Flow Measurement Response Time, Filtering, Smoothing

With all flowmeters, it is important to consider the tradeoff between measurement response time and smoothness or stability - to optimize your flow monitoring application. Response time is how fast or slow the flowmeter measurement updates to changes in actual flow. Whether you want the meter to respond fast or slow depends upon your application.

For example, consider a juice dispensing machine which fills small bottles as they move along a conveyor belt. This is what we call a 'batch application' where the flow has very short durations or batches. In this case, you certainly want the flowmeter to catch the start and stop of the flow as soon as it happens so the bottles are filled perfectly in a short time. Alternatively, perhaps your process has long durations of unstable flow, but you require a very stable measurement. In this case, you would want your measurement to respond slowly to changes in flow, so that it does not fluctuate. Fortunately for both cases, there is a control in your flowmeter that can be used to adjust how your flowmeter responds to changes in actual flow. This is the time constant setting.

The time constant setting is accessed within your flowmeter’s mobile App (Orcas Flowmeter App) in the Setup Menus -> Advanced Settings menu. The time constant has units of seconds and communicates how long the flowmeter should take to respond to a change in flow. If you are working with short durations of flow (like the juice manufacturer), you should set the time constant to zero which tells the flowmeter to respond to any change in flow as fast as it can. If you are on the other end of the spectrum (measuring long durations of flow), and perhaps have unsteady flow, then you may want to increase the time constant to smooth the measurement. Let’s consider two images to better understand how the time constant works.

Figure 1 This figure illustrates the flowmeter response to a step change in flow. The red line is actual flow, and the black lines are the flowmeter’s response for three different time constants.

Let’s consider a step change in flow, and the flowmeter’s response for three different time constants. A step change in flow would be the case when there is no flow, and then suddenly a pump kicks on at full throttle. Figure 1 displays what this looks like, the red line is the actual step up in flow, and the black lines are the flowmeter measurements for three different time constants (2, 9, and 31 seconds). You can see that for the smallest time constant, the flowmeter responds much faster to the actual flow, while the longer time constant takes longer to respond.

You probably noticed that the flowmeter takes longer to reach the actual flow value than the time constant itself. This has to do with how the time constant is defined. We use the traditional definition, which is ‘the time required to reach 63% of the actual value’. So, reaching the full 100% of actual flow will take longer than the 'time constant' value itself.

Figure 2 This figure displays an unstable flow (red lines) and the flowmeter's smoothed response for three different time constants.

In the second example shown in Figure 2, the red line is actual flow, which is rather unstable, oscillating up and down. The black lines are again the flowmeter measurements for various time constants. You can see right away that the larger the time constant is, the more stable the measurement.

It is important to note that you can't have it both ways. There is a tradeoff between fast response and smoothness or stability. You cannot have a flowmeter respond instantaneously AND at the same time provide perfectly smooth measurements. You can see this by comparing the two figures provided. In Figure 2, the large time constant provides good smoothing, but provides a slow response in Figure 1. Alternatively, in Figure 1, the small-time constant provides a fast response, but then is unsteady in Figure 2. You can find a middle ground by adjusting the time constant for your specific application.

Until now, we've been talking about static filtering, where you pick a time constant and that is the one the flowmeter will use. However, by default our flowmeters are configured to use a dynamic time constant, which is our own way of trying to get around the problem of having the best of both worlds (fast time response with smooth flow). It does a pretty good job, and we have been using it by default for about eight years. In this mode, our flowmeter will consider the actual instantaneous values compared to its filtered value and decide to respond quickly or instead slowly to smooth the flow. It works well, but in those edge cases where you want perfect filtering or perfect response time, it is better to go with the static time constant.

As always, we would love to speak with you. Call us if you would like more information about our flowmeter products!