PCB Milling Anomaly Detection Using Speed and Current Signals

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional milling machines fail to detect anomalies and potential failures in real-time, leading to unexpected downtimes and reject products, as they only recognize errors after they occur, causing delays and inefficiencies in manufacturing processes.

Innovation Solution

A method and device using a trained adaptive algorithm, specifically a neural network, to monitor the rotational speed and electrical supply current of milling machines, detecting anomalies and predicting failures, allowing for proactive maintenance and minimizing operational downtimes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional error detection methods are used, then errors are detected after they occur, but this leads to unexpected downtimes and reject products

Engineering Contradiction:
Improveerror detection accuracyVSAvoiddowntime
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of operational data using a trained learning algorithm to detect anomalies before they develop into actual errors. By continuously monitoring parameters like rotational speed and supply current, the system identifies deviation patterns early, enabling preventive maintenance before machine failure or product rejection occurs

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements continuous feedback loops where operational data is constantly fed into the learning algorithm, which compares current operations against learned normal patterns. When deviations are detected, the system provides immediate feedback through notifications to operators, enabling real-time corrective action before errors manifest

Inventive Principle:
Principle #23Feedback

2Reliability

If real-time monitoring with learning algorithms is implemented, then failures can be predicted early, but this increases device complexity

Engineering Contradiction:
Improvefailure prediction capabilityVSAvoidmonitoring system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The learning algorithm performs self-training by automatically learning normal operational patterns from historical data without requiring manual programming of detection rules. The system self-adjusts to different machine states and workpiece types, reducing the need for complex configuration and maintenance while improving prediction accuracy over time

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The monitoring system is designed to handle multiple machine types and operational parameters through a single unified learning algorithm framework. The system can adapt to different milling machines, workpiece materials, and operational conditions using the same core technology, reducing overall system complexity while maintaining versatility

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3899679B1Method and device for monitoring a milling process
Publication Date: 2022.12.28 SIEMENS AG
  • EP3899679B1 patent drawingFigure 1~3

AI summary

The invention relates to a method for monitoring a milling process of a printed circuit board, having the steps of: (a) detecting (S1) the rotational speed of a milling head (2) of a milling machine (1) and at least one other operating parameter of the milling machine (1) during the milling process, wherein the other operating parameter is an electric supply current for operating the milling machine, and (b) analyzing (S2) the detected rotational speed and the detected operating parameter using a trained adaptive algorithm for detecting anomalies during the milling process.