Ultra Sonc Flow Rate Measurements n Large Ppes and open Channels Dr. Mchael Teufel I Summary Correlaton s a mathematcal tool whch, enabled by developments of fast and powerful mcroprocessors n last few years, has found ts way nto the sgnal evaluaton n the feld of flow rate measurement. Wthn the varous flow measurement methods t s especally appled as a non contactng technology n multphase flows. In ths connecton the correlaton s used wth varous sgnal transmtters, ths report s confned to an acoustc sgnal transmtter and a recever though. To portray the correlaton n smplfed terms t can be consdered as the consstent development of smple CW Doppler systems and Pulse Doppler systems. To pont out the dfferences these technologes wll be dscussed as well. The acoustc Doppler effect s a well known to be used for velocty or flow measurement n fluds and gases. Frst applcatons n ths feld worked wth contnuously operatng sound generators and separate recevers (CW - Doppler) determnng the Doppler shfted frequency of scatters n the lqud. The velocty of the reflectng partcles n the flud (gas bubbles, sold substances or smlar) can be calculated from the dfference between transmtted and receved frequency. Intellgent evaluaton procedures enable plausble profle assumpton, so the Doppler velocty measurement method can be used as flow measurement n connecton wth level measurement. The Pulse Doppler method represents a further development of the contnuous Doppler method and provdes an addtonally spatal resoluton of the measured partcle velocty. Ths postonng of a spatal measurement wndow s ganed from measurng the transt tme of a short ultrasonc pulse. So the maxmum velocty s measured that s converted nto the mean velocty and usng ths, the flow rate s calculated. Addtonally the pulse method allows the transmtter and recever to be dentcal. The cross correlaton method provdes much hgher accuracy for the velocty measurement due to the turbulent velocty fluctuatons a further mprovement wll not offer addtonal nformaton. The spatal resoluton s more than 0 tmes lower than for a Pulse Doppler system. Though the current velocty profle can be determned. In combnaton wth an ntegrated level measurement an accurate calculaton nto the flow rate s feasble. Ths opens up measurement ranges n partally and full flled ppes wth large dameters and n applcatons wth open channels and flumes. Such applcatons could not be covered suffcently n the past or caused extremely hgh expenses.
II I II Table of Contents Summary Table of Contents.) Introducton.) Doppler Method 3.) Impulse Doppler 4.) Cross Correlaton 5.) Applcatons.) Introducton Flow measurement n full flled ppes mostly attans very hgh accuracy, partcularly n ppes wth dameters < 00 mm. Many supplers and many varous measurement methods normally guarantee nexpensve measurement devces here. The stuaton changes as soon as large partally or full flled ppes and open channels wth rregular velocty profles, backwater, low fllng levels and low flow veloctes are consdered. Here generally the attaned accuracy s not comparable, but nevertheless prces ncrease often very clearly n connecton wth the channel dmensons. It s plan to see the ncreasng need for accurate and nexpensve measurement methods n ths area. Ultrasonc meters for flow measurement have been practcally used for some tme. Intal stages are operated wth a contnuous ultrasound Doppler method whch s used n connecton wth level measurement. Ths procedure s brefly descrbed n chapter. Chapter 3 outlnes mprovements whch come n use especally n full flled ppes. Here, usng the Pulse Doppler method enables the spatal allocaton of measured veloctes. The measurement accuracy s than be mproved by usng the cross correlaton. Ths method s descrbed n chapter 4. By usng the latest developments n the feld of electroncs measurement devces a new sensor could be developed that provdes very hgh accuracy n velocty measurement as well as hgh spatal resoluton. Ths enables a relable velocty profle determnaton. Hence, n combnaton wth an ntegrated level measurement wthn our compact sensor, relable flow rate measurements are feasble. Chapter 5 ponts out several applcatons.
.) Doppler Method When a car passes by t emts a sound sgnal that s changng ts frequency. As long as the car s approachng us, the frequency s hgher as f t s movng away from us. In smple words: f the vehcle comes towards us, the sound waves are "compressed" causng a hgher ptch. If the car moves away the sound waves are stretched causng a lower ptch. Ths frequency shft s a perceptve effect only: for the drver the sound s the same all the tme. Ths effect (descrbed by Chrstan Doppler n 84) s practcally used n the felds of astronomy, medcne and ndustral applcatons. The followng explanatons descrbe how the Doppler effect can be used for flow rate measurements f transmtter and recever are frmly fxed wth a flowng medum and movng and reflectng partcles. As depcted n Fg. there are two pezo crystals ntegrated wthn the Doppler sensor. One of them s operatng as contnuous transmtter, the other one as contnuous recever. Both crystals are acoustcally decoupled aganst each other on ther rears. Towards the medum (due to mechancal and chemcal protecton) they are solated wth a cast resn layer that has good acoustc couplng and transmsson capactes. Sound Velocty c: (at 0 C) Water = 480 m/s Ar = 330 m/s Steel = 5900 m/s Meth. Alc. = 00 m/s Fg. : Doppler Effect n Flow The transmtter crystal contnuously transmts sound waves wth the defned frequency f and a fxed angle α nto the water (contnuous wave operaton). The waves ht partcles or gas bubbles contaned n the medum, whch move through the acoustc area wth the velocty v. The recever crystal measures the Doppler shfted frequency f that s reflected by the partcles accordng to the formula: f f c = c v cosα From ths measured frequency f the frequency shft can be calculated. f = f f v f cosα = c 3
In practce the hgh frequency transmtter sgnal s heterodyned wth the receved frequency, than the envelope frequency s determned. Ths s to attan a hgher resoluton. Example: Transmttng Frequency: Recevng Frequency: Frequency Dfference:.000.000 Hz.000.00 Hz 00 Hz The frequency dfference s a lnear measurement for the partcles velocty n the medum and thus representatve for the flow velocty of the medum tself, f the partcles are small enough. The determned Doppler dfference frequency s converted nto dgtal mpulses. The used perod duraton measurement has, n comparson wth the drect frequency measurement, the advantage to measure even very low frequences (.e. low flow veloctes) quck and relable. Accumulated data are saved and sent to the measurement transmtter s statstcal data evaluaton. Fg. : Frequency Dagram The statstcal data evaluaton subdvdes the entre frequency spectrum nto a number of frequency groups. For each of these groups the frequency of the measured sngle frequences s determned. A defned number of sngle frequency measurements are sorted nto ths set of statstcs. Ths s how the average value of the frequency groups and thus the resultng average velocty can be calculated from the determned frequency group maxmum. Dependng on densty and number of the partcles n the medum as well as the used sound frequency the Doppler method enables ntruson depths from a few mllmetres up to 5 m. Unfortunately t s not possble to poston or to lmt the ntruson depth. Ths may possbly result n reflectng the sound waves from an uneven water surface f number or sze of the partcles n the medum are badly fluctuatng. In ths case the movement 4
of the waves on the water surface may eventually be msnterpreted as partcle velocty. Above all ths procedure s used n sewage treatment plants for nfluent and dscharge measurements, to record recrculaton and return sludge volumes as well as for dscharge measurement and control n storm water and ran retenton basns. Recordng drect or ndrect nfluent n feedng sewer systems or measurements and control n pump statons are wdespread applcatons too. 3.) Pulse Doppler The ultrasonc pulse Doppler method shows a further development of the known Doppler measurement technology descrbed above. Fgure 3 shows the setup. Lke the CW Doppler method (contnuously workng transmtters) the pulse Doppler method emts short ultrasonc frequency bursts. Snce the temperature compensated sound velocty n the medum and the beam angle of the ultrasonc sgnal are known the measurement pont can be determned by precsely controllng the tme duraton between transmsson and recepton. Ths enables a defned measurement wndow n the medum to be assgned to the receved sgnal. As descrbed above the frequency shft of the receved ultrasonc sgnal can be used to calculate the flow rate. Due to the used pulse tact between transmttng pulse and recevng the measured velocty can be located spatally. Echoes from partcles n other areas do not affect the velocty measurement. Fg. 3: Pulse - Doppler 5
Ths measurement method s restrcted to be appled n full flled crcular ppes due to the calculaton procedure t s based on. The pulse Doppler method wth ts spatally defned sngle pont - measurement offers two dfferent optons for evaluaton and converson of the measured flow velocty nto the average flow velocty:. v max (preferred method) The measurement wndow s postoned at the centre of the ppe (rm = 0). v average (more naccurate method, used mostly n very large dameters) Measurement wndow s postoned to the pont of average flow velocty (rm = 0,77 x r) A next measurement pulse s sent out wth an mpulse repeat rate whch results from sgnal ntruson depth, flow velocty and dameter (Table ). The settng s made automatcally, but can be fxed manually too. Velocty Max. Intruson Max. Ppe Ø [DN] v max Max. Ppe Ø [DN] v avg. Repeat Frequency [m/s] Depth at 750 khz / Method / [mm] Method / [mm] [mm] [mm] 0 340 680.800 463 0 70 340.500 97 0 4 85 70 750 5954 Table 4.) Cross Correlaton Correlaton s a mathematcal evaluaton method whch, enabled by the development of fast and powerful mcroprocessors, has found ts way nto the sgnal evaluaton n the feld of flow measurement technology. Wthn the varous flow measurement methods t s especally appled as a noncontactng technology n multphase flows. In connecton wth ultrasonc technology t can be appled to cover measurement ranges n full or partally flled ppes wth large dameters as well as n open channels, whch otherwse cannot be covered suffcently, wth unsatsfactory accuracy only or by causng extremely hgh expenses. Correlaton s the comparson between two ultrasonc records wth a temporary offset. The prerequste for ts use are reflectng partcles (sold and floatng partcles or gas bubbles) n the water. 6
Image Image Memory Correlator Fg. 4: Correlaton These partcles are scanned by ultrasonc pulses wth defned angle and ther echoes are stored as mage or echo patterns. Resultng from the pulses travel tme (or the echoes of these mpulses) a spatally allocated measurement can be derved lke usng the Pulse Doppler method. Wthn about a mllseconds a second scan s taken whose echo pattern s stored as well. A travel tme determnaton ensures the measurement has be taken n the same tme/spatal wndow. The correlaton of stored sgnals enables the determnaton of the tme shft, that s drectly converted nto the flow velocty and the volume flow rate. Sgnal,00,50,00 0,50 Value 0,00-0,50 -,00 -,50 -,00 0 00 00 300 400 500 600 Tme Fg. 5: Ultrasonc Echo Sgnal 7
Dgtalserte Sgnal => Sgnfunkton,50 35,00 30 0,50 5 0,00 0-0,50 0 -,00 5 -,50 0 5 0 5 0 5 Fg. 6: Partal Ultrasonc Echo Sgnals, dgtalzed and correlated Cross correlaton 0,5 0,0 Correlaton value 0,5 0,0 0,05 0,00 -,0-0,8-0,6-0,4-0, 0,0 0, 0,4 0,6 0,8,0 Tme delay Here the basc correlaton equaton Fg. 7: Correlaton Functon ϕ ( τ ) f, g f ( t) g ( t) ϕ T + f, g ( τ ) = lm f ( t) g( t + τ ) T T T dt 8
s dgtally expressed as follows: wth ϕ( T ) = = N = N = f g ( T ) N = N g f g ( T ) N f = f = st dgtal Image g = nd dgtal Image The fgures 5-7 outlne the look of an ultrasonc echo sgnal and how a cross correlaton can be dgtalzed and evaluated by usng a smple sgn functon. Ths results n a maxmum concernng the tme shft whch wll be processed further. Ths measurement setup enables velocty measurements n up to 6 wndows and hence the velocty profle can be determned. Ths can be used to determne the flow e.g. n full flled ppes (by usng the usual average velocty calculaton from the path velocty, as usual done for example for ultrasonc clamp-on measurement methods). It s advantageous to obtan n addton to the average path velocty a velocty profle measurement whch can be used n smple cases to assess the qualty of a measurement place. Further t can be used for measurement value correcton too. Fg. 8: Integrated and/or external Level Measurement 9
By usng the velocty profle from 6 wndows above all flow measurement n partally flled ppes and channels s feasble. To acheve ths, an addtonal level measurement s requred whch can be carred out by an ntegrated or an external sensor. Wth a combnaton sensor n partal fllngs besdes the flow velocty the flow level s determned too. Ths s acheved by usng a water ultrasonc sensor whch determnes the water level by evaluatng the travel tme of an ultrasonc pulse up to the flud surface. Connectng an external level sensor t s possble too and even recommended under certan condtons,.e. wavy surface or a lot of ar bubbles n the water. Dependng on the fllng level the wndows are postoned varably and ther sze s adjusted. At low fllng levels the number of the wndows s decreasng. Consderng the accuracy we can see some errors effectng the total error accordng to the common quadratc error calculaton: f = Tme Meas. = Velocty Meas. = 0,0 % of Measurement Value f = Installaton Angle (5 ) = 0,75 % f3 = Velocty Profles = 0-0 % typcally % f4 = Length of Meas. Path n Velocty Profle = 0-0 % typcally % f5 = Number of Measurement Wndows = / N for 4 Gates 0,5 % f6 = Black Areas = 0-50 % typcally % Total Error (typcally) = f =,96 % 6 = The devce s depcted n fgure 9. In fgure 0 some sensor varatons can be seen whch are to be used n varous applcatons. 0
Fg. 9: OCM Pro Fg. 0: wedge-shaped Sensor Immerson Sensor Ppe Sensor wth Ball Pvot 5.) Applcatons.) Wastewater Treatment Plant: Influent, Dscharge, Influent to Aeraton Basns, Return Sludge, Recrculaton, Surplus Sludge, Dgester Dscharge and Coagulant Dosage.) Statonary measurements at RÜB s 3.) Temporary Measurements n Sewer Systems 4.) Leakage nvestgaton or nvestgaton of extraneous water 5.) Drect feed control 6.) Industral sewer channel systems 7.) Flow rate measurement n chemcal ndustry 8.) Irrgaton systems 9.) Coolng water nfluent, dscharge and crculaton 0.) Sluce stages n rvers.) Hydro power plants, thermal power statons