Kernel Regression Prognostics Using Sequential Pattern Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Kernel regression models for predictive condition monitoring and prognostics do not effectively utilize time domain information, leading to inaccurate fault detection and prognosis, as they treat data in disconnected time-contemporaneous patterns without considering the order of sensor signals over time.
Innovation Solution
Incorporating time domain information into kernel regression models by using input pattern arrays with temporally-related vectors, allowing the system to generate virtual or inferred estimate values for future time points, and extending vector-to-vector operations to matrix-to-matrix operations to improve accuracy in fault detection and prognosis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If kernel regression models treat data in disconnected time-contemporaneous patterns, then the computational complexity is reduced and the model is simpler to implement, but the time domain information is lost and the fault detection accuracy deteriorates
Solution Approach 1:
The patent transitions from treating data as disconnected vectors to organizing data into sequential pattern arrays that preserve temporal dimensionality. By arranging sensor readings into arrays where temporal relationships are maintained, the model can process data in its natural time-ordered form without excessive computational complexity.
Solution Approach 2:
The patent performs preliminary organization of sensor data into sequential pattern arrays before feeding them to the kernel regression model. This preprocessing step structures the data to preserve time domain information while maintaining computational efficiency, as the temporal relationships are established before the main computation occurs.
2Measurement precision
If time domain information is incorporated into kernel regression models using input pattern arrays, then the fault detection accuracy is improved, but the computational complexity and data processing requirements increase
Solution Approach 1:
The patent segments the input data into sequential pattern arrays, where each array contains a sequence of sensor readings that preserve temporal relationships. This segmentation allows the model to process time-domain information in manageable units without overwhelming computational requirements.
Solution Approach 2:
The patent introduces dynamic temporal relationships into the static kernel regression framework by using sequential pattern arrays. The model dynamically processes time-ordered data while maintaining the computational efficiency of kernel methods, adapting the temporal structure to the prognostics problem.
Data Source
AI summary
A method for determining the future operational condition of an object includes obtaining reference data that indicates the normal operational state of the object, and obtaining input pattern arrays. Each input pattern array has a plurality of input vectors, while each input vector represents a time point and has input values representing a plurality of parameters indicating the current condition of the object. At least one processor generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values, in the form of an estimate matrix, include at least one estimate vector of inferred estimate values, and represents at least one time point that is not represented by the input vectors. The inferred estimate values are used to determine a future condition of the object.