Predictive Maintenance for Medical Imaging Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical imaging devices, with thousands of complex components, generate vast amounts of log data, leading to challenges in predictive maintenance due to data quality issues, ambiguity in service call data, and the need for efficient failure mode and resolution feature extraction.
Innovation Solution
A predictive maintenance alerting method and device that processes time-stamped machine and service log data to derive built failure mode and resolution features, applying machine-learned analytical models with embedded statistical remaining useful lifetime models to generate maintenance alerts for medical imaging devices, reducing component-level and group-level ambiguities and dependencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine log data and service log data are collected from medical imaging devices, then predictive maintenance capability is improved, but data quality issues and ambiguity in service call data worsen
Solution Approach 1:
The patent introduces an intermediary processing layer between raw log data and predictive maintenance models. This layer includes natural language processing components that act as mediators to clean, standardize, and disambiguate service call data before it reaches the predictive models, thereby resolving the contradiction between utilizing comprehensive data and maintaining data quality
Solution Approach 2:
The patent replaces traditional manual data cleaning and analysis methods with automated natural language processing and machine learning systems. This substitution enables the handling of large volumes of ambiguous service call data at scale, transforming the mechanical process of data quality improvement into an automated computational process
2Reliability
If service call data is used for failure prediction, then maintenance insights are improved, but ambiguity in human decision factors worsens
Solution Approach 1:
The patent substitutes human interpretation of ambiguous service call data with natural language processing and machine learning models. These computational systems automatically extract meaningful patterns from text descriptions of service calls, replacing the subjective and ambiguous human decision-making process with consistent, scalable automated analysis
Solution Approach 2:
The patent transforms unstructured service call text data into structured numerical features that can be processed by predictive models. By changing the parameter representation from ambiguous human language to quantifiable features, the system enables reliable maintenance insights while overcoming the limitations of human decision factor ambiguity
3Measurement precision
If component-level features are extracted for maintenance prediction, then prediction accuracy is improved, but computational complexity and dependencies worsen
Solution Approach 1:
The patent segments the complex task of component-level failure prediction into hierarchical levels: device-level features, component-group-level features, and individual component features. This segmentation allows the system to achieve high prediction accuracy through progressive refinement while managing computational complexity by processing features at appropriate levels of granularity
Solution Approach 2:
The patent introduces a hierarchical dimension to the feature extraction process, organizing components into groups and processing features at multiple levels of abstraction. This dimensional organization reduces computational complexity by avoiding redundant processing while maintaining the ability to generate accurate component-level predictions when needed
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A predictive maintenance alerting device (40) comprises a server computer (42) operatively connected with an electronic network (46) to receive time stamped machine log data (30) and time stamped service log data (32) from a medical imaging device (10), and to transmit maintenance alerts (44) to a service center (12). The predictive maintenance alerting method includes deriving features from the received log data, and applying a set of models (64) of component groups to the derived features to generate the maintenance alerts. Each model may comprise a heterogeneous model including a machine learned analytical model (70) representing the component group with embedded statistical remaining useful lifetime models (72) of the components of the component group. Each component may belong to a single component group. Some derived features may be built failure mode features or built failure resolution features.