Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.632838
Title: A study of the mechanics of current meters under steady and dynamic flow conditions
Author: Jepson, P
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 1965
Availability of Full Text:
Access through EThOS:
Abstract:
Since current meters are generally calibrated by towing in still water, their ability to measure the mean flow velocity under the sometimes totally different field conditions has come under considerable question. In an effort to answer some of the questions posed the work described in this thesis was undertaken. The analytical and experimental studies described are divided into three main chapters. (1) An Investigation into the Principles of Current Meters under Steady Flow Conditions. The analysis attempted in this chapter has been carried out because the existing theories appeared inadequate, and totally unconfirmed experimentally. A fundamental study is therefore made of the forces experienced by such meters, and a theory developed and used, in conjunction with some unique experimental measurements, to predict quantitatively the shape of the steady state calibration curves. The agreement between predicted and measured calibration curves for the three OTT meters tested is generally good. The theory is then further developed to show that temperature changes and symmetrical velocity gradients can cause registration errors, but under normal working conditions it is shown that these are unlikely to be excessively large. It is also noted with interest that asymmetrical velocity profiles can cause this type of meter to overestimate the mean flow velocity. (2) The Errors of Current Meters in Pulsating Flow. A meter situated in a stream can be exposed to fluctuating velocities both parallel and transverse to the meter axis, and since such flow conditions contravene the calibration conditions the question arises as to the accuracy of such measurements. It is shown in the initial experiments of this chapter that sinusoidal flow fluctuations parallel to the mean velocity vector cause the meter to over-register the mean liquid velocity, and transverse fluctuations generally cause the meter to underestimate. From the results of these tests general suggestions are made concerning the choice of meter design and station in order that such errors can be minimised. A theoretical attempt at predicting the measured errors is also undertaken, and it is found, using simple aerodynamic principles, that all the trends experienced by the meter when subjected to axially pulsating flow could be estimated. The agreement between this completely theoretical approach and the experiments is found to be good; and from this investigation further comment is made about the optimisation of the meter design. However it is found, after much work, that this and a subsequent approach somewhat oversimplifies the problem, since it is shown using a more rigorous approach, along similar lines to the initial theory, that the quantitative agreement with the measured error under axially pulsating conditions is not as good as that previously found. In this rigorous approach the equation of motion is formed using experimentally measured torques, and the overestimation obtained by solving this equation on an Analogue Computer. It was thought that this deviation, between experiment and theory, could have been caused either by the viscous effects of the liquid or by its inertia, of which no account had been taken in the theory. In order to minimise these effects similar tests were conducted in air flow, and the agreement between measured and estimated errors, under axially pulsating conditions was found to be surprisingly good; which partially proved the hypothesis. A rough estimation is also made at predicting the error due to transverse pulsations; again using the Analogue Computer to solve the relevant differential equation. It is conclusively shown in the appendices of this chapter that current meters tend to overestimate more when the axial flow fluctuations contain large negative accelerations; and also that the measured error is unaffected by prolonged vibration over quite considerable lengths of time. It would appear for this work that, for most practical cases of flow measurement where no large amplitudes and high frequency pulsations are present, the current meter system is surprisingly accurate. Indeed it is shown that a sinusoidal axial fluctuation in flow of;l; 101'0 about the mean velocity can only cause a maximum overestimation of ~. However, if the fluctuations are large then it is shown that a rough estimation of the % error incurred can be made if the approximate mean velocity, size and shape of the fluctuations, and the rotor torque characteristics are known. (3) The Transient Response of Current Meters and Turbine Type Flowmeters. Because of the unusual results of the preceding chapter it was thought tbat a more fundamental study should be mode into dynamic effects on current meters. It was decided to study the response of propeller type flowmeters to rising step-inputs since this appeared to be the simplest way of subjecting the meter to dynamic flow conditions. There was also a need for such an analysis because of its direct application in the mechanics of rocket propulsion servo-systems, where propeller flowmeters are used. In this work, using some novel experimental techniques, it is shown that a theory which neglects the fluid inertia can only be used accurately to predict the response times of such meters under certain limited conditions. It is again found that the response under gaseous flow conditions (where the fluid inertia term is negligible compared with th~t of the rotor) can be fully explained; and with liquid flow the inertia term causes the theory to underestimate the actual time constant at a given flow velocity. The amount or deviation from the theory is shown to depend on the dimensions and density of the rotor and on the fluid density. In the appendices of this chapter it is shown that the effect of symmetrical velocity profiles on the time constant is, for most practical cases, unlikely to be excessive. A possible system for measuring the rotor revolutions is also discussed, which seems to have a number of advantages over the mechanism now used in laboratory current meters. From the results of these three chapters general conclusions are drawn and a brief outline is given of a proposed future research programme.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.632838  DOI: Not available
Share: