Elevator Health Scoring for Predictive Maintenance Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing elevator systems lack effective predictive maintenance solutions to anticipate and address potential issues based on operational data, leading to inefficiencies and potential downtime.
Innovation Solution
A predictive maintenance advisory system utilizing machine learning models trained with elevator data and sensor alerts to determine health scores and identify specific components requiring maintenance, communicated to technicians for timely intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional reactive maintenance is used for elevators, then maintenance is performed only after failures occur, but this leads to increased downtime and operational disruptions
Solution Approach 1:
The system performs preliminary actions by continuously monitoring elevator component data and predicting potential failures before they occur. The machine learning model analyzes historical and real-time data to identify patterns indicating future failures, enabling maintenance to be scheduled in advance rather than reacting to actual breakdowns, thus reducing downtime and improving reliability
Solution Approach 2:
The system implements feedback by continuously collecting data from elevator sensors, comparing actual performance against predicted patterns, and adjusting maintenance schedules based on real-time health assessments. The feedback loop enables dynamic updates to maintenance timing, ensuring interventions occur optimally before failures manifest, thereby minimizing operational disruption
2Measurement precision
If comprehensive sensor monitoring is implemented across all elevator components, then more data is available for predictive analysis, but system complexity and data processing requirements increase
Solution Approach 1:
The system extracts only the most critical features and data points from the comprehensive sensor data using machine learning dimensionality reduction techniques. By identifying and focusing on key predictive indicators rather than processing all raw sensor data, the system maintains high measurement precision while reducing computational complexity and data processing requirements
Solution Approach 2:
The machine learning model serves as an intermediary between the complex sensor network and the maintenance decision-making process. It processes and interprets multi-sensor data, transforming complex raw measurements into simplified health scores and failure predictions, thereby enabling precise component monitoring without requiring direct complex system management
3Measurement precision
If machine learning models are trained with extensive historical elevator data, then prediction accuracy improves, but training time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-processing and organizing historical data into structured formats with engineered features before training begins. Data cleaning, normalization, and feature extraction are completed in advance, reducing the actual model training time while maintaining the benefits of extensive historical data for improved prediction accuracy
Solution Approach 2:
The system uses partial action by implementing incremental or online learning approaches where the model is trained on representative subsets of data or updated periodically with new data. This allows the system to achieve sufficient prediction accuracy without requiring exhaustive training on entire historical datasets, reducing training time while maintaining effective predictive capability
Data Source
AI summary
Embodiments of the present disclosure are directed to a computing device comprising processors configured to receive training data comprising elevator data associated with a plurality of elevator assemblies indicating a series of actions performed by an elevator car, receive alert data associated with the elevator assemblies indicating alerts generated by sensors, determine health scores for the elevator assemblies based on the alert data, label the training data based on the health scores to generate first labeled training data, train a first machine learning model, using the first labeled training data and supervised learning techniques, to predict a health score for an elevator based on input elevator data, receive first elevator data associated with a first elevator, input the first elevator data into the trained first machine learning model, and determine a health score associated with the first elevator based on an output of the first machine learning model.