Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hybrid Model-Data Fault Diagnosis for Wafer Handler Robots: Tilt and Broken Belt Cases

Tim van Esch, Farhad Ghanipoor, Carlos Murguia, Nathan van de Wouw

TL;DR

This work tackles fault detection, isolation, and estimation for wafer handler robots by blending model-based fault estimation with a data-driven classifier. Faults are treated as additive signals within a state-space framework, enabling an ultra-local fault representation and an ISS-stable augmented observer whose LMIs yield robust $H_\infty$ and $H_2$ performance. The data-driven stage uses SVMs trained on fault estimates to achieve high fault isolation accuracy, and experiments show the hybrid FDIE outperforms pure data-driven approaches, particularly for smaller faults like tilt. The approach has practical impact by enabling more reliable, scheduled maintenance in semiconductor manufacturing and reducing downtime.

Abstract

This work proposes a hybrid model- and data-based scheme for fault detection, isolation, and estimation (FDIE) for a class of wafer handler (WH) robots. The proposed hybrid scheme consists of: 1) a linear filter that simultaneously estimates system states and fault-induced signals from sensing and actuation data; and 2) a data-driven classifier, in the form of a support vector machine (SVM), that detects and isolates the fault type using estimates generated by the filter. We demonstrate the effectiveness of the scheme for two critical fault types for WH robots used in the semiconductor industry: broken-belt in the lower arm of the WH robot (an abrupt fault) and tilt in the robot arms (an incipient fault). We derive explicit models of the robot motion dynamics induced by these faults and test the diagnostics scheme in a realistic simulation-based case study. These case study results demonstrate that the proposed hybrid FDIE scheme achieves superior performance compared to purely data-driven methods.

Hybrid Model-Data Fault Diagnosis for Wafer Handler Robots: Tilt and Broken Belt Cases

TL;DR

This work tackles fault detection, isolation, and estimation for wafer handler robots by blending model-based fault estimation with a data-driven classifier. Faults are treated as additive signals within a state-space framework, enabling an ultra-local fault representation and an ISS-stable augmented observer whose LMIs yield robust and performance. The data-driven stage uses SVMs trained on fault estimates to achieve high fault isolation accuracy, and experiments show the hybrid FDIE outperforms pure data-driven approaches, particularly for smaller faults like tilt. The approach has practical impact by enabling more reliable, scheduled maintenance in semiconductor manufacturing and reducing downtime.

Abstract

This work proposes a hybrid model- and data-based scheme for fault detection, isolation, and estimation (FDIE) for a class of wafer handler (WH) robots. The proposed hybrid scheme consists of: 1) a linear filter that simultaneously estimates system states and fault-induced signals from sensing and actuation data; and 2) a data-driven classifier, in the form of a support vector machine (SVM), that detects and isolates the fault type using estimates generated by the filter. We demonstrate the effectiveness of the scheme for two critical fault types for WH robots used in the semiconductor industry: broken-belt in the lower arm of the WH robot (an abrupt fault) and tilt in the robot arms (an incipient fault). We derive explicit models of the robot motion dynamics induced by these faults and test the diagnostics scheme in a realistic simulation-based case study. These case study results demonstrate that the proposed hybrid FDIE scheme achieves superior performance compared to purely data-driven methods.

Paper Structure

This paper contains 30 sections, 7 theorems, 65 equations, 12 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Modeling of the Robot Dynamics
  3. Planar Dynamics
  4. Power Transmission Dynamics
  5. Total Planar Dynamics
  6. State-Space Representation of the Healthy Dynamics
  7. Fault Scenarios and Associated Fault Models
  8. State-Space Formulation of the Healthy Dynamics with Fault Signals
  9. Model-based Fault Estimation
  10. Ultra Local Lumped Fault Representation
  11. Augmented Dynamics
  12. Fault Estimator
  13. Fault Estimator Error Dynamics
  14. Fault Estimator Synthesis
  15. Simulation-Based Case Study Results on Fault Estimation
  16. ...and 15 more sections

Key Result

Theorem 1

(Fault Estimator Synthesis) Consider the augmented dynamics eq:augmented, the fault estimator filter as defined in eq:observer, the corresponding estimation error dynamics eq:error_system, and the transfer functions eq:tfs. To design parameters of the optimal mixed $H_2/H_{\infty}$ fault estimators, given $\epsilon, \bar{\gamma}_{max} > 0$, $\bar{C}_a$ in eq:c_bar_a, and the remaining matrices as

Figures (12)

  • Figure 1: An example of SCARA manipulator zhang2007calibration.
  • Figure 2: Schematic representation of a SCARA manipulator.
  • Figure 3: Planar schematic overview of the WH robot in the XY-plane.
  • Figure 4: Schematic overview of the power transmission model.
  • Figure 5: Schematic representation of tilt in the WH robot.
  • ...and 7 more figures

Theorems & Definitions (7)

  • Theorem 1
  • Proposition 1
  • Proposition 2
  • Proposition 3
  • Lemma 1
  • Lemma 2
  • Lemma 3