1、37.1CHAPTER 37MEASUREMENT AND INSTRUMENTSTerminology . 37.1Uncertainty Analysis 37.3Temperature Measurement 37.4Humidity Measurement 37.10Pressure Measurement. 37.13Air Velocity Measurement 37.15Flow Rate Measurement 37.20Air Infiltration, Airtightness, and Outdoor Air Ventilation Rate Measurement.
2、37.24Carbon Dioxide Measurement. 37.25Electric Measurement 37.27Rotative Speed and Position Measurement 37.28Sound and Vibration Measurement 37.29Lighting Measurement . 37.31Thermal Comfort Measurement . 37.31Moisture Content and Transfer Measurement 37.32Heat Transfer Through Building Materials . 3
3、7.34Air Contaminant Measurement 37.35Combustion Analysis 37.35Data Acquisition and Recording 37.35Mechanical Power Measurement. 37.37VAC engineers and technicians require instruments for bothH laboratory work and fieldwork. Precision is more essential inthe laboratory, where research and development
4、 are undertaken, thanin the field, where acceptance and adjustment tests are conducted.This chapter describes the characteristics and uses of some of theseinstruments.1. TERMINOLOGYThe following definitions are generally accepted.Accuracy. Ability of an instrument to indicate the true value ofmeasur
5、ed quantity. This is often confused with inaccuracy, which isthe departure from the true value to which all causes of error (e.g.,hysteresis, nonlinearity, drift, temperature effect) contribute.Amplitude. Magnitude of variation from its equilibrium or aver-age value in an alternating quantity.Averag
6、e. Sum of a number of values divided by the number ofvalues.Bandwidth. Range of frequencies over which a given device isdesigned to operate within specified limits.Bias. Tendency of an estimate to deviate in one direction from atrue value (a systematic error).Calibration. (1) Process of comparing a
7、set of discrete magni-tudes or the characteristic curve of a continuously varying magni-tude with another set or curve previously established as a standard.Deviation between indicated values and their corresponding stan-dard values constitutes the correction (or calibration curve) for infer-ring tru
8、e magnitude from indicated magnitude thereafter; (2) processof adjusting an instrument to fix, reduce, or eliminate the deviationdefined in (1). Calibration reduces bias (systematic) errors.Calibration curve. (1) Path or locus of a point that moves so thatits graphed coordinates correspond to values
9、 of input signals andoutput deflections; (2) plot of error versus input (or output).Confidence. Degree to which a statement (measurement) isbelieved to be true.Dead band. Range of values of the measured variable to whichan instrument will not effectively respond. The effect of dead band issimilar to
10、 hysteresis, as shown in Figure 1.Deviate. Any item of a statistical distribution that differs from theselected measure of control tendency (average, median, mode).Deviation. Difference between a single measured value and themean (average) value of a population or sample.Diameter, equivalent. The di
11、ameter of a circle having the samearea as the rectangular flow channel cross section.Deviation, standard. Square root of the average of the squares ofthe deviations from the mean (root mean square deviation). A mea-sure of dispersion of a population.Distortion. Unwanted change in wave form. Principa
12、l forms ofdistortion are inherent nonlinearity of the device, nonuniformresponse at different frequencies, and lack of constant proportional-ity between phase-shift and frequency. (A wanted or intentionalchange might be identical, but it is called modulation.)Drift. Gradual, undesired change in outp
13、ut over a period of timethat is unrelated to input, environment, or load. Drift is gradual; ifvariation is rapid and recurrent, with elements of both increasing anddecreasing output, the fluctuation is referred to as cycling.Dynamic error band. Spread or band of output-amplitude devi-ation incurred
14、by a constant-amplitude sine wave as its frequency isvaried over a specified portion of the frequency spectrum (see Staticerror band).Emissivity. Ratio of the amount of radiation emitted by a real sur-face to that of an ideal (blackbody) emitter at the same temperature.Error. Difference between the
15、true or actual value to be measured(input signal) and the indicated value (output) from the measuringsystem. Errors can be systematic or random.Error, accuracy. See Error, systematic.Error, fixed. See Error, systematic.Error, instrument. Error of an instruments measured value thatincludes random or
16、systematic errors.Error, precision. See Error, random.Error, probable. Error with a 50% or higher chance of occur-rence. A statement of probable error is of little value.Error, random. Statistical error caused by chance and not recur-ring. This term is a general category for errors that can take val
17、ues oneither side of an average value. To describe a random error, its dis-tribution must be known.Error, root mean square (RMS). Accuracy statement of a sys-tem comprising several items. For example, a laboratory potentiom-eter, volt box, null detector, and reference voltage source haveindividual a
18、ccuracy statements assigned to them. These errors aregenerally independent of one another, so a system of these units dis-plays an accuracy given by the square root of the sum of the squaresof the individual limits of error. For example, four individual errorsof 0.1% could yield a calibrated error o
19、f 0.4% but an RMS error ofonly 0.2%.Error, systematic. Persistent error not caused by chance; system-atic errors are causal. It is likely to have the same magnitude and signfor every instrument constructed with the same components andThe preparation of this chapter is assigned to TC 1.2, Instruments
20、 andMeasurements.37.2 2017 ASHRAE HandbookFundamentalsprocedures. Errors in calibrating equipment cause systematic errorsbecause all instruments calibrated are biased in the direction of thecalibrating equipment error. Voltage and resistance drifts over timeare generally in one direction and are cla
21、ssed as systematic errors.Frequency response (flat). Portion of the frequency spectrumover which the measuring system has a constant value of amplituderesponse and a constant value of time lag. Input signals that havefrequency components within this range are indicated by the mea-suring system (with
22、out distortion).Hydraulic diameter Dh. Defined as 4Ac/Pwet, where Acis flowcross-sectional area and Pwetis the wetted perimeter (perimeter incontact with the flowing fluid). For a rectangular duct with dimen-sions W H, the hydraulic diameter is Dh= 2HW/(H + W ). Therelated quantity effective or equi
23、valent diameter is defined as thediameter of a circular tube having the same cross-sectional area asthe actual flow channel. For a rectangular flow channel, the effectivediameter is Deff= .Hysteresis. Summation of all effects, under constant environ-mental conditions, that cause an instruments outpu
24、t to assumedifferent values at a given stimulus point when that point is ap-proached with increasing or decreasing stimulus. Hysteresis in-cludes backlash. It is usually measured as a percent of full scalewhen input varies over the full increasing and decreasing range. Ininstrumentation, hysteresis
25、and dead band exhibit similar outputerror behavior in relation to input, as shown in Figure 1.Linearity. The degree of straightness of the transfer curvebetween an input and an output (e.g., the ideal line in Figure 1); thatcondition prevailing when output is directly proportional to input(see Nonli
26、nearity). Note that the generic term linearity does notconsider any parallel offset of the straight-line calibration curve.Loading error. Loss of output signal from a device caused by acurrent drawn from its output. It increases the voltage drop acrossthe internal impedance, where no voltage drop is
27、 desired.Mean. See Average.Median. Middle value in a distribution, above and below whichlie an equal number of values.Mode. Value in a distribution that occurs most frequently.Noise. Any unwanted disturbance or spurious signal that modi-fies the transmission, measurement, or recording of desired dat
28、a.Nonlinearity. Prevailing condition (and the extent of its mea-surement) under which the input/output relationship (known as theinput/output curve, transfer characteristic, calibration curve, or re-sponse curve) fails to be a straight line. Nonlinearity is measuredand reported in several ways, and
29、the way, along with the magni-tude, must be stated in any specification.Minimum-deviation-based nonlinearity: maximum departurebetween the calibration curve and a straight line drawn to give thegreatest accuracy; expressed as a percent of full-scale deflection.Slope-based nonlinearity: ratio of maxi
30、mum slope error any-where on the calibration curve to the slope of the nominal sensitivityline; usually expressed as a percent of nominal slope.Most other variations result from the many ways in which thestraight line can be arbitrarily drawn. All are valid as long as con-struction of the straight l
31、ine is explicit.Population. Group of individual persons, objects, or items fromwhich samples may be taken for statistical measurement.Precision. Repeatability of measurements of the same quantityunder the same conditions; not a measure of absolute accuracy. Itdescribes the relative tightness of the
32、distribution of measurementsof a quantity about their mean value. Therefore, precision of a mea-surement is associated more with its repeatability than its accuracy.It combines uncertainty caused by random differences in a numberof identical measurements and the smallest readable increment ofthe sca
33、le or chart. Precision is given in terms of deviation from amean value.Primary calibration. Calibration procedure in which the instru-ment output is observed and recorded while the input stimulus isapplied under precise conditions, usually from a primary externalstandard traceable directly to the Na
34、tional Institute of Standards andTechnology (NIST) or to an equivalent international standards orga-nization.Range. Statement of upper and lower limits between which aninstruments input can be received and for which the instrument iscalibrated.Reliability. Probability that an instruments precision a
35、nd accu-racy will continue to fall within specified limits.Repeatability. See Precision.Reproducibility. In instrumentation, the closeness of agreementamong repeated measurements of the output for the same value ofinput made under the same operating conditions over a period oftime, approaching from
36、both directions; it is usually measured as anonreproducibility and expressed as reproducibility in percent ofspan for a specified time period. Normally, this implies a longperiod of time, but under certain conditions, the period may be ashort time so that drift is not included. Reproducibility inclu
37、desFig. 1 Measurement and Instrument Terminology4HW Measurement and Instruments 37.3hysteresis, dead band, drift, and repeatability. Between repeatedmeasurements, the input may vary over the range, and operatingconditions may vary within normal limits.Resolution. Smallest change in input that produc
38、es a detectablechange in instrument output. Resolution, unlike precision, is a psy-chophysical term referring to the smallest increment of humanlyperceptible output (rated in terms of the corresponding increment ofinput). The precision, resolution, or both may be better than theaccuracy. An ordinary
39、 six-digit instrument has a resolution of onepart per million (ppm) of full scale; however, it is possible that theaccuracy is no better than 25 ppm (0.0025%). Note that the practicalresolution of an instrument cannot be any better than the resolutionof the indicator or detector, whether internal or
40、 external.Sensitivity. Slope of a calibration curve relating input signal tooutput, as shown in Figure 1. For linear instruments, sensitivity rep-resents the change in output for a unit change in the input.Sensitivity error. Maximum error in sensitivity displayed as aresult of the changes in the cal
41、ibration curve resulting from accu-mulated effects of systematic and random errors.Stability. (1) Independence or freedom from changes in onequantity as the result of a change in another; (2) absence of drift.Static error band. (1) Spread of error present if the indicator(pen, needle) stopped at som
42、e value (e.g., at one-half of full scale),normally reported as a percent of full scale; (2) specification or rat-ing of maximum departure from the point where the indicator mustbe when an on-scale signal is stopped and held at a given signallevel. This definition stipulates that the stopped position
43、 can beapproached from either direction in following any random wave-form. Therefore, it is a quantity that includes hysteresis and nonlin-earity but excludes items such as chart paper accuracy or electricaldrift (see Dynamic error band).Step-function response. Characteristic curve or output plotted
44、against time resulting from the input application of a step function(a function that is zero for all values of time before a certain instant,and a constant for all values of time thereafter).Threshold. Smallest stimulus or signal that results in a detect-able output. Resolution and threshold are som
45、etimes used inter-changeably.Time constant. Time required for an exponential quantity tochange by an amount equal to 0.632 times the total change requiredto reach steady state for first-order systems.Transducer. Device for translating the changing magnitude ofone kind of quantity into corresponding
46、changes of another kind ofquantity. The second quantity often has dimensions different fromthe first and serves as the source of a useful signal. The first quantitymay be considered an input and the second an output. Significantenergy may or may not transfer from the transducers input to output.Unce
47、rtainty. An estimated value for the bound on the error (i.e.,what an error might be if it were measured by calibration). Althoughuncertainty may be the result of both systematic and precisionerrors, only precision error can be treated by statistical methods.Uncertainty may be either absolute (expres
48、sed in the units of themeasured variable) or relative (absolute uncertainty divided by themeasured value; commonly expressed in percent).Zero shift. Drift in the zero indication of an instrument withoutany static change in the measured variable.2. UNCERTAINTY ANALYSISUncertainty SourcesMeasurement g
49、enerally consists of a sequence of operations orsteps. Virtually every step introduces a conceivable source of uncer-tainty, the effect of which must be assessed. The following list isrepresentative of the most common, but not all, sources of uncer-tainty.Inaccuracy in the mathematical model that describes the physicalquantityInherent stochastic variability of the measurement processUncertainties in measurement standards and calibrated instru-mentationTime-dependent instabilities caused by gradual changes in stan-dards and instrumentationEffects of environmental factors such as
