Bayesian Time-Series Prediction for Missing Manufacturing Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manufacturing systems face challenges in predicting and adjusting parameters using multivariate time series data, particularly due to missing values and asynchronous data acquisition, which affects the accuracy and efficiency of equipment maintenance and operational settings.
Innovation Solution
A method and system that utilize a Bayesian model to interpolate missing values in multivariate time series data, store prediction data with uncertainty, optimize model parameters, and adjust manufacturing system parameters based on predicted values, enabling efficient prediction and maintenance scheduling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to handle missing values in time series data, then data completeness is maintained, but prediction accuracy deteriorates
Solution Approach 1:
A Bayesian imputation model is introduced as an intermediary mechanism to estimate and fill missing values in time series data. The model uses observed data points to infer probable values for missing entries, thereby maintaining data completeness while preserving prediction accuracy. This intermediary process transforms incomplete raw data into a complete dataset suitable for accurate predictions.
2Measurement precision
If complex models are used to predict multivariate time series, then prediction accuracy is improved, but computational cost increases
Solution Approach 1:
The system dynamically adjusts model parameters based on data characteristics and prediction requirements. By optimizing parameters such as model complexity, time window size, and imputation precision, the system achieves accurate predictions while minimizing computational overhead. This parameter tuning allows the model to adapt to different operational contexts without requiring consistently high computational resources.
3Loss of information
If frequent data collection is performed, then data availability is improved, but data acquisition time increases
Solution Approach 1:
The system collects and processes only the necessary subset of data required for accurate predictions, rather than continuously gathering all possible data points. By identifying and focusing on critical time windows and key parameters, the system achieves sufficient data availability for reliable predictions while reducing overall data acquisition time and processing overhead.
4Adaptability or versatility
If manual adjustment of manufacturing parameters is performed, then system adaptability is maintained, but productivity decreases
Solution Approach 1:
The system implements automated feedback loops where prediction results directly inform parameter adjustments in the manufacturing process. The Bayesian model continuously monitors time series data, predicts future states, and automatically adjusts operational parameters based on these predictions. This closed-loop feedback system maintains high adaptability to changing conditions while eliminating manual intervention, thereby maximizing productivity.
Data Source
AI summary
The present disclosure describes a method of controlling a manufacturing system using multivariate time series, the method comprising: recording data from one or more devices in the manufacturing system; storing the recorded data in a data storage as a plurality of time series, wherein each time series has a first recorded value corresponding to a first time and a final recorded value corresponding to an end of the time series; interpolating, within a first time window, missing values in the plurality of time series using a Bayesian model, wherein the missing values fall between the first and end time of the respective time series; storing the interpolated values as prediction data in a prediction storage, wherein the interpolated values include the uncertainty of each interpolated value; loading the recorded data that fall within a second time window from the data storage; loading prediction data from the prediction storage that fall within the second time window and for which no recorded data are available; optimizing the parameters of the Bayesian model using the loaded recorded data and the prediction data; predicting, using the Bayesian model, values for each of the time series for which loaded recorded and prediction data are not available; storing the predicted values as prediction data in the prediction storage, wherein the prediction values include the uncertainty of each prediction value; and adjusting one or more of the devices that generate the recorded data based on the prediction data within the second time window.