Dissipative Latent Residual Physics-Informed Neural Networks for Modeling and Identification of Electromechanical Systems

TL;DR

This paper introduces DiLaR-PINN, a physics-informed neural network that models dissipative effects in electromechanical systems while ensuring physically consistent energy behavior, validated on a helicopter system.

Contribution

The paper proposes a dissipative latent residual PINN with a skew-dissipative structure for physically consistent modeling of dissipative effects, improving stability and accuracy.

Findings

DiLaR-PINN more accurately captures dissipative effects.

It achieves superior long-horizon extrapolation performance.

It demonstrates stable training with partial state measurements.

Abstract

Accurate dynamical modeling is essential for simulation and control of embodied systems, yet first-principles models of electromechanical systems often fail to capture complex dissipative effects such as joint friction, stray losses, and structural damping. While residual-learning physics-informed neural networks (PINNs) can effectively augment imperfect first-principles models with data-driven components, the residual terms are typically implemented as unconstrained multilayer perceptrons (MLPs), which may inadvertently inject artificial energy into the system. To more faithfully model the dissipative dynamics, we propose DiLaR-PINN, a dissipative latent residual PINN designed to learn unmodeled dissipative effects in a physically consistent manner. Structurally, the residual network operates only on unmeasurable (latent) state components and is parameterized in a skew-dissipative form that guarantees non-increasing energy for any choice of network parameters. To enable stable and data-efficient training under partial measurability of the state, we further develop a recurrent rollout scheme with a curriculum-based sequence length extension strategy. We validate DiLaR-PINN on a real-world helicopter system and compare it against four baselines: a pure physical model (without a residual network), an unstructured residual MLP, a DiLaR variant with a soft dissipativity constraint, and a black-box LSTM. The results demonstrate that DiLaR-PINN more accurately captures dissipative effects and achieves superior long-horizon extrapolation performance.