The enlargement of the impeller throttling gap width (W) by wear causes an increase of the throttling gap flow. Consequently, there is an alteration of the pump characteristic curves; the questions, how do they alter and to what extent, then arise. Part Two of this article describes the testing methods used to investigate the impact the enlarged W has on characteristic curves, as well as the test results for each of the test pumps.
By Jurgen H. Timcke
While tests have been done to show the result of these alterations for a multistage centrifugal pump with a balance piston, they did not allow for a statement regarding the influence of the enlarged W on the characteristic curves of a single stage volute casing pump with axial inlet (hereby referred to as volute casing pump), nor the influence of the specific speed (nq). Such information, when referring to a particular case, must be dealt with care, as it can very easily be misinterpreted.
All three test pumps were measured at n = 2900 [1/min]. To avoid the influence of scatter of the measured H- and P-values (seen in Figure 2) on the Q-η curves, they were not calculated with the H- or P- values, instead they use the required numerical values taken from the drawn Q-H- and Q-Pcurves. To get a better overview, the measuring points themselves are not indicated in the diagram. Measurements to determine the Q-NPSH3% curves at an enlarged W were not carried out, as NPSH3% is not always the decisive criterion for the operation reliability, but instead the altered operating point concerning Q and H. The influence of enlarged W on the suction performance is given in detail in Reference 1.
In the last test (without a test number and presented with broken lines in the diagrams, see Figure 8) the objective was to turn down the impeller running rings of test 5 completely until the impeller inlet diameter and vane tips were at the inlet angle β1a. Occasionally, the remaining vane tips were not removed.
Eccentric position alteration of the impeller running rings caused by shaft deflection were not considered as the setting of the eccentricity ‘e’ of the impeller running rings during the tests may not be measured with acceptable technical expenditure or may be ignored at turbulent flow in the throttling gaps. At tests 1 through 5 at all test pumps always existing. The influence quantities included: impeller weight, size of the radial force (amongst other facts dependent on the position of the Q-measuring points referred to QBEP), diameter and material of the shaft, and bearing ratio distance.
Figure 8 shows how the characteristic curves are altered with an increasing of W.
Independent of size and nq altering the Q-H-curves as W increases, test 2 to 5 curves were a decreasing equidistant to those at the starting state of test 1. At the starting point, the test pumps were ‘taken from stock,’ meaning no special treatment was given before testing.
Dependent on nq showing the Q-Pcurves of the test 2 to 5, compared with the starting state, test 1, with increasing of W, they experienced a conspicuous phenomenon; they turn themselves about a ‘point’ (figure 8, black triangles with tip to the top) which is positioned at:
- < low nq (TP1) on the right of QBEP
- < medium nq (TP2) about at QBEP
- < high nq (TP3) on the left of QBEP
On the left of this ‘point’, the Q-Pcurves always rise while at the right of it, dependent on nq, they drop increasingly more and more.
The best efficiency point, ηBEP, drops with increasing of W. Even when considering that the unavoidable scatters shows that the test results of this investigation are a definite deviation from the theoretical statement explained in Reference 2. Simply put: increasing nq less ηBEP – drop.
The numerical values, seen in Chart 1, based on the test results, do not show a decrease of ΔηBEP [%] = f (nq), but referred to the nq – range of the test pumps, the contrary. At this point in the test, both sides the impeller running rings do not alter the existing QBEP. This can be observed at all test pumps and tests (see figure 8). Measuring inaccuracies or errors of measurements as a cause can be ruled out. However, if the impeller running rings on both sides are completely turned down, then the result is an nq-dependent reduction of QBEP: ≈ 7,4 [%] (TP1), ≈ 15,2 [%] (TP2) and at TP3 ≈ 14,3 [%].