Abstract:To optimize structures and monitor their health, it is essential to build an accurate dynamic analysis model. However, traditional modeling methods based solely on physical information or data-driven techniques may not suffice for many engineering applications. While physical models can accurately simulate complex equipment, they may also incur high computational time. On the other hand, data-driven models may improve computational efficiency but are subject to significant deviations due to the influence of training data. To address these challenges, the Physics-Informed Neural Network (PINN) has gained popularity for imposing physical constraints during the training process, leading to better generalization capabilities with fewer data samples. This paper proposes a physics-informed hybrid modeling (PIHM) approach that combines a reduced-order model, kernel functions, and dynamic equations to predict dynamic output with limited training data and physical information. The method integrates prior physics information into function approximation by incorporating the reduced dynamic equation into a surrogate modeling framework. The loss function considers inertial and damping effects, ensuring physical plausibility. Unlike traditional PINN applications, the proposed modeling method is more explainable, as the trained model can be expressed in function form with engineering interpretation. The approach is verified with a real-world engineering example (telehandler boom) under complex load conditions, demonstrating accuracy, efficiency, and physical plausibility. Overall, the proposed method offers promising capabilities in solving problems where high-fidelity simulation is challenging.

Application of Physics-Informed Machine Learning for Excavator Working Resistance Modeling

Modeling and Control of General Hydraulic Excavator for Human-in-the-loop Automation

Performance Evaluation of TBM Using an Improved Load Prediction Model

Toward an Online Monitoring of Structural Performance Based on Physics-Informed Hybrid Modeling Method

Enhanced Wear Prediction of Tunnel Boring Machine Disc Cutters for Accurate Remaining Useful Life Estimation Using a Hybrid Model

In Situ Soil Property Estimation for Autonomous Earthmoving Using Physics-Infused Neural Networks

Research on calculation and measurement method of hydraulic excavator load force-time process

Predictive Modeling of Loader's Working Resistance Measurement Based on Multi-Sourced Parameter Data

Predictions of Deep Excavation Responses Considering Model Uncertainty: Integrating BiLSTM Neural Networks with Bayesian Updating

Physics-Informed Machine Learning for Seismic Response Prediction OF Nonlinear Steel Moment Resisting Frame Structures

Research on Predictive Modeling Method of Loader Working Resistance in a Sensor-Less Environment

Predicting the Strut Forces of the Steel Supporting Structure of Deep Excavation Considering Various Factors by Machine Learning Methods

Physics-informed machine learning and its structural integrity applications: state of the art

Auto machine learning-based modelling and prediction of excavation-induced tunnel displacement

When Physics Meets Machine Learning: A Survey of Physics-Informed Machine Learning

Physics-Informed Deep Learning for Tool Wear Monitoring

Machine learning-based soil–structure interaction analysis of laterally loaded piles through physics-informed neural networks

Physics-guided deep learning for driving force estimation in synchronous tunnel boring machines under missing cylinders

Physics-informed machine learning for reliability and systems safety applications: State of the art and challenges

Physics-informed Deep Learning Method for Predicting Tunnelling-Induced Ground Deformations

Physics-Informed Extreme Learning Machine framework for solving linear elasticity mechanics problems