etc.Modularity will make it simple and easy to upgrade this tool for predicting the properties of materials before they are synthesized. The proposed NNAMPP software engine will consist of a kernel that will oversee the tasks performed for each type ofproperty, a number of property analysis modules, each performing a distinct function, and a neural network select materials to synthesize that would have desired properties. The neural network will take the material properties as input, and work backwardto material composition and synthesis. The NNAMPP engine will benefit the national welfare by synthesizing and predicting material properties of national importance such as flame resistance, toughness, and even superconduction. In Phase I we will developand demonstrate a preliminary feasibility prototype NNAMPP software engine. Using the proposed NNAMPP, new materials can be synthesized for submarine hulls and spacecraft skins, superconductive cables, and spacecraft or submarine windows. Strongermaterials will bring us lighter and stronger automobiles, and apparel and pharmacological synthesis can lead to new classes of medicines.