Reducer Maintenance Timing Using Iron Powder Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing maintenance systems for reducers in actuators lack the ability to determine the optimal time for maintenance, leading to potential inefficiencies and increased downtime due to uncertainties in iron powder concentration in grease, which affects the reducer's performance and lifespan.
Innovation Solution
A control system incorporating a machine learning apparatus that correlates rotational speed, torque, and iron powder concentration data to build a concentration estimation model, allowing for the derivation of a recommended maintenance time based on these parameters, and further refining this through an increasing estimation model and modified estimation equations to ensure accurate and timely maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional maintenance scheduling based on rated lifetime and operational parameters is used, then maintenance can be performed periodically, but the timing is not optimal leading to increased downtime and reduced productivity
Solution Approach 1:
The patent transforms the maintenance scheduling approach by changing from fixed periodic intervals to dynamic parameter-based prediction. It uses operational parameters (rotational speed, torque) and degradation indicators (iron powder concentration) to dynamically determine maintenance timing, thereby optimizing productivity and minimizing unnecessary downtime.
Solution Approach 2:
The patent replaces traditional mechanical/time-based maintenance scheduling with an information-processing system. It uses sensors, databases, and machine learning models to predict maintenance needs based on actual operational conditions and degradation patterns, substituting periodic mechanical checks with intelligent prediction.
2Reliability
If maintenance is performed more frequently to ensure reliability, then reducer performance is maintained, but operational time is reduced and productivity decreases
Solution Approach 1:
The patent implements a feedback mechanism where sensors continuously monitor operational parameters and iron powder concentration, feeding this information back to the maintenance prediction system. This closed-loop feedback enables real-time assessment of reducer health, allowing maintenance to be performed only when actually needed, thus maintaining reliability while maximizing operational time.
Solution Approach 2:
The patent performs preliminary assessment of maintenance needs by continuously monitoring degradation indicators before actual failure occurs. The machine learning model predicts future iron powder concentration based on current trends, enabling proactive scheduling of maintenance at the optimal moment before performance degradation becomes critical.
3Measurement precision
If iron powder concentration is monitored to determine maintenance timing, then maintenance can be scheduled more accurately, but measurement and data processing complexity increases
Solution Approach 1:
The patent makes the control system multi-functional by integrating it with sensors, database management, machine learning model execution, and maintenance scheduling capabilities. This universal control system handles multiple functions (monitoring, prediction, decision-making) that would otherwise require separate systems, reducing overall complexity despite the advanced capabilities needed for precise measurement.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A control system for an actuator includes a database and a first model builder. In the database, first information, second information, and third information are stored and correlated with each other. The first information is associated with a rotational speed of a reducer included in the actuator. The second information is associated with a torque acting on the reducer. The third information indicates a concentration of iron powder in grease in the reducer. The first model builder is configured to perform machine learning based on the first information, the second information, and the third information stored in the database to build a concentration estimation model indicating a relationship between the first information, the second information, and the third information.