Who's afraid of control in the field? Part 2

The possibility of placing control function blocks in field devices is an important feature of Foundation Fieldbus technology, but what does it mean in practice? What else can go wrong?

Something comes along and chews up the fieldbus cable to the controller: I lose the control in all the devices on the segment. Godzilla is in.

Figure 1: Example of field control system in a distillation column. The field I/O was used for motor control and simple flow control (flowmeter and valve) was located in the valve positioner, as was the temperature control in the column (temperature sensor, pressure sensor and valve positioner).

Cable breaks can occur in any system and the consequences are always that at least one loop is lost. What is quicker, to hook up the two bus wires or locate and repair an analog connection?

1. Since power is normally lost, the mechanical override on the valves will cause them to move to the fail-safe position selected.
2. Since communication is lost, any loops in other segments requiring data from the affected devices will flag bad input status and default to the programmed fail-safe state.
3. If the loop was designed with power redundancy only, the back-up LAS will take over control and all loops within the affected segment will function normally. External loops will react as in Item 2.
4. If power, controller (hardware) and control (application) redundancy has been provided, the back-up controller will take over control and all loops will act normally.

Psycho

The controller is taken out, I lose all control.

In other systems yes, but with Foundation Fieldbus it depends on the control strategy you have selected. Assuming there are no function blocks being executed in the controller:

1. Normally, the back-up LAS in each segment will become active and all loops within the connected segments will run normally (assuming there is still power). Any cross-segment loops will fail and default to the programmed fail-safe state.
2. If controller redundancy has been provided, the back-up controller will take over control and all loops will act normally. If there are function blocks being executed in the controller:
3. Normally, the back-up LAS in each segment will become active and all loops within the connected segments will run normally (assuming there is still power) with the exception of those with blocks being executed in the controller. These, together with any cross-segment loops, will fail and default to the programmed fail-safe state.
4. If controller and control redundancy have been provided, the back-up controller will take over control and all loops will act normally.

Headless horseman

The HMI system goes. I am blind and have no idea what is going on.

If this has ever happened to you, then you will have provided a back-up system. High-speed ethernet is part of the Foundation Fieldbus specification, so depending on your control system, this does not have to be expensive, since components-off-the-shelf (COTS) can be used to support network redundancy.

1. Losing the HMI has no effect on the controller and devices: all loops will continue to work.
2. Providing that the OPC Server is also redundant, if a back-up HMI is present, it will take over.
3. It is perfectly feasible to provide central or local emergency panels and buttons. FF function blocks include discrete inputs that allow manual override of critical loops, switching them to fail-safe mode.

Where does control in the field make sense?

Although it might sometimes be a case of preference, control in the field usually makes sense when there are economical or technical benefits to the user. When comparing the cost of Foundation Fieldbus to other systems, every control block used in a field device replaces an analog channel on a DCS. Within a Foundation Fieldbus system itself, however, there is no obvious economic benefit from running the control blocks in the devices, unless the controller is limited in the number of blocks or links it can handle, a high degree of loop integrity is required or the controller is not necessary at all.

Figure 2: Examples of control strategies and the effect of location of function blocks.

Where blocks are running, control in the field does have an effect on the macro cycle length. By careful design of both segment and control strategy, most applications can be effectively solved. Limitations are set by the number and type of blocks available and the number of links that can be supported. These in turn depend on the field device and host system.

The examples that follow indicate what control in the field looks like in practice, see Figure 1. The host and devices used supported multivariable optimisation.

Simple feedback flow control

The optimal location for the PID block of a control valve is the positioner itself, see Figure 2. Not only does this reduce the number of links and loop integrity, it also ensures that the block assumes fail-safe mode when the input is bad. There is no intervention by an operator or other devices in the network.

Temperature control in a distillation column

As a continuous process, the regulation of temperature in distillation columns (or boilers)provides an ideal application for control in the field. In the application shown, see Figure 1, solvents were distilled in batches, meaning that there also had to be some kind of ramp control over the initial boiler heating phase. This was done by means of a characteriser and arithmetic block which, together with a constant block for tracking, were run in the valve positioner. In all, five external publisher links were required. Had the application been implemented in the controller more than twice as many links would have been required, making a noticeable impact on macro cycle time and decreasing the degree of loop integrity. When many loops are necessary, the reduction in cost becomes more apparent because it is not necessary to replicate local controllers that execute each temperature control loop independently.

Motor control from a field I/O device

As mentioned earlier, currently Foundation Fieldbus is not very strong in logical control. The general method of managing analog and digital I/O signals is to add I/O cards to the controller. There are, however, several possibilities to integrate PLC platforms into Foundation Fieldbus or to use FF field I/O devices. The latter, mounted locally on the H1 segment, was the solution selected. By using a flexible function block, a small degree of logical control is available in the field. This was suitable for the motor control centre in the distillation plant just described.

Simple is beautiful

Does control in the field have a future? As control systems have increased in size and capabilities, the remoteness and complexity generated by a proliferation of smaller, faster and cheaper technological components in a large system has been forgotten. Could this be one of the main reasons why people are 'afraid'? Is it the thought of relying on thousands of components to work together reliably - components that no longer deliver information to one source, but which, in order to control a process, distribute it, apparently without supervision, across the field? From the technical point of view, there is no longer a need for a centralised control room.

Figure 3: Possibilities for providing logical control in a Fieldbus system. 1. Local I/O connected to the field controller; 2. Field I/O with flexible function blocks for discrete control and interlocks; 3. PLC connected via ethernet for sequential control.

Today's control systems are advanced enough that the control room is not needed to coordinate the control. Perhaps the way ahead is to go back to the field and create small, manageable units, capable of operating by themselves, but reporting regularly to the supervisory system. Foundation Fieldbus is suitable for such systems, since a segment can also be operated without a 'controller', indeed without control, when required. At the same time there is a need for 'simplicity'. Simplicity does not preclude users from undertaking skilled tasks such as connecting up the equipment, rather it implies that they would like to see a lot more plug and play in device integration, configuration and design of the control strategy.

Undeniably, Fieldbus is an 'enabling' technology, but let us not forget the people who must work with it. Manufacturers can make Fieldbus reality much simpler and more effective. As a result, fears and concerns will no longer find the nourishment to sustain them. When the Zombies can be exchanged and put to rest at the push of a button and the plant manager can sleep well at night, the goal will have been reached.