BioMatNet Item: FAIR-CT95-1198 - AMOCO: Advanced monitoring and control of the operation of wastewater treatment processes of the wood industry in order to improve the process efficiency

FAIR-CT95-1198
AMOCO: Advanced monitoring and control of the operation of wastewater treatment processes of the wood industry in order to improve the process efficiency

Contacts
Summary Information

To find similar Items, click on a keyword below:
FAIR Area 2.3 - Control Systems : Process Engineering : Wood (Lignocellulose)

	Contract No:	FAIR-CT95-1198
	Date Prepared:	January 2000
	Source:	Final Report Abstract

Final Report Abstract Objectives

This project was directed to enhance monitoring and control systems in order to meet European constraints and to optimise the biological treatment of the effluents of the agricultural and food industry in general and of the wood industry in particular before rejection in the rivers. It is aimed at developing, applying and scaling up an advanced monitoring and control system in order to enhance performance of the on-line operation and to improve stable operation of a fibreboard manufacturing wastewater treatment plant.

Activities

The project work was organised in a number of tasks as follows:

Tasks 1 & 3: the preliminary step of the project is to characterise the dynamics of the involved wastewater treatment processes on the basis of pilot-scale and industrial-scale experiments via a dynamic model of the process.

Task 2: dedicated sensors will be developed and applied to the process to increase the number of available informative measurements for monitoring and control.

Task 4: the above mentioned information will be included in a mathematical framework to estimate additional parameters of the process.

Tasks 5, 6 & 7: a decision support based system will then combine both quantitative knowledge (basically reaction networks and mass balance models) and qualitative knowledge (such as physiological state and causality links between biological variables) in order to reach different objectives :

to optimise process efficiency
to improve its stable operation
to guarantee the wastewater quality
to quickly detect disturbances in process operation.

Task 8 : an important aspect of this project is the scale-up of the involved processes and methodologies that will be looked for as early as possible in order to apply them at an industrial wastewater treatment plant and to quantify their impact on the quality of the effluent treatment.

Results

An experimental protocol was determined by all the partner in order to reproduce the influent wastewater variations typical of industrial operating conditions. The protocol consists in applying to the anaerobic digester at a steady state different organic concentrations and different influent flow rates and to observe its transient behaviour to reach the new steady state. This protocol was used for experiments both at the pilot-scale and at the industrial-scale. A lot of variables were measured with on-line sensors (gas flow rates and compositions, pH, temperature, liquid flow rates) or with off-line measurements (Volatile fatty acids (VFA), COD, alkalinity, nutrients, etc). For certain conditions of high organic loading rate a destabilisation of the fermenter was observed (VFA storage and strong pH decrease). The key role of the alkalinity was pointed out : it guarantees a buffer capacity that avoids destabilisation of the fermenter.

New sensors have been studied and developed to monitor the wastewater treatment process. For the determination of alkalinity, two sensors have been developed, tested and validated on both plants. For the pilot-scale plant, a measuring device giving the values of the total alkalinity, the partial alkalinity, the concentration of volatile fatty acids, and the TOC (Total Organic Carbon) has been developed, tested, and is presently integrated on the pilot plant at the INRA in Narbonne. For the industrial plant, a titrating automatic flow system was developed, using a tubular proton selective electrode and a planar surface electrode with flow-cell. Different methodologies for titration execution were evaluated, namely point-by-point, second derivative and Gran transformed techniques. A prototype has be tested on-line, and validated on the industrial scale plant at TAFISA in Pontevedra. An automatic analytical system for phenol determination has been developed and tested with standards and real samples. Concerning COD automatic determinations the study of the oxidation phase using a flow microwave digester has been tested and flow titrating systems based in MCFS has been evaluated.

Software sensors have also been designed. They consists in a set of linear ordinary differential equations based on the mass balance equations model. The predictions of these sensors have been tested through extensive simulations. Both designed software sensors (the asymptotic observer designed on the basis of the mass balance model, the fuzzy observer) have been tested on the plants. Their predictions have been successfully compared with actual measurements for pilot- scale and industrial experiments. The asymptotic observer, which gives on-line estimation of the process components in the process (total alkalinity, volatile fatty acids, organic matter) and the concentrations of organic matter and volatile fatty acids in the influent, are integrated in the monitoring and control system, both at the pilot scale and at the industrial scale.

Different control strategies have been designed in order to meet the control objectives that have been defined by the partners (to maintain the total alkalinity larger than 3g/l and the ratio of intermediate alkalinity over total alkalinity lower than 0.3) and tested in numerical simulation : adaptive linear controllers and fuzzy controllers. These have been tested and validated both at the pilot scale and at the industrial scale. They are integrated in the monitoring and control system.

Different fault classes based on the physical structure of both, the pilot plant and the industrial scale plant had been defined as a framework to the "Fault Detection and Isolation (FDI)" strategy. They are related to the faults present in sensors, in local regulation loops and in the biological process. The related developed diagnosis module has been tested and integrated in the monitoring and control system. It is based on a number of rules (basically 32 rules) allowing to give explanations to the operator about problems existing in the process and equipment and the action to be taken in the system. The Troubleshooting Guide that has been prepared to allow during the operation of the process to check the whole system in order to correct the problems, has been used to design the expert system, that is now part of the supervision system.

The different parts of the project (information from the sensors, software sensors, controllers, diagnosis system, expert system) are integrated in the whole monitoring and control system in the format of a prototype software written in Matlab. The system software has been tested and validated both at pilot scale and at industrial scale.

A special session dedicated to the AMOCO project was organised at the IWA conference WATERMATEX held in Gent (Belgium) in September 2000.

Achievements

4 sets of experimental data are available (2 at pilot-scale, 2 at industrial-scale).
An alkalinity sensor has been developed and tested.
Sensors for nitrate, nitrite, phenol, and a COD measuring device have been tested.
Software sensors have been designed and validated.
Controllers have been designed and validated.
Fault detection systems have been designed and validated.
A supervision system has been developed and validated.
The whole monitoring and control system has been integrated, tested and validated.
The prototype software was implemented and validated at the industrial scale.