Machine Prediction Modeling for Cyclic Ambient Conditions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction systems for machines under cyclically variable ambient conditions suffer from reduced precision due to changing environmental factors, and existing machine-learning algorithms struggle to adapt quickly and accurately.
Innovation Solution
A method utilizing a Generalized Additive Model (GAM) to generate prediction systems by determining correlations between ambient and performance parameters, allowing for rapid generation of precise models even with limited training data, and enabling extrapolation and interpolation across different time intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional machine-learning algorithms are used for prediction, then the system can handle complex patterns, but the learning rate is impeded by changing ambient conditions and deployment time is extended
Solution Approach 1:
The patent segments the prediction task by separating ambient parameters from performance parameters in the training data. This segmentation allows the model to learn distinct patterns for each parameter type, improving adaptation to changing ambient conditions while maintaining faster training speeds through modular learning
Solution Approach 2:
The patent implements dynamic adaptation by continuously updating the prediction model with new training data that reflects current ambient conditions. The system dynamically adjusts its parameters and retraining schedule based on environmental changes, enabling rapid adaptation without full model retraining
2Measurement precision
If comprehensive training data covering all ambient conditions is collected, then prediction precision improves, but data collection time and computational requirements increase
Solution Approach 1:
The patent applies preliminary action by pre-processing training data to separate ambient and performance parameters before model training. This preliminary organization of data enables more efficient training computations and reduces the time required to process comprehensive datasets
Solution Approach 2:
The patent extracts and isolates ambient parameters from the comprehensive training data, treating them as separate input features. This extraction allows the model to focus on learning the relationship between ambient conditions and machine performance without being overwhelmed by the full complexity of all training parameters simultaneously
Data Source
AI summary
A method for generating a prediction system for a machine, the machine being subjected to cyclically variable ambient conditions, the prediction system being configured to predict at least one process variable of the machine. The method comprises a first step, in which a set of training data is provided, the set of training data comprising a plurality of ambient parameters of the machine, a plurality of performance parameters of the machine, and at least one process variable of the machine. Furthermore, the method comprises that a correlation value between the at least one process variable and each of the plurality of ambient parameters and each of the performance parameters of the machine is determined for a first time interval. In a further step, at least one model relevant ambient parameter and at least one model relevant performance parameter are determined based on the corresponding correlation values. Still further, the at least one model relevant ambient parameter, the at least one model relevant performance parameter, and the process variable are being fed into a model generating algorithm and generating a prediction model for the first time interval, the prediction model for the first time interval being a part of the prediction system. In the disclosed method, the model generating algorithm is a Generalized Additive Model (GAM).