Skip to main navigation Skip to search Skip to main content

A data-based KPI prediction approach for wastewater treatment processes

  • School of Astronautics, Harbin Institute of Technology
  • Bogazici University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, the Benchmark Simulation Model No. 1, which is designed for the purpose of simulating actual wastewater treatment processes, is introduced and implemented in SIMULINK environment. Then the partial least squares (PLS) model and its kernel version is studied, and wavelet transform is used to carry out the so called multi-scale kernel partial least squares (KPLS). By means of multi-scale KPLS, the prediction of key performance indicator (KPI)-the COD concentration in effluent-is implemented. Simulation results show that this prediction model has strong generalization ability under the condition that the data collected during the wastewater treatment processes are distributed unevenly and coupled tightly.

Original languageEnglish
Title of host publicationProceedings - 2015 International Conference on Man and Machine Interfacing, MAMI 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509002252
DOIs
StatePublished - 19 Apr 2016
Externally publishedYes
EventInternational Conference on Man and Machine Interfacing, MAMI 2015 - Bhubaneswar, Odisha, India
Duration: 17 Dec 201519 Dec 2015

Publication series

NameProceedings - 2015 International Conference on Man and Machine Interfacing, MAMI 2015

Conference

ConferenceInternational Conference on Man and Machine Interfacing, MAMI 2015
Country/TerritoryIndia
CityBhubaneswar, Odisha
Period17/12/1519/12/15

Keywords

  • BSM1
  • KPI prediction
  • Multi-scale KPLS
  • Wastewater treatment

Fingerprint

Dive into the research topics of 'A data-based KPI prediction approach for wastewater treatment processes'. Together they form a unique fingerprint.

Cite this