Skip to main content

Advanced Intelligent Multisensor System for Control of Boilers and Furnaces

Objective

The objective of this project is to develop, implement and test a general-purpose multi-sensor system to optimise the operation of industrial fuel-burning systems, and to control and assist the maintenance of industrial furnaces and boilers.
The objective of this project was to develop, implement and test a general purpose multisensor system to optimise the operation of industrial fuel burning systems, and to control and assist the maintenance of industrial furnaces and boilers. The project was aimed to assure optimum conditions of industrial burning systems through the maintenance of efficient flame characteristics, which are achieved with the implementation of adequate control mechanisms in 2 different levels. Low level control aims at the definition of simplified dynamic models and proper control strategies and includes sensor integration and the implementation and optimization of new sensors, either physical and chemical or optical. It has been given particular attention to the development of a vision system to acquire data on flames and to classify them according with previously learned standards. Other sensors under development included viscosity meters for heavy fuel oil and molten glass and a volumetric fuel flowmeter for fuel oil.
High level control aims at the definition and implementation of actions based on: sensorial information, specific knowledge driven by sophisticated mathematical models of the flow and heat transfer characteristics of industrial burning systems and empirical knowledge of subsystem not considered by the mathematical models.
The AIMBURN project was essentially industry driven, but the progress of the work revealed a scientific potential which was partially exploited. Considerable innovative and original work include: use of computer vision for flame classification, experimental characterization of turbulent flames in mutual interaction, numerical simulation of turbulent flame images, implementation of furnace control strategies based on physically derived models, sensor development for molten glass and heavy fuel oil viscosity, dedicated hardware integrating recently produced digital signal processing (DSP) chips in purpose built architectures, development of physical submodels to predict nitrogen oxide emissions from industrial combustion environments, application of expert systems to real time processes, and standardization of procedures for knowledge acquisition in continuous plants.
The impetus for the work stems from the fact that industrial competitiveness within the European Community over the next decade will undoubtedly depend on the efficient use of information technologies to optimise energy consumption. Since energy consumption and pollution are frequently linked, the implications for pollution abatement may have also to be considered in the application of IT to improve industrial competitiveness. The project is aimed to assure optimum conditions of industrial burning systems through the maintenance of efficient flame characteristics, which are achieved with the implementation of adequate control mechanisms in two different levels, as follows:

Low-level control

- Definition of simplified dynamic models and proper control strategies.
- Implementation and optimization of new sensors, either physical and chemical or optical. It has been given particular attention to the development of a vision system to acquire data on flames and to classify them according with previously learned standards. Other sensors under development included viscosity meters for heavy fuel-oil and molten glass and a volumetric fuel flowmeter for fuel-oil.
- Sensor integration.

High-level control

Definition and implementation of actions based on:
- Sensorial information
- Specific knowledge driven by sophisticated mathematical models of the flow and heat transfer characteristics of industrial burning systems.
- Empirical knowledge of sub-system not considered by the mathematical models.

The AIMBURN project has been essentially industry-driven, but the progress of the work has revealed a scientific potential which has been naturally exploited. Considerable innovative and original work includes:

- use of computer vision for flame classification
- experimental characterisation of turbulent flames in mutual interaction
- numerical simulation of turbulent flame images
- implementation of furnace control strategies based on physically-derived models
- sensor development for molten glass and heavy fuel-oil viscosity
- dedicated hardware integrating recently produced DSP chips in purpose-built architectures
- development of physically sub-models to predict NOx emissions from industrial combustion environments
- application of expert systems to real time processes
- standardisation of procedures for knowledge acquisition in continuous plants.

Coordinator

ASSOCIACAO PARA O DESENVOLVIMENTO DO INSTITUTO SUPERIOR TECNICO
Address
Avenida Rovisco Pais, 1
1096 Lisboa
Portugal

Participants (10)

ASSOCIACAO PARA O DESENVOLVIMENTO DO INSTITUTO SUPERIOR TECNICO
Portugal
Address
Avenida Rovisco Pais, 1
1096 Lisboa
ELECTRICIDADE DE PORTUGAL
Portugal
Address
Rua Cidade De Goa, 4
2685 Sacavem
FABRICA DE VIDROS BARBOSA & ALMEIDA
Portugal
Address
Apartado 27 - Avintes
4407 Vila Nova De Gaia
IMPERIAL COLLEGE OF SCIENCE, TECHNOLOGY AND MEDICINE
United Kingdom
Address
Exhibition Road
SW7 2BX London
INSPECCION Y GARANTIA DE CALIDAD
Spain
Address
Avda. De Europa, 26
28224 Madrid
INTELLIGENT DECISION SYSTEMS
Spain
Address
Avd. Albufera, 153 4
28038 Madrid
MAGUE
Portugal
Address

2616 Alverca
SERVOTROL
Portugal
Address
Praceta A-avenida Cidade Lourenco Marques Lote, 532/10
1800 Lisboa
TRION PRAEZISIONSELEKTRONIK GMBH & CO KG
Germany
Address
Voltastraße 5
13355 Berlin
UNISOFT
Portugal
Address
Rua Actor Antonio Silva, 7
1600 Lisboa