
On transfer learning of neural networks using bifidelity data for uncertainty propagation
Due to their high degree of expressiveness, neural networks have recentl...
read it

Simulation free reliability analysis: A physicsinformed deep learning based approach
This paper presents a simulation free framework for solving reliability ...
read it

Transfer learning enhanced physics informed neural network for phasefield modeling of fracture
We present a new physics informed neural network (PINN) algorithm for so...
read it

A physicsinformed operator regression framework for extracting datadriven continuum models
The application of deep learning toward discovery of datadriven models ...
read it

Physics Informed Data Driven model for Flood Prediction: Application of Deep Learning in prediction of urban flood development
Flash floods in urban areas occur with increasing frequency. Detecting t...
read it

A deep learning framework for solution and discovery in solid mechanics
We present the application of a class of deep learning, known as Physics...
read it

Fleet Prognosis with Physicsinformed Recurrent Neural Networks
Services and warranties of large fleets of engineering assets is a very ...
read it
Transfer learning based multifidelity physics informed deep neural network
For many systems in science and engineering, the governing differential equation is either not known or known in an approximate sense. Analyses and design of such systems are governed by data collected from the field and/or laboratory experiments. This challenging scenario is further worsened when datacollection is expensive and timeconsuming. To address this issue, this paper presents a novel multifidelity physics informed deep neural network (MFPIDNN). The framework proposed is particularly suitable when the physics of the problem is known in an approximate sense (lowfidelity physics) and only a few highfidelity data are available. MFPIDNN blends physics informed and datadriven deep learning techniques by using the concept of transfer learning. The approximate governing equation is first used to train a lowfidelity physics informed deep neural network. This is followed by transfer learning where the lowfidelity model is updated by using the available highfidelity data. MFPIDNN is able to encode useful information on the physics of the problem from the approximate governing differential equation and hence, provides accurate prediction even in zones with no data. Additionally, no lowfidelity data is required for training this model. Applicability and utility of MFPIDNN are illustrated in solving four benchmark reliability analysis problems. Case studies to illustrate interesting features of the proposed approach are also presented.
READ FULL TEXT
Comments
There are no comments yet.