INSTRUMENT TECH MUST-KNOWS: MEASUREMENT ERROR
To err is human; to forgive, devine. - Alexander Pope
The last PTOA Segment #99 featured the importance of accuracy with respect to process variable measurement.
The accuracy of the measurement read by a Process Operator is impacted by both instrument-related measurement error and on-line process variable measurement error.
Sources of instrument-related measurement error are:
- The Inherent Instrument Measurement Error which will be stated on the manufacturer's product data sheet as a quantifiable number.
- The Instrument Response Measurement Lag which is the amount of time between a changing process variable and the moment that the instrument sensor can actually sense the change.
- Instrument Reliability, aka the tendency of the instrument to continue operating and not malfunction.
Sources of on-line measurement error include:
- Miscalibration of the instrument by an Instrument Tech human being.
- Misread instrument reading by a Process Operator human being.
- Instrument wear and tear.
- Electrical interference causing signal measurement noise.
The information in this PTOA Segment 100 applies to the measurement of any process variable, not just Temperature.
INSTRUMENT-RELATED MEASUREMENT ERROR QUANTIFIED
All instruments rarely yield "true" readings; all measuring instruments have an amount of inherent instrument measurement error.
Focus once more on the radial distance that the blue measurement dots extend from the exact center of the bullseye.
The spray of blue dots from dead center illustrates the range of inherent instrument measurement error with respect to all of the measurements taken by the instrument.
Each instrument manufacturer will define the error of the instrument on the manufacturer's instrument/product data sheet.
Inherent Instrument Measurement Error in terms of Span
Typically the the accuracy of measurement instruments is given in terms of "percent of span."
For example, the manufacturer of a filled system thermometer (such as the bulb-capillary featued in PTOA Segment #98) might state that the newly purchased thermometer has a span of -185°C to 650 °C and an accuracy "plus or minus 1.25 percent of span."
Span is the accurate range of measurement for an instrument.
The inherent instrument measurement error can be converted into an easier format to understand ... °C ... using these three steps:
1.Determine the span of the instrument in total number of degrees:
The span of this thermometer is 650+185 = 835 °C.
2. Convert the percent error into decimal form by dividing the stated percent instrument error by 100:
1.25/100 = 0.0125
3. The inherent instrument measurement error is thus determined by multiplying the stated error in decimal form with the span:
Error in terms of °C = (835 °C)*(0.0125) = 10.4 °C (18.7 °F)
This result means that a temperature measurement of a process stream given as 700 °C (1292 °F) might really be as low as 689.6 °C (1273.3 °F) or as high as 710.4 °C (1310.7°F) .
In some industrial process applications, this stated inherent instrument measurement error would not greatly impact final product quality or process efficiency.
Specified Amount for the Inherent Measurement Error
Sometimes instrument manufacturers state the inherent instrument measurement error as a specific amount.
For example, assume the manufacturer of a liquid-in-glass thermometer states the device has a span of -120 to 320 °C "with an accuracy of plus or minus 1 °C."
That means that a 100 °C true temperature may be measured as 99 °C or 101 °C.
The maximum percent error for this liquid-in-glass thermometer can be determined from the following steps:
The instrument span is determined:
The span of the instrument is 320+120 = 440
The decimal form of the inherent instrument measurement error is determined by dividing the 1 °C stated error by the span.
1/440 = 0.0023
The maximum percent error for this instrument is determined by multiplying the decimal form of the error by 100:
(0.0023) * (100) = 0.23%
In summary, Instrument Technicians need to be alert as to how the manufacturer characterizes the inherent instrument measurement error so that an error comparison between two different instruments is comparing apples to apples.
Always take the time to determine what the stated error range means in the real world application of the instrument.
Be mentally wary of the instrument that has a percent error greater than plus or minus 0.5 (aka .005). Always do the math; given a span of 1000 °C (1800 °F) the statement of accuracy translates into plus or minus 5 °C which is 9 °F. Is that sufficiently accurate for the process?
FYI:
Each instrument in a process variable control loop that helps send a signal to the controller has an associated error.
The collective error of each instrument impacts the ability to successfully control the process variable of concern (temperature, pressure, flow, or level).
INSTRUMENT RESPONSE LAG
How many PTOA Readers and Students have already figured out that the working liquid in a liquid-in-glass thermometer and the working fluid in a volume expansion device (like a bulb and capillary) can only get heated to what is eventually sensed as a measured temperature after the protected sheath, bulb and capillary have been heated?
All that heat transfer via conduction and convection has to happen before the working fluid can expand ... and that takes time.
The amount of time it takes is called the response lag.
Use of a bulb and capillary expansion device would be a poor choice for a process stream temperature measurement service if the process stream had a dynamically changing temperature that needed to be controlled.
The outcome would be a poorly controlled process, mostly due to the response lag of the temperature measurement.
Otherwise stated, when the total sum of instrument-related measurement error comes mostly from the response lag component, the Process Operator might as well chase his/her tail as try to control the process.
INSTRUMENT RELIABILITY
Instrument Reliability is the ability of an instrument to operate and not malfunction.
The reliability of any man-made device is directly related to how well the device was engineered and manufactured.
And the reliability of process control instrumentation is directly related to the expertise of the Engineering and Design firm selected by the Plant Owner that built the processing complex.
The below information will be included in the manufacturer's data sheet which was hopefully scrutinized by the E&D firm prior to instrument selection.
- The functionality of the instrument with respect to how the instrument will be used, the range of operation, accuracy, and power requirements.
- The physical limitations with respect to the instrument's internal and external dimensions, how the instrument will be mounted and connections to the instrument.
- The physical environment that will surround the instrument will impact the materials of construction and whether or not it needs to be encased or otherwise isolated. Typically the NEMA ratings for the instrument will be included in the manufacturer's data sheet. NEMA = National Electrical Manufacturers Association.
Design engineers typically forget about the physical environment in which the instrument will be installed because they work on their isolated piece of the facility design project from the vantage point of their cubicle at the firm.
The possibility that a more hostile processing environment could likely exist in the real world nearby the application of their particular design focus is rarely explored.
SOURCES OF ON-LINE MEASUREMENT ERROR
Once the measuring instrument is installed, sources of on-line measurement error also contribute to total measurement error.
One source of on-line measurement error that is caused by human beings is the miscalibration of an instrument by an Instrument Technician.
The Instrument Tech that performs field calibration of an instrument must always read-for-understanding the calibration instructions provided by the manufacturer.
Another source of on-line measurement error caused by human frailty is misreading the measurement.
For example, the response from some measurement devices is not directly linear, a subject that will be detailed in future PTOA Segments.
When the recorded measurements from these devices are shown on strip chart recorders, the measurement readings on the low end of the scale range are more difficult to clearly read.
PTOA Readers and Students are already aware that the instrument can malfunction due to normal wear and tear and/or ... and, yes improper manhandling.
When an Outside Process Operator overpressures a pressure gauge, the designed elasticity of the bourdon tube is ruined and the gauge will never again measure a pressure accurately.
The very hardest source of error to isolate and eliminate is noise due to an undesirable additional electrical interference.
For example, all the wires crowded into a conduit will interfere with each other without proper insulation and grounding.
TAKE HOME MESSAGES: Total measurement error has contributions from instrument-related measurment error and on-line measurement error.
The contributions to instrument-related measurement error are:
- inherent instrument measurement error
- instrument response lag
- instrument reliability
Whether or not the range of inherent instrument measurement error for an instrument will work in the instrument's intended service application can be quantified using the stated instrument error and span.
Instrument response lag exists because heat must be transferred into a temperature measuring device before it can sense and measure temperature.
Instrument reliability of an instrument is improved when its functionality, physical limitations, and physical environment are all taken into consideration prior to purchase.
On-line contributions to measurement error include miscalibration of the instrument and misreading of measurements, both of which can be reduced by training and awareness.
Normal instrument wear and tear and electrical interference/noise also contribute to on-line measurement error.
©2016 PTOA Segment 00100
PTOA Process Variable Temperature Focus Study Area
PTOA Process Industry Automation Focus Study Area
You need to login or register to bookmark/favorite this content.