Pumps play a significant role in industrial plants and need continuous monitoring to minimise loss of production. To date, there is limited published information on the application of Acoustic Emission (AE) to incipient pump cavitation. This paper presents a case study where AE has been applied for detecting incipient cavitation and determining the best efficiency point (BEP) of a 60KW centrifugal pump. Results presented are based on NPSH (Net Positive Suction Head) and performance tests. In conclusion the AE technique was shown to offer early detection of incipient cavitation, furthermore, the technique has demonstrated the ability to determine the BEP of a pump.

Adhesion and other mechanical behaviour of coatings is at present routinely tested in industry and research organisations using the scratch test. The test has been the subject of a European Standard prEN 1071-3:2000 established by the European Standards Committee CEN TC184 WG5. A previous European project on the development and validation of test methods for thin hard coatings-FASTE-revealed that uncertainties in the Rockwell C stylus tip shape represent a major source of error for the scratchtest method. Therefore a follow-up project-REMAST-was started to develop and certify a reference material as a quality control tool for the testing and qualification of scratch-test instruments. In addition, a considerable effort was devoted to improve the scratch stylus manufacturing process. This paper presents the results of the REMAST project. During a feasibility study, two candidate materials-titanium nitride (TiN) and diamond-like carbon (DLC) coatings both on high speed steel substrates-were evaluated. Because of the lower sensitivity to styli tip shape variations and higher data scatter observed for the TiN coating, DLC was chosen as the reference material to be certified. One thousand samples were produced, as well as qualified styli to carry out the certification campaign, including homogeneity and stability testing. Nine independent laboratories were involved in the certification exercise, and the data were statistically analysed to obtain the certified critical load values and their uncertainty ranges. A certified reference material BCR-692 is available for verification purposes. This presents three repeatable failure events at certified critical load intervals and can provide a good indication of overall scratch-test instrument performance, including stylus condition and calibration. It is also useful as a diagnostic tool, providing a means of sensitive monitoring of machine and stylus performance over extended periods. Considerable improvement of the quality of scratch styli was achieved by the strict control of all manufacturing steps.

A review of the available methods used to predict the closed valve head of a centrifugal pump is presented. A database of pump geometry was assembled and used to assess the methods of prediction for accuracy on the basis of their specific speed and configuration. The methods based only on Euler correction factors are found to be the least accurate and give an indication of the possible closed valve head value within 20 per cent. Further methods linked to the geometry through empirical coefficients prove more successful. Correction coefficients are suggested that improve the accuracy of the prediction methods for the analysed pumps within the database.

Predicting the head of a centrifugal pump operating at closed valve remains a difficult task. The nature of the flow regime and the influence of geometric features on this flow is uncertain. In this paper both the flow regime and the influence on that regime by geometry is investigated using a commercially available CFD RANS code. A CFD methodology is presented that takes account of the difficult boundary conditions. This methodology is then used to present the flow regime in a volute pump against the background of available research. The theory of solid body rotation, as a major influence on the closed valve head, is shown by the CFD simulations to be analogous to but not representative of the actual flow regime. The nature of the flow is impulsive and unsteady with fluid interchange occurring between the pump collector and the impeller vane passages. At the inlet the pump impeller experiences a strong steady outflow from the impeller blade tip. This spiralling flow must be accommodated within the computational solution. The impeller outlet is filled with a vortex driven by the flow which cannot be accommodated within the stalled stator passageways. The annular gap between the impeller and collector is filled with a pulsating flow whose frequency is determined by the number of vanes within the impeller. Model validation was carried out by reference to experimental papers and time averaged closed valve head values obtained under standard performance testing.

No abstract