A neural-network method of predicting defect formation on the surface of thin ITO films under mechanical load

A method for determining the number of defects arising under compressive and tensile stress in bended thin transparent conducting coatings on polymer substrates is proposed. This algorithm is based on the use of mathematical methods of artificial neural networks. The network is trained for calculating the average defect density per unit length at the input parameters corresponding to film and substrate sizes, surface resistance of the conducting coating, and bending radius. The application of this method allows one to determine the average defect density with high accuracy.