Method for building fractal dimension-based multi-phase mixed effect prediction model

Abstract

The invention discloses a method for building a fractal dimension-based multi-phase mixed effect prediction model, which comprises the following steps of: (1) acquiring a multi-phase stirring mixing real-time pattern by using electron tomography or a high-rate camera; (2) computing the fractal dimension of the acquired real-time pattern by using a compiled box dimension computation program; (3) acquiring previous relative data of each period at every interval by using the method in step (2) to obtain a corresponding time sequence; (4) computing a time delay variable, an embedded dimension and a bandwidth by using the time sequence obtained in step (3) and by using a phase reconstruction method; (5) processing the time sequence in the step (3) by using a one-step-forward method and kernel regression prediction to obtain an entire time sequence; and (6) testing and verifying a prediction error by using an experimental method. The invention provides a reliable and practical prediction method for predicting the mixed effect of chemical engineering experiments and theoretically instructing stirring reactor design, stirring operation and correction experiments.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

The technical effect of this inventor's new approach described by its patents includes improving both classical models (ARM) used for data analysis and estimating parameters from them over time without requiring significant modifications or rebuilding effort. It also allows for easy tuning when there may have problems with certain variables that affect their performance during operation. Overall, these improvements make better predictive tools more efficient than current techniques like linear regression and partial least squares.

Problems solved by technology

This patented technical problem addressed in this patents relates to improving the accuracy of analyzing complex systems like industrial processes due to their temporal dynamics over time. Current methods involve linear regression, multidimensional integration, partial state equation simulation, advanced machine learning algorithms, wavelets transform, spectral analysis, spatial filtering, and others. These approaches aim to identify relevant aspects within each frame period while considering both internal and outside influences.

Images