Ice Detection System Using N-Dimensional Statistical Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional ice detection systems are unable to differentiate between Appendix C and Appendix O icing conditions, leading to insufficient de-icing protocols and potential safety concerns for aircraft and other vehicles.
Innovation Solution
A hierarchical statistical model is used to classify icing conditions by plotting data inputs in n-dimensional space and dividing them into regions representing different icing conditions, allowing for real-time detection and differentiation between dry, Appendix C, and Appendix O conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional ice detection systems are used, then detection capability is provided, but differentiation between Appendix C and Appendix O icing conditions is lost
Solution Approach 1:
The detection system is segmented into multiple independent sensing elements that measure different physical parameters (capacitive coupling, inductive coupling, temperature). Each sensor provides specific information about icing conditions, and their combined data enables differentiation between Appendix C and Appendix O conditions without requiring a single complex sensor system.
Solution Approach 2:
The patent transitions from single-dimensional temperature-based detection to multi-dimensional detection by incorporating multiple sensing modalities (capacitive, inductive, temperature). This dimensional expansion in the measurement space enables the system to distinguish between different icing conditions that temperature alone cannot differentiate.
2Measurement precision
If multiple sensors are used to differentiate icing conditions, then measurement precision improves, but device complexity increases
Solution Approach 1:
The sensing system uses multi-functional sensor elements that can operate in different modes (capacitive coupling mode, inductive coupling mode, temperature measurement mode). Each sensor serves multiple detection purposes, reducing the need for separate dedicated sensors for each measurement type and thereby managing system complexity while maintaining high measurement precision.
3Speed
If real-time on-board analysis is performed, then detection speed improves, but computational requirements and system complexity increase
Solution Approach 1:
The system pre-establishes the relationships between multiple sensor parameters and icing condition types during system design and calibration. By pre-configuring the analysis framework and decision logic, the system minimizes real-time computational complexity while maintaining fast detection response. The complex analysis is partially performed offline, with only final classification requiring real-time processing.
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 accurate and timely detection of icing conditions, reducing safety events and costs by allowing for improved certification and compliance with international standards, enabling vehicles to avoid unsuitable conditions before damage occurs.
Implementation Method 1
monitoring electric field variation in capacitive elements
Implementation Method 2
magneto-restrictive detection e.g. by monitoring the frequency change on magnetic probes
Data Source
Figure 1~3B
Figure 2
Figure 4
AI summary
A method and system for detecting and determining icing conditions comprising: creating a statistical model from test data (1, 2) in which set of data (1, 2) from n data inputs (12) are each plotted as a point in n-dimensional space, and wherein the n-dimensional space is divided into a plurality of regions each representative of a different icing condition (26) according to the data (1, 2) in that region; for a set of current data (1, 2) from n data inputs (12), using the model to classify the icing condition (26) indicated by the current data (1, 2), by: obtaining current data (1, 2) from n data inputs (12); providing the current data (1, 2) as input data (12) to the model; determining the region of the model in which the current data (1, 2) set is located; and identifying the icing condition (26) indicated by the current data (1, 2) set according to the determined region.