NC Machining Reinforcement Learning for Cutting Rate Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-product, variable quantity production sites, operators face challenges in optimizing machining programs efficiently due to time constraints, leading to suboptimal cutting rates and increased cycle times, which negatively impact production efficiency.
Innovation Solution
A machine learning device is employed to perform reinforcement learning with a numerical control device, acquiring state and action information to optimize cutting amounts and rates. The device includes a state information acquisition unit, an action information output unit, a reward calculation unit, and a value function update unit to adjust machining parameters based on real-time data and feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If operators manually optimize machining programs based on experience, then machining precision and safety are maintained, but time consumption increases and production efficiency decreases
Solution Approach 1:
The system enables self-service by allowing the numerical control device to automatically optimize machining programs through reinforcement learning. The device acquires state information (cutting amounts, cutting rates), determines optimal actions, and updates its value function autonomously without requiring operator intervention, thus maintaining precision while eliminating time-consuming manual optimization
Solution Approach 2:
The system implements feedback mechanisms where the numerical control device receives determination information (pressure magnitude, waveform shape, time period) from actual machining operations and uses this feedback to update its value function. This closed-loop feedback enables continuous improvement of machining programs while reducing the need for manual time consumption
2Reliability
If cutting rates are reduced excessively to ensure safety in changed machining conditions, then machining safety is maintained, but cycle time increases and production efficiency decreases
Solution Approach 1:
The system dynamically changes cutting parameters (cutting amounts, cutting rates) based on learned value functions that consider safety constraints. By optimizing these parameters through reinforcement learning rather than using fixed conservative values, the system maintains machining safety while avoiding excessive reductions in cutting rates that would harm production efficiency
Solution Approach 2:
The system transitions from static, pre-determined cutting parameters to dynamic optimization where cutting rates and amounts are continuously adjusted based on real-time determination information and the updated value function. This dynamic approach allows the system to maintain safety while maximizing productivity by avoiding overly conservative fixed parameter settings
3Ease of operation
If machining programs are reused across different machines and workpieces, then operator workload is reduced, but optimization quality decreases and production efficiency is compromised
Solution Approach 1:
The system replaces the mechanical process of manual program optimization with an intelligent system based on reinforcement learning. The numerical control device automatically adapts and optimizes machining programs by learning from determination information, substituting operator intellectual labor with an autonomous optimization system that maintains high productivity
Solution Approach 2:
The system enables self-service by allowing the numerical control device to automatically optimize machining programs through reinforcement learning. The device acquires state information (cutting amounts, cutting rates), determines optimal actions, and updates its value function autonomously without requiring operator intervention, thus maintaining precision while eliminating time-consuming manual optimization
Data Source
AI summary
A machine learning device for performing machine learning with respect to a numerical control device which causes a machine tool to operate, and is provided with: a state information acquisition unit that causes the machine tool to perform cutting work, in which a cutting amount and a cutting rate are set, and acquires state information including the cutting amount and cutting rate; an action information output unit that outputs action information; a reward calculation unit that acquires determination information that is information about the strength of pressure applied to a tool at least during cutting work, the shape of the waveform of the pressure applied to the tool, and time it has taken to perform work, and outputs a reward value in reinforcement learning; and a value function update unit that updates a value function on the basis of the reward value, the state information, and the action information.