Fail Open

<a href="http://www.jashaw.com/pid/action/">http://www.jashaw.com/pid/action/</a>

Controller Action and Fail Open Valves

As you probably know, the action of a controller is defined as:

• If an increase in the measured variable results in an increase in output, the controller is direct acting.
• If an increase in the measured variable results in a decrease in output, the controller is reverse acting.

The action of the controller depends upon the action of the process. The controller action is the opposite of the process action. For example, a flow controller reduces its output when the measured flow increases.

However, what do we do if the valve is fail open (also known as "air-to-close")?

For most control loops, an increase in signal from the controller (electrical or air pressure) causes the valve to open. However, if we follow the concept of fail safety we design control loops so that the most likely failure (loss of electric power or loss of instrument air pressure) will result in the less dangerous condition. Therefore, certain valves, such as those that admit cooling water to a process, will be configured to be fully open with no air pressure (or signal from the controller) and to close with increasing air pressure or signal. These valves reverse the action of the loop--the controller output increases in order to close the valve.

So what do we do about the controller action when the valve is fail open?

Answer: It Depends.

Typically, older analog controllers required that the action of the controller be changed if the a fail open valve is used. For example, a flow loop controlling the flow of cooling water would use a fail open valve and therefore direct action.

Digital controllers, such as DCSs, typically have a separate inversion function to correspond to the valve action. Their control algorithm action is not affected by the valve. A flow controller will be set to reverse action.

A Little Bit of History

The analog days

When the process industries used pneumatic and electronic controllers, mounted in the field or on panel boards, the output of the PID control function would be connected directly to the valve (or to an I to P converter to convert the current to pressure). Our example, the flow loop for cooling water, would have a direct acting controller. To prevent operator confusion, there would be some means of inverting the output display (such as turning the output gage upside down) and inverting the operator’s manual control (such as reversing the labels on the OPEN and CLOSE buttons). For decades, through the evolution of controls from field mounted pneumatic boxes to small electronic panel mounted controllers, this standard was used.

The digital evolution

In the mid-1970’s computer control gradually entered the process industries. Digital "mini-computers" would power the PID control loops as well as more advanced control methods. The outputs (sometimes up and down pulse trains or incremental positive and negative currents) were connected to “computer/manual” stations that would let the operator switch the process off of computer control and operate the valves directly. (These were used when the computer failed--not an uncommon event.)

These "computer/manual" stations (and other devices that converted a signal from the computer to an analog signal) provided the means to invert the signal for use with fail open valves. Therefore, the action of the control algorithm in the computer was not affected by the action of the valve. Flow loops would be configured to reverse action, and, if the valve were fail open, the computer/manual station would be set to invert the signal.

Modern DCS control

Now, digital control has almost completely replaced analog control. Modern distributed control systems (DCSs) perform basic regulatory control (i.e. the typical PID control loop) using software function blocks to execute the PID control algorithms and do other functions. DCS installations replace both the function of the control computer and the function of the analog output module. Functions within the DCS define the analog output and control the conversion of the output of the PID control function to the 4-20 ma. (or digital bus signal) for transmission to the valve.

So what do we do now?

Most, if not all, DCSs allow the configuration of the PID control function to be based on the process, without regard to the action of the valve. In order to invert the output to loops with reverse acting (air to open or fail close) valves, either the inversion will be handled by an option in the analog output definition or by the use of a function to invert (convert 0% to 100% and vice-versa) in the software definition of the loop. Received on Sat Dec 02 03:20:00 2006