Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

An Incremental Spatiotemporal Learning Method for Online Modeling of Distributed Parameter Systems

A learning method and distributed technology, applied in general control systems, control/regulation systems, instruments, etc., can solve the problems of online updating of space-time synthetic sets, data identification of space-time synthetic sets, time-consuming problems, etc., and achieve remarkable calculation results , broad application prospects, the effect of reducing time and memory usage

Active Publication Date: 2021-07-27
CENT SOUTH UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Due to the increasing amount of training data, the method of batch processing needs to retrain the entire spatio-temporal synthetic set. With the continuous inflow of data, the number of training sets is getting larger and larger. On the one hand, it will consume a lot of time; burden
Because the existing technology cannot collect all the training data separated by time and space, and can not use all the data to identify the space-time composite set from scratch, there is a problem that the online update of the spatio-temporal composite set is difficult

Method used

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
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An Incremental Spatiotemporal Learning Method for Online Modeling of Distributed Parameter Systems
  • An Incremental Spatiotemporal Learning Method for Online Modeling of Distributed Parameter Systems
  • An Incremental Spatiotemporal Learning Method for Online Modeling of Distributed Parameter Systems

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0143] This example presents an incremental spatiotemporal learning method for online modeling of distributed parameter systems, including:

[0144] (1) After adding new data to the incremental data set, incrementally update the spatial basis functions;

[0145] (2) Update the time coefficient and identify the new timing model;

[0146] (3) Reconstruct historical data through the old space-time composite set, the updated spatial basis function in step (1) and the time series model identified in step (2), and predict future output;

[0147] (4) Repeat steps (1)-(3) to complete the online update of the spatio-temporal synthetic set.

[0148] Among them, the data increment set is the collected continuous data flow with a specific time step, the spatial basis function is the basis function of n-dimensional space, and the basis function of n-dimensional space is separated by time and space, with the time step as L, learned from the training data. , Incremental update refers to t...

Embodiment 2

[0156] In order to prove the performance of the incremental space-time learning method for online modeling of distributed parameter systems in Example 1, we take the catalytic reaction rod experiment as an example, and compare Example 1 with the traditional batch processing mode method to illustrate Feasibility and advantages of incremental learning.

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

PUM

No PUM Login to View More

Abstract

The example of the present invention provides an incremental space-time learning method for online modeling of a distributed parameter system. The method first adds new data to the data incremental set, then incrementally updates the spatial basis function, and updates the time coefficient. Identify the new time series model, and then reconstruct the historical data through the old space-time synthetic set and the updated spatial basis function and the identified time series model to predict future output. The method of the embodiment of the present invention makes up for the shortcomings of the existing methods, reduces the calculation time and the amount of memory used by the device, is simple and easy to implement, and has universal applicability in industrial modeling. Both theoretical analysis and experimental results prove that the incremental space-time learning method can It achieves good online performance, and at the same time, the calculation effect is remarkable, and the application prospect is broad.

Description

technical field [0001] The invention belongs to the technical field of industrial process control, and in particular relates to an incremental space-time learning method for online modeling of a distributed parameter system. Background technique [0002] Distributed parameter system (distributed parameter system, DPS) widely exists in the field of industrial processes, such as semiconductor manufacturing, nanotechnology, bioengineering and chemical engineering, and is usually expressed by partial differential equation (PDE). Since the state of DPS changes continuously with the position of time and space, it needs to be described by an infinite-dimensional state space, while PDE is finite-dimensional, so model decrement is unavoidable in practice. Proper modeling of DPS complex systems is very important for industry Simulation, control and optimization are critical. [0003] The space-time separation method is a method that can effectively simplify the modeling of unknown DP...

Claims

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

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G05B13/04
CPCG05B13/042
Inventor 李涵雄李旭昊王志
Owner CENT SOUTH UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com