Physics-Informed Neural Networks for Non-linear System Identification for Power System Dynamics

Stiasny, Jochen; Misyris, George S.; Chatzivasileiadis, Spyros

doi:10.48550/arxiv.2004.04026

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Remark 3.1: The structure of the friction compensator ( 9) is embedded in the PGNN-II structure defined in (17). This can be seen by using a direct feedthrough for the physics-guided layer, i.e., T (x 1 ) = x 1 in ( 16), choosing hidden layer weights zero, i.e., W 3,2 = 0, and physics-guided layer weights W 3,PGL = 0 m fv fv fc fc .…”

Section: Pgnn-ii: Physics-guided Structure Designmentioning

confidence: 99%

“…Therein, the physical guidance consists of using a physics-based loss function (cost to be minimized during training), which penalizes the deviation of NN outputs from compliance with an available physicsbased model. This approach has also been termed as physicsinformed neural networks (PINNs) in [17] (see also the references therein), where it was applied to identification of nonlinear power systems dynamics. Recently, a physics-guided architecture for NNs was also proposed in [18] and applied to a lake temperature estimation problem.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Physics-Guided Neural Networks for Inversion-based Feedforward Control applied to Linear Motors

Bolderman¹,

Lazăr²,

Butler³

2021

Preprint

View full text Add to dashboard Cite

Ever-increasing throughput specifications in semiconductor manufacturing require operating high-precision mechatronics, such as linear motors, at higher accelerations. In turn this creates higher nonlinear parasitic forces that cannot be handled by industrial feedforward controllers. Motivated by this problem, in this paper we develop a general framework for inversion-based feedforward controller design using physicsguided neural networks (PGNNs). In contrast with black-box neural networks, the developed PGNNs embed prior physical knowledge in the input and hidden layers, which results in improved training convergence and learning of underlying physical laws. The PGNN inversion-based feedforward control framework is validated in simulation on an industrial linear motor, for which it achieves a mean average tracking error twenty times smaller than mass-acceleration feedforward in simulation.

show abstract

Section: Pgnn-ii: Physics-guided Structure Designmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%