An original concept of modular evolved control system, seamless and with gradual integration into the primary control system is proposed.

The aim of the application is to integrate the concepts of evolved control algorithms, portability of software modules, real time characteristics of the application.

The target systems are the large scale distributed control systems with optimum granularity architecture.

The first part of the life cycle phases of the new control system, from conception to validation stage, the new control system lives "hiding in the shadow" of the control system it will replace, and after validation the old system will be replaced by the new one.

The identification, modeling, control and validation stages of the life cycle of the system, will be done on-line (the new system uses a real image of the I/O process data), without affecting the existing control system.

Because of high level of interconnectivity between system components, it is necessary to provide the highest independence between communication modules on one-hand and the control modules on the other hand. In order to obtain high ability of integration, the communication modules have to cover the widest possible area of industrial communication interfaces and protocols.

One item of the application is to offer a unified API of extended generality and extendibility in order to unify access and information retrieval from various wireless and wired technology and communication interfaces (RS 232, RS 485, fieldbus: Profibus / Interbus, Ethernet IP, TCP/IP, etc).

Applications could properly adapt to changes in the network connections.

The design and implementation of a solution to hide the embedded communication network problems from the application system programmers is included.

One of the main objectives of the application is to supply an integrated solution of systems, which should support all the phases of the life cycle: modeling, simulation, development and implementation.

For parameter tuning, for validation and also for embedding a large number of industrial communication protocols, multi-disciplinary simulation environments are developed which generate instruments for control, I/O data consistency check, and defect detection.

In the end, real-time advanced control applications are developed, with seamless and gradual integration into the existing distributed control system.

A software package for evolved control includes a method based on fuzzy model predictive control.

By using the basic concept of decomposition-coordination in a large-scale system theory, the fuzzy model predictive controller design can be accomplished through a two-layer iterative design process.

The design is decomposed into the derivation of local controllers. The subsystems regulated by those local controllers will be coordinated to derive a globally optimal control policy.

In order to provide the real-time characteristic, we choose a multitasking environment for the application (WINDOWS Operating System).

From structural point of view we propose a Client / Server architecture for fuzzy Controller (FC) (Andone et al., 2006):

Client — is a Windows application representing the implementation of the graphical user interface (GUI). The Client enables the operator to control the system in two modes: manual/automatic, to monitor the system response, etc. The Client has also the ability to connect and communicate with the Server application.

Server — is an ActiveX EXE application containing the implementation of the Fuzzy Controller (FC) kernel.

The Server includes a collection of objects, these objects cover the tasks of both data processing and the communication between dedicated applications for input and output data.

The Client application will have a thread pool architecture.

The Server application will have a real multithreading architecture (each active object having assigned its own execution thread).

The Server have also a multi-layer structure: at the higher level are implemented upper FC and the communication classes (using different transmission mechanisms — DDE, OPC, HLI, ActiveX, Winsocket, Pipes), at the lower level are implemented the controllers for the subsystems corresponding to the low level FC.

The Server’s application as real multithreading architecture, provides the FC Kernel the real­time response characteristic, required for the industrial process control.

4. Conclusions

Control of SG water level strongly affects nuclear power plant availability.

The control task is difficult for a number of reasons, the most important among them being the nonlinear plant dynamics and the non-minimum phase plant characteristics.

There has been a special interest in this problem during low power transients because of the dominant reverse thermal dynamic effects known as shrink and swell.

The SG level control problem was viewed as a single input/single output control problem with the feed-water as the manipulated variable, the level as the controlled variable and the turbine steam demand as disturbance.

It has been shown that in the case of nonlinear processes, the approach using fuzzy predictive control gives very promising results.

The process non-linearity was addressed by scheduling the model (and the controller) with the power level.

The SG system is modeled by Takagi-Sugeno’s fuzzy modeling methodology, where the system output is estimated based on gradient. The complex shrink and swell phenomena associated with the SG water level are well captured by the model.

The predictive controller based on fuzzy model is designed in a hierarchical control design.

An original concept of modular evolved control system, seamless and gradual integration into the existing distributed control system is proposed in the chapter.

A unified API of extended generality and extendibility in order to unify access and information retrieval from various wireless and wired technology and communication interfaces is developed in order to ensure independence between communication and control modules of the designed systems.

A Client / Server architecture for evolved controller that runs on the Windows

environment, with real-time characteristics is proposed.

5. Acknowledgment

Parts of this chapter are reprinted from Hossu, D. Fagarasan, I., Hossu, A., Iliescu, S. St.,

Evolved fuzzy control system for a steam generator, Int. Journal of Computers, Communications

and Control, IJCCC, ISSN 1841-9836, E-ISSN 1841-9844, Vol. V (2010), No.2, pp. 179 — 192.