Overview
As this is a guide for designers, we start with a description of what the design of a
control system implies. The description can also work as an introduction to the different
concepts that are used in this guide.
A designer's starting point is, of course, the process that is to be controlled by the system,
and the first task is to learn about the process and to figure out the best way to control
it: what control loops are needed, what interlockings, how is the startup and shutdown
carried out, how are operators and maintainers going to work with the system. This is
summed up in a Design specification.
At the same time you have to consider what information about the process the control system
needs to perform its task, i.e. which sensors should be placed in the plant. The control
system also has to influence the process in various ways, for example with valves and engines.
This ends in a Signal list, that is a list of all input and output signals of the system.
At this point the question of which control system is to be used is raised, and one alternative
is of course ProviewR. You also have to decide which I/O-system to use and how to divide the
function among the different process stations.
IO-systems
The task of the I/O-system is to bring signals from the process to the control system, and to
put out signals to influence the process. The signals are usually digital or analog, but
there are also other types as integers and counters. You can choose between a rack and card
system connected to the process station, or distributed I/O, e.g. profibus.
Configure the system
When it's time to start to configure the system, you first create a new project in the
Administrator. The Administrator is a tool for creating order among all the projects, as there
can be plenty of them in time.
The configuration of a system is mainly done by creating objects in a database, the Workbench.
There are a large amount of objects to configure everything from IO-channels to PLC programs.
The ProviewR Object Reference Manual contains over 800 different types of objects. The objects
are placed in a tree structure, and you use a tool called the Configurator to create objects and
to navigate in the object tree.
The object tree is divided in two parts, the Plant hierarchy and the Node hierarchy.
The Plant hierarchy reflects the different functions in the plant and in the process, while
the Node hierarchy reflects the hardware of the control system, with computers, I/O-racks and
I/O-cards.
When the control system is started later in runtime, a copy of the object tree is created in
a realtime database, rtdb. The transfer from the Workbench to rtdb is done with so called
loadfiles, files that are generated from the Workbench and contain all its objects.
Control program
ProviewR offers a graphical programming language in which logic, Grafcet sequences and control
loops are programmed. It is named the PLC program. Also the PLC program is a part of the object
tree. It is configured by placing specific program objects, PlcPgm, in the Plant hierarchy.
When opening a PlcPgm you enter the Plc Editor, in which the graphical programming is preformed.
Function blocks are created, and connected in a signal flow of digital and analogous signals,
where the input signals are fetched at the left side, transformed in different function blocks,
and finally stored in output signals at the right side.
A complement to the PLC program are application programs, that are written in the c, c++ or
java language. Applications are written and started as separate programs and connected to
the realtime database through an API.
Simulation
The realtime database, the PLC program and possible applications kan easily be started in the
development station. This makes it possible to test the programs in direct connection with
the programming. You can even write special simulation programs which read the output signals,
simulate the outputs influence of the process, calculate values of different sensors and put
these values in the input signals.
The configuration and programming of the system is then a process where you switch between
configuring/programming and testing. The result is carefully debugged programs and a fast and efficient
commissioning of the plant. It also results in better programs, and more thoroughly worked
through functions, because the feedback is greater in the creating process than the construction
of a process control system implies.
At the simulation and commissioning it is of great importance to have access to tools that
makes it possible to monitor and examine the system, and quickly localize possible
malfunctions. In ProviewR this tool is called Xtt. Xtt contains a lot of functions to examine
the content of the realtime database, to flow the signal flow, to log fast or slow sequences,
etc.
Operator interface
There are a number of different groups of professionals that shall gain access to the system,
operators running the plant on a daily basis, maintainers occasionally correcting some
malfunctions, process engineers requesting various process data. All have their demands on the
system interface. Furthermore the limits between the various groups of professionals might be
fluid, operators that are both operators and maintainers, and perhaps even process engineers.
This puts high demands on the functionality and flexibility of the operator interface. You
can rapidly and easily trough so called methods that are activated from popup menues fetch all the
information of various objects that are stored in the realtime database, or in different server
systems.
Process graphics
Process graphics are built in a graphical editor. The graphics are vector based, which makes all
graphs and components freely scalable. Components have a preprogrammed dynamic to change
color and shape, depending of signals in the realtime database, or to respond to mouse clicks
and set values in the database. In each component that is sensible to mouse click or input,
you can state access, and selectively allow or hinder users to influence the system.
Supervision
If any malfunction arise in the system, the operator has to be noticed. This is done with
special supervisory objects that are configured in the plant hierarchy or in the PLC program,
and that originate alarms and events. The alarms have four priority levels: A, B, C and D,
and are displayed to the operator in the alarm list, the event list and the historical event list.
The alarmlist contains unacknowledged alarms and alarms in alarm state. An alarm normally has
to be acknowledged before it disappears from the list. If the alarm state is active, the alarm
remains in the list as long as it is active.
Alarms are also registered in the event list, that displays events in a chronological order.
The historical eventlist is a database, also registering events. Here you can search for events,
stating criteria as priority and process part.
If a plant part is shut down, it is possible to block the alarms to avoid to distract the
operator. Blocked plant parts are displayed in the Block list.
Data Storage
You often want to follow the change of a signal over time, in the shape of a curve. In ProviewR
there are three kinds of functions for this, DsTrend, DsFast and SevHist.
DsTrend is a trend that is stored in the realtime database. The value of a signal is stored
continuously with an interval of 1 second and upwards. For each curve there is space for about
500 samples, so if you choose to store a new sample every second, you have a trend of the
signal of about 8 minutes.
SevHist stores signals in a similar way in a database on disk, which makes it possible to store
values a longer period of time than DsTrend.
DsFast stores more rapid sequences, where the storage is started on a trigger condition, and
continues a specified time. When the sequence is finished, it is displayed as a curve.