Model and control of processes and systems

The ADERSA products are tools such as modelling, system analysis,identification and control which have been developed by ADERSA for industrial projects. These tools of general purpose got to a professional stage thanks to the improvements and to the inclusion of user friendly CAD modules (Computer Aided Design).

These products are :

applied by ADERSA to perform industrial projects,
left to the industrial users for maintenance purposes,
transferred to Engineering Companies and control systems manufacturers (technology and know-how transfer),
some of these products are distributed by third-part companies.

IDENTIFICATION
The identification is the evaluation of the parameters of a model from specific recorded experiments. The identification itself is performed by an algorithm after some pre-processing of the measured signals.		ADERSA developed several identification algorithms which are dedicated to different cases : SISO or MIMO systems, multi or monovariable case, structured representation (eg. Laplace transform) or non structured (impulse or step response).

 

Product	Platform	System	Monovariable	Multivariable	Representation
MONOREG	PC	MS-DOS	yes	yes	step response
GLIDE	PC & Workstation	MATLAB	yes	(*)	state, Laplace, others (*)
HIDENT	PC	WINDOWS	yes	yes	step response
PCR	PC	PCVUE-WINDOWS	yes	(*)	Laplace

 (*) this product makes possible to define a model structure in a user module. This structure may define a relationship between several input variables and an output.

(**) PCR is dedicated to the control of chemical reactors. In one of the configurations of the cooling equipment, the identification is to identify simultaneously two transfers.

   

MODEL BASED	PREDICTIVE	CONTROL
This type of controller is based on a model of the process, identified beforehand. The use of models allows the simulation of the full closed loop(controller and simulated process) in a CAD environment. Tuning and simulation tests yield the best conditions for performances and robustness.		ADERSA developed also several control algorithms which may be applied to different kinds of multi or monovariable processes: - furnaces, heaters, distillation columns - chemical reactors - robots, mounts, rolling mills - evaporators, critallisations - fluidised bed dryers, ...

 

Product	CAD System	Monovariable	Multivariable	Target platform
MONOREG	PC (MS-DOS)	yes	no	Supervision
PFC	PC & Workstation (MATLAB)	yes	(*)	PLCs or supervision
IDCOM-HIECON	PC (WINDOWS)	yes	yes	Supervision
PCR	PC (PCVUE-WINDOWS)	yes	no	PLCs or supervision

 (*) the way the control is formulated allows multivariable (MIMO) control; the CAD is designed for monovariable structures (SISO).

 

MAGE AND DATA ANALYSIS
The results of the work done in the field of perception and decision have been capitalised in software. Those are available, depending on their nature, either on Personal Computer or on scientific workstation.		They include one library of image analysis dedicated for industrial applications, one professional application and two general libraries of image analysis and multidimensional data analysis.

 

Product	Description	Host computer	Operating system
PIMWIN32 (*)	Software and library for image analysis	PC	MS-WINDOWS95 & NT
PIMCOMPT	Software for counting for microbial colonies out of PETRI boxes
PIMUX	Software and library for image analysis	Workstation	UNIX
KDTREE	Software and library for  multidimensional data analysis	Workstation	UNIX

(*) The library  PIMWIN32 is distributed by the Company   IMASYS under the name  IMALIB.

Model based predictive control, optimisation

The methodology and the tools are applied to very different industrial sectors such as food industry, oil, steel or chemical industry.

Model based predictive technology is used at levels zero and one in the hierarchical control of industrial plants.

Hierarchical control considers several levels:

level zero controls basic loops such as flow rates,

level 1 controls 'process' variables (temperatures, qualities, concentrations,...).

This is mainly at these two first levels that ADERSA operates.

Level 2 makes static optimisation and level 3 corresponds to production planning.

ADERSA applies MBPC and takes benefit of its experience of industrial projects to fulfil part or all of the following objectives :

Maximisation and minimisation are performed by the dynamic optimisation which is globally considered with the control objectives.

Integration, quality methodology, project management

ADERSA performs the integration of the developed systems into the target equipment ofthe site. The designed algorithms that have been adjusted and validated on simulators are integrated with operating procedures that are specific of a given unit (safety procedures, control mode switching,...).

Along the whole project, from the pre-study to the control starting, ADERSA follows a 'Quality' approach which includes project follow-up documents, commissioning definition and operator guide.

ADERSA can install the tools that are required for the evaluation of the performances in real time (production and process).