Machine Thermal Displacement Compensation With Model Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing thermal displacement compensation systems in machines fail to accurately account for individual differences in machines, leading to inaccuracies in machining due to thermal expansions and changes in operating conditions.
Innovation Solution
A thermal displacement compensation system that switches learning models based on individual machine differences, using a condition specifying section, state quantity detecting section, inferential calculation section, compensation executing section, learning model generating section, and learning model storage section to infer and compensate for thermal displacements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a general-purpose machine learning device is introduced to accommodate various machine situations, then adaptability improves, but device complexity and parameter volume increase significantly
Solution Approach 1:
The patent divides the machine learning process into two distinct segments: a learning phase that generates machine-specific learning models, and an inference phase that uses pre-stored models. This segmentation eliminates the need for a complex general-purpose learning device during operation, as the learning functionality is separated into a preliminary setup phase.
Solution Approach 2:
The system performs learning actions in advance by generating machine-specific learning models during a preliminary learning phase. These models are then stored in storage units for later use during the inference phase, eliminating the need for complex real-time learning capabilities during operation.
2Productivity
If machine learning is performed without considering individual machine differences, then processing speed improves, but manufacturing precision deteriorates due to inaccurate thermal displacement compensation
Solution Approach 1:
The patent creates customized learning models tailored to each individual machine's characteristics rather than using a universal model. Each machine receives local optimization through its own learning model that accounts for its specific thermal expansion patterns, ensuring high precision without sacrificing processing efficiency during inference.
Solution Approach 2:
The system changes the parameters by creating machine-specific learning models with customized parameters for each individual machine. This allows the inference process to use optimized parameters specific to each machine's thermal characteristics, achieving high precision compensation while maintaining efficient processing speed.
3Device complexity
If a single learning model is used for all machines, then device complexity is reduced, but measurement precision deteriorates due to inability to capture individual machine characteristics
Solution Approach 1:
The patent creates multiple copies of learning models, each customized for a specific machine. Instead of using a single universal model, the system generates and stores separate learning model copies for each machine, allowing each to capture its individual thermal characteristics while maintaining simple model structures.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables highly efficient machine learning and improved accuracy in thermal displacement compensation by selecting appropriate learning models tailored to individual machine conditions, preventing over-learning and enhancing machining precision.
Implementation Method 1
heat generated by the motor, frictional heat created by a rotation of a bearing, and frictional heat in a contact portion between a ball screw and a ball nut
Implementation Method 2
heat generated by the motor
Implementation Method 3
the feed screw and the spindle expand and a machine position changes due to heat generated by the motor
Implementation Method 4
the use of a coolant also cause a change in column and bed temperatures
Data Source
AI summary
A thermal displacement compensation system detects a state quantity indicating a state of a machine, infers a thermal displacement compensation amount of the machine from the detected state quantity, and performs a thermal displacement compensation of the machine based on the inferred thermal displacement compensation amount of the machine. The thermal displacement compensation system generates a learning model by machine learning that uses a feature quantity, and stores the generated learning model in association with a combination of specified conditions of individual difference of the machine.