If an instrument has an accuracy claim of 0.5% of full scale, for instance, you should recognize that the actual accuracy diminishes as the operating conditions fall below the full-scale setting. Sometimes, though, you will hear claims that a meter offers 0.5% of reading over full range. Although the difference may sound insignificant, it could be very costly to the owner. Imagine, for instance, that a paddle-wheel flowmeter claims to have an accuracy of ±0.5%. Suppose, further, it is a percent of full range, and the full range is 50 feet per second (ft/sec).

If the flow range you will use it in is 6 ft/sec, which is common in treatment plants, the actual accuracy is much different than you might expect: 0.005 x 50 ft/sec = ±0.25 ft/sec .If you apply this accuracy against a flow rate of 6 ft/sec, you see that the actual accuracy is: ±0.25/6 ft/sec = ±0.0417, or 4.17% . Comparing a magnetic flowmeter with an accuracy of 0.5% of reading to a Doppler flowmeter with an accuracy of 0.5% of full range yields a similar result.

A common problem occurs when a city or municipality uses two different types of flowmeters. Imagine one meter is a highly accurate magnetic flowmeter located in a meter vault to monitor the plant’s effluent flow, and the other is a Doppler meter monitoring the influent flow; this meter’s accuracy diminishes as the flow ate drops. Case histories have shown that the plant appears to be either generating wastewater, because the effluent is more than the influent, or something is evaporating the wastewater.

We know in both cases that neither of these conditions really exists. What is really happening is that the Doppler meter is not matching the accuracy of the magnetic meter. The difference between 0.5% of 12 million gallons a day (Mgd) and 4.17% of 12 Mgd is substantial: 4.17% – 0.5% x 12 Mgd = 0.44 Mgd, or 305 gal/min .Matters are made even worse if the Doppler meter is used for pacing chemical feed into the wastewater with the same inaccuracies, resulting in either overdosing or underdosing.

Water treatment plants have low, average daily, and high peak demand flows, and further, low and average daily flows occur more frequently. This demonstrates the importance of being cautious in choosing meter types for those flow variables. Many types of flowmeters suffer in performance as the flows decrease and approach the lower end of their viable flow range; therefore, pacing during low flow periods may be highly suspect. Chemicals are becoming more costly, analytical instruments for measuring the effects of these chemicals are becoming costly, and corrosion due to underdosing or overdosing wastewater can be costly to equipment. All of these may contribute to effluent that is a danger to wildlife and, in extended cases, human life.

Repeatability

Another tool in evaluating equipment is repeatability, defined as the quantity that characterizes the ability of an instrument to give identical indications or responses for repeated applications of the same value of the quantity measured under the same conditions of use. In the past, when equipment operated on motion balance, where equipment used linkages and temperature compensation values, repeatability was critical. Today, however, a number of field instruments work on force balance techniques, such as piezoelectric crystals, capacitance, and strain gauges.

These all work on the principle that if you put a force on an instrument, there should be no motion, though an electric signal is generated on the output of that instrument. There are still flow, level, and chemical measuring devices that do not work on the force balance principle, and for these types, looking at the repeatability of that piece of equipment is still important. A steady widening of the repeatability is an indication that something is going wrong with the instrument.

Although some might believe good repeatability is a measure of accuracy, that is incorrect. To understand the difference between accuracy and repeatability, imagine an archer shooting at a conventional archery target. Suppose one archer hits the bulls-eye consistently. Because he was always accurate, the shots were repeatable. Now imagine an archer that hits the target but misses the bulls-eye consistently. Although the archer has good repeatability, the archer was not accurate. Good repeatability does not guarantee accuracy. If you do not see a proper accuracy statement on equipment but only a repeatability statement, be cautious.

Rangeability and uncertainty

One of the most common problems with instrumentation equipment is the exaggeration of its range. How many times have you heard a meter can read flow rates at velocities of 1–100 ft/sec, giving the impression that you can read flows accurately through that total velocity range? What often goes unmentioned is that the particular meter’s accuracy has a 10:1 turndown ratio. This means that a meter sized to measure a range of 0–30 Mgd has a true accuracy over the full range 3–30 Mgd. Below 3 Mgd, the meter accuracy diminishes. Additionally, different types of meters have different turndown ratios over their full range.

It is common for a Venturi tube, for example, to have two transmitters measuring the flow. This is because a Venturi tube with one transmitter measures accurately with a 6:1 turndown ratio over the full range. So if we look at a range of 0–30Mgd, the meter’s accuracy diminishes below 5 Mgd. The range over which the instrument meets the stated linearity of uncertainty requirements is its “rangeability.” Uncertainty is the range of values within which the true value lies with a specified probability. Uncertainty of ±1% at 95% confidence means the instrument will give the user a range of ±1% for 95 readings out of 100.

Another common error occurs during the equipment sizing. In the water reclamation industry, it is a common practice to assume that solids in wastewater will settle out around a velocity of 2 ft/sec. A magnetic flowmeter reads accurately if the minimum velocity is above 2 ft/ sec, but below this, settling is likely to occur— and who can then say what the accuracy really is?

Typically, designers size plants to handle increased flow capacities for 20 years. For this reason, designers often oversize pipes for early life-cycle flow, and there is corresponding settlement inside the pipe. This settling can also occur in the inner liner of the meters. Because these meters are velocity sensing devices with an assumed constant cross section, they will give a false reading if the inner liner becomes coated with sludge. A solution may be to reduce the size of the meter to increase velocity by utilizing a pipe reducer on the inlet side and a pipe expansion section on the discharge side of the meter.

If possible, avoid connecting the reducer and expander directly onto the meter. Manufacturers recommend that when you reduce the pipe, the flowmeter has a minimum of six to 10 pipe diameters upstream from an elbow or valve and at least two pipe diameters downstream of a pipe elbow or valve. This provides a less distorted flow profile for the meter to read. Be certain you can afford to lose the pressure head when you reduce the meter. Maximum velocities should not exceed 15 ft/sec. By maintaining a minimum scouring effect inside the pipe, your sludge buildup inside pipes and any in-line equipment will diminish, helping avoid measurement errors and costly maintenance downtime.

Misconceptions and truths

Some people will ask for the accuracy of a certain flowmeter, level, or pressure-measuring device and, upon hearing a low number, think that everything involved with the flowmeter will be of the same accuracy. However, the meter accuracy is not the accuracy for the entire flow system. A mathematical equation known as the root mean square (RMS) correctly determines the accuracy of the complete system.

Consider the case of a magnetic flowmeter that records flow locally, sending an analog signal to an operator’s workstation via a programmable logic controller (PLC). You must look at each component’s accuracy: a magnetic flowmeter (±0.5%), a magnetic flowmeter transmitter (±0.5%), a wire connection to the recorder (±0.01%), a wire connection to a local control panel terminal block (±0.01%),and the I/O card of the PLC (0.4%). Each component in the system has its own measurement errors and uncertainties, which contribute to the overall accuracy of the complete system. In real cases, there could be more components attached to a control system. To use the RMS method, first square each number, yielding 0.000025, 0.000025,0.00000001, 0.00000001, and 0.000016.

Second, add the numbers. Then find the square root of the sum. The entire system has an accuracy of approximately ±0.813% instead of 0.5%. This accuracy equation works for any individual chemical, pressure, level, temperature, or flow loop. Remember, too, that no two flowmeters or instruments will have exactly the same accuracy. For this reason, the accuracy statement should indicate a ± component. When choosing an instrumentation control strategy, look at all the manufacturers’ equipment literature regarding accuracy; consider the range, repeatability, turndown ratio, and piping constraints; choose similar equipment types; and utilize the RMS equation.

This article was written and submitted by

John Davis is a senior member of ISA who operates Davis
Instrumentation Services. His address is nira47@aol.com.