Electricity + Control January 2018

DRIVES, MOTORS + SWITCHGEAR

for the development of an accurate simulation. The simulation incorporated both the High Lift and Low Lift pumping stations. Verification of the simulation model indicated a deviation of less than 10% com- pared to the results of the actual procedure. Discussion of tests and simulation results The Peak Clip test was completed first in order to determine how the system reacts to a control change at the High-lift pump station. The most im- portant outcome of this test was to determine the necessary control inputs that would result in con- sistently high Balancing dam levels. From Figure 4 it is clear that it was possible to keep the Balancing Dam’s levels constant at around 2,5 m for the du- ration of the test. This was achieved by reducing the speed of the Low-lift pump station VSDs for the duration of the test to match the Low- and High-lift pump station outflows.

week into peak, standard and off-peak hours. This can then be utilised to generate electricity cost sav- ings for the client by reconfiguring the control phi- losophy of Pump Station A. The reconfigured con- trol will focus on pumping more during the standard and off-peak hours than the peak hours. It was pro- posed to conduct tests to better understand the system and develop an optimised control strategy before a prolonged intervention is implemented. The first test conducted was a peak clip test to determine the interaction between the sections of Pump Station A. In addition, this test also indicated the maximum load reduction that can be achieved during the Eskom evening peak hours as the speed of the pumps were reduced to their operational min- imum. For the peak clip test, the VSD speed of the High Lift pumping station was set on the minimum setting five minutes before the start of the Eskom evening peak period. The VSD speed of the low lift was set such that constant balancing dam levels could be achieved. For the duration of the evening peak period the set points were not altered. Five minutes after the evening peak period ended, the set points were reset to their pre-test values. The VSD speed values, station flow, power consump- tion and balancing dam values was recorded every fifteen minutes for the duration of the test. The next test is named the RPMs vs Power vs Station flow (RPS) Test. This test was conducted to determine the system characteristics of the High Lift pumping station. It was decided to only focus on the High Lift pumping station, as the Low Lift pumping station will mirror its control to keep the Balancing dam levels constant. The main outcome of the test is to determine the influence of VSD speed and the number of running pump sets on the total station outflow and power. This test was conducted by first increasing the VSD speed set points with increments of 50 RPMs from mini- mum to maximum on one pump set. When the maximumVSD speed was reached, the procedure was repeated with two pump sets and then three sets. Prior to increasing the VSD speed, all neces- sary data points were recorded. An overall scheme and site investigation led to the identification of possible constraints and lim- itations, such as lost communications and break- downs. This investigation developed the best pro- cedures to conduct the necessary tests, which ensured optimum data acquisition and test results. The results from the tests and investigation allowed

1800 1600 1400 1200 1000

Level [m] 3,0 2,5 2,0 1,5 1,0 0,5 0,0

800 600 400 200 0

RPM

19:15

18:15

17:30

17:45

19:30

19:45

20:15

18:30

18:45

19:00

18:00

20:30

20:00

Time

BD-Levels HL Speed LL Speed

Figure 4: Balancing Dam levels and VSD speeds for PC Test.

During the PC test there was a deficit in water transference compared to the normal pumping schedule. It is important to quantify this loss so that the necessary comeback load needed for a successful load-shift can be determined. Table 1 shows the average flow before and during the test. These two values are used to calculate the accumulative transfer deficit during the test.

Table I: Station flow during PC test.

HL-PS Flow [m 3 /s]

Average beforeTest

5,13

Average during test

3,36

Transfer deficit (m 3 )

12 762,30

It is clear from Table 1 that a significant volume of water was not transferred due to the test and this will accumulate from day to day if it is repeated.

18 Electricity + Control

JANUARY 2018