Educational Simulators for Industrial Process Control - Simulation and Modeling Methodologies, Technologies and Application - page 151

mechanism to avoid the well-known “bumpless” of the auto/man controller commuta-
tions. With respect to the algorithm to calculate the control signal, it is possible to use
diverse algorithm and PID structures and the PID calculus use normalized SP, PV and
OP values.
Pressing the setting (“AJUSTE”) button of the GUI of each PID, the user accesses
to the tuning parameters: proportional gain (Kp), reset time (Ti), derivative time (Td),
sampling period (Tm), SP and PV time constant filters, PV and OP span values in
Engineering Units (PVEULO: Process Variable Engineering Units Low, PVEUHI:
Process Variable Engineering Units High, OPLO: Output to Process Low, OPHI:
Output to Process high). Additionally, there are three menus to select the type of algo-
rithm (Ideal or Interactive), the PID equation (PID, PI-D, I-PD, I) and the action con-
troller (direct or reverse) that affects to the sign of the controller gain.
The previous experiment can be repeated with other FC1 controller parameters, for
instance: Kp=5 and Ti=0.1. Then the control structure doesn't work well (Fig. 8).

Fig. 8. Bad FC1 tuning: process and control structure response

4 Software Structure and Development Tools

When selecting a particular module a SCADA system is started. This SCADA is
called EDUSCA [15] and it is the simulation module GUI. EDUSCA starts the simu-
lation program linked to the selected module. The development of each module GUI
involves the EDUSCA setting, which is done by a drag & drop strategy through a
setting tool (Fig. 9).
The communication between EDUSCA and the simulation program is performed
by the OPC (OLE for Process Control) communications standard for process control
applications for Windows environments [16]. EDUSCA acts as an OPC client and the
simulation program as an OPC server.
The simulation models have been performed using EcosimPro. EcosimPro belongs
to the so called object oriented modeling languages (OOML). Many of the EcosimPro
characteristics are similar to the modeling tools that implement Modelica [17]. In the
sense that it supports non-causal models able to be modified automatically
according to the context in which they are used. Its simulation language, called EL