The case for Real time (Hard real time vs. Soft real time)
Real time inputs and feedback of operating conditions is
probably a control engineers dream come true. However, with the latency introduced in
automated systems due to various factors like -
Signal conditioning and Transmission,
Data Sorage and retrieval,
Processing of arithmetic and logic,
Priority assignment and priority based interruption,
Some processes or loops in a process like manipulating the field
magnets in a magnetic spectrometer may require latencies or response times as low as 10ms
and with a repeatability within ±100ms, and are categorized as applications requiring
hard real time control. However, other processes needing soft real time, like turning on
an agitator on a reactor vessel, can suffice with a response time of 10ms and with a more
relaxed norm on repeatability.
The case for Embedded OS and the market offerings
Most operations in manufacturing, involve minute control of a multitude of loops. Any medium sized manufacturing unit of today employs as
many as 1000 loops. If one were to use human control inputs, the complexity probably would
rival that of building the Sphinx, on a minute-to-minute basis. To avoid such
administratively challenging tasks, was born the embedded loop controller with very
scientific algorithms like Proportional, integral, derivate or a combination, applied to
any loop in the global process.
With the scale of economies that silicon today achieves, it is
possible to combine the function of many loop controllers into one integrated unit,
marshaled by an operating system rather than low level assembly instructions that are very
complex to maintain and develop. Such operating systems, with very minute footprints, are
usually an integral part of the hardware of the control system, and for reasons of
reliability are usually embedded in the memory on this system.
Case for integration with distributed I/O
Compare with Loop controller, PLC, DCS and Hybrid
The evolution of the process control industry has seen the simplest of automation taks, i.e. a
loop, being automated by a loop controller. Today loop controllers are specialized and
very integrated with the process they are built to control. However, like all mass
manufactured and scaled up processes, it is usually necessary to incorporate loop control
systems for a larger batch operation like a distillation column or a batch reactor.
Programmable logic controllers (PLCs) have now take this role of automating such tasks and
ideally are suited to handle such operations, with configuration usually build by ladder
programming or IEC 1131 programs.
A majority of todays sensors (or transducers) are capable of producing an analog signals to
indicate their state. When multiple signals of such nature are integrated into a control
system, the cost of a PLC based system is usually higher that some computer based systems.
Distributed control systems (DCS) or hybrid systems are commonplace in most plants today, where a
supervisory network co-ordinates the various loop controllers or PLCs (in a hybrid
configuration).
The ideal combination as is evident is to incorporate the deterministic operation (hard real
time control) of the loop controller and the supervisory capabilities of a modern PC based
network. These systems, which are increasingly gaining popularity in industrial control
applications, are called distributed I/O systems, where the I/O subsystem is installed to
automate a particular or several operations. A supervisory industrial grade computer
connected to various such I/O subsystems forms the building block of the evolving
distributed I/O control architecture. In a typical industrial configuration, may such
distributed I/O systems could be monitored by a central operations center, which in turn
could be integrated with an MIS system prevalent in the organization.
Fig. The distributed I/O based automation architecture
Industry Verticals to service |