Time-Series Signal Control for Fast Anomaly-Based Maintenance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current predictive maintenance strategies for systems like HVAC, robotic assemblies, and vehicles are inefficient due to the need for extensive training data and time-consuming model development, which hinders quick adaptations and optimal maintenance scheduling, especially when dealing with mixed-age systems where some require frequent maintenance while others need it less often.
Innovation Solution
A controller that compares input time-series data with rotated test signals using a sliding three-dimensional window method to produce statistics efficiently, allowing for optimal system operation by computing valid statistical observations and excluding invalid ones, thereby reducing computational complexity and improving maintenance scheduling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If predictive maintenance uses extensive training data and complex model development, then measurement precision and reliability improve, but loss of time and productivity worsen
Solution Approach 1:
The patent creates a reference signal representing normal system operation and compares incoming sensor data against this reference. Instead of training complex models on extensive historical data, the system copies the expected normal behavior pattern and directly compares actual measurements to this copy, enabling quick adaptation without lengthy training periods while maintaining accurate system state determination
Solution Approach 2:
The patent segments the maintenance monitoring approach by separating the reference signal creation phase from the ongoing comparison phase. The reference signal is established once based on normal operation, then the system continuously segments incoming data into comparisons against this reference, allowing rapid maintenance scheduling decisions without retraining models for each new maintenance scenario
2Reliability
If predictive maintenance develops custom models for each system in a group, then reliability improves by accounting for system-specific characteristics, but device complexity and loss of time worsen
Solution Approach 1:
The patent creates a universal reference signal comparison mechanism that can be applied to multiple systems in a group. Instead of designing and training separate custom models for each system, the system establishes a reference signal representing normal operation patterns that can be universally compared across different systems, maintaining reliability by accounting for system-specific characteristics through the flexibility of the reference signal creation process while significantly reducing device complexity
3Ease of operation
If preventive maintenance uses fixed time intervals, then ease of operation improves, but productivity worsens due to unnecessary maintenance on new systems
Solution Approach 1:
The patent transitions from static fixed-interval maintenance scheduling to dynamic condition-based scheduling. The system continuously compares real-time sensor data against the reference signal to dynamically determine when maintenance is actually needed. This allows new systems to operate longer without maintenance when they are performing normally, while still providing simple operation through automated comparisons and clear maintenance triggers when deviations are detected
Data Source
AI summary
A controller for controlling an operation of a system is disclosed. The controller receives an input signal indicative of the operation of the system and rotates a test signal multiple times with different circular shifts to produce different rotations of the test signal forming a matrix data structure with the input signal. The input signal and the test signal are time-series data having values monotonically measured over time. The controller is further configured to apply a sliding three-dimensional (3D) window method to the matrix data structure to produce statistics of the input signal with respect to the rotations of the test signal. The sliding 3D window method iteratively moves window over the matrix data structure to compute a value of the statistics for a segment of the matrix data structure within the window. Furthermore, the controller controls the operation of the system according to the statistics of the input signal.