Recipe Independent Fault Diagnosis Mapping

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional fault diagnosis systems in manufacturing are limited by their reliance on recipe-specific and metric-specific fault classes and signatures, making them ineffective for use across different processes and machines, leading to inefficiencies in fault detection and diagnosis.

Innovation Solution

The development of a method and apparatus for diagnosing faults using metrics-independent and recipe-independent fault signatures, allowing fault signatures generated with one metric to be applied to faults detected by another metric, and fault classes to be shared across multiple recipes and processes, enabling more comprehensive and flexible fault analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional model-specific fault classes and signatures are used, then fault diagnosis is simple for a specific model metric, but the fault classes and signatures cannot be used with other model metrics or recipes

Engineering Contradiction:

Improvefault class reusabilityVSAvoidfault diagnosis system complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a mapping relationship between fault classes of different model metrics, enabling a fault class generated from one model metric (e.g., T2) to be applied to faults detected by other model metrics (e.g., SPE, Q). This universalization allows fault classes to serve multiple functions across different metrics and recipes, resolving the contradiction between reusability and system complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Adaptability or versatility

If recipe-specific fault classes are generated, then fault diagnosis accuracy is high for that specific recipe, but the fault classes cannot be applied to other recipes or processes

Engineering Contradiction:

Improvefault class applicabilityVSAvoidfault diagnosis accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent creates a mapping relationship that allows fault classes generated from one recipe to be copied and applied to other recipes. The mapping preserves the essential fault characteristics while adapting them to different recipe contexts, enabling broad applicability without completely sacrificing diagnostic accuracy for specific recipes.

Inventive Principle:

Principle #26Copying

3Reliability

If multiple model metrics are used for fault detection, then fault detection comprehensiveness is improved, but fault diagnosis becomes more complex due to metric-specific fault classes

Engineering Contradiction:

Improvefault detection comprehensivenessVSAvoidfault diagnosis complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent merges the fault class information from multiple model metrics by creating a mapping relationship that allows fault classes to be shared across metrics. This combining approach maintains the comprehensive fault detection capability of multiple metrics while simplifying the diagnosis process by using unified fault classes instead of separate metric-specific classes.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS8010321B2Metrics independent and recipe independent fault classes

Publication Date: 2011.08.30 APPLIED MATERIALS INC

AI summary

A method and apparatus for diagnosing faults. Process data is analyzed using a first metric to identify a fault. The process data was obtained from a manufacturing machine running a first recipe. A fault signature that matches the fault is identified. The identified fault signature was generated using a second metric and/or a second recipe. At least one fault class that is associated with the fault signature is identified.