Water monitoring
TABLE 3. ACCURACY TEST RESULTS, PPB. ONLY THREE PAIRS WERE NOT WITHIN EXPECTED TOLERANCE. ULR readinga
GS#1 60 60 50 GS#2 40 80 50 GS#3 70 40 AVG 57 60 50 STD 12 16 0 23 16 20 Σ LODb 28 28 28 vs. ULR 33.3 44.1 30.4
Grab Sample takenc
12/28/2020 14:00
12/31/2020 12:45
1/8/2021 16:50
a: Readings correspond to the grab sample time; b: Refer to Table 2 for match criteria; c: The grab sample was taken at the recorded time and two or three analyses were conducted consecutively, using the same sample.
RO system influent, after GAC beds and Sodium Metabisulfite (MBS) injection, with source water (city tap water) containing 3 to 4ppm chlorine before GAC (Figure 3).
After conducting the MBS response test (Figure
4), the plant personnel made the first observations, calculations, and preliminary conclusions leading to the extension of the test to learn more about the analyser and its capabilities. The main results of the first three weeks of testing showed the analyser demonstrating stable and accurate readings and fast reaction to changes in the MBS feed (Figure 4). This facility normally calculates membrane
life span based on the manufacturer’s recommendations to maintain chlorine level < 100ppb and tries to keep it below 80ppb with the target set at 30ppb. The existing grab sample analysis method detects and measures chlorine above 20 ppb, and was used to verify the performance of the ULR analyser in a comparative test conducted in the extended trial (Figure 5). Insufficient sample flow can affect the
performance of any process analyser, and therefore, the intermittent operation of RO skids, being a normal case, can present a big challenge. The internal flowmeter of the new ULR analyser helped to overcome this challenge and maintained the instrument’s operation by placing the analyser on standby when the sample flow was insufficient, and automatically restarting its operation when the flow was restored. This ensured the accuracy
of the analyser readings recorded in the internal logs, which were thoroughly analysed to arrive at the right conclusions. From the analysis of chlorine and flow data,
graphically represented in Figure 5, it was clear that once the MBS feed was adjusted to lower rates based on the grab sample results, the discrepancy between grab sample and online analyser readings fell out of expected tolerance (Table 2). This can be explained by comparing grab sample analysis details and specifications for both methods (Table 3). Table 3 shows there were several grab samples
taken for each comparison falling out of expected tolerance and the spread between the results for the same sampling was quite significant, up to 40ppb. This indicates either fluctuations in the sample, or accuracy of the lab analysis, or both. Therefore, the comparison between ULR chlorine readings (LOD = 8ppb) and lab results (LOD = 20ppb) should be considered marginally matching. Mainly, such discrepancies can be attributed to a higher probability of deviations in conducting grab sample analysis, because any test involving human interaction increases chance of a random error. Based on this logic, statistics, and specifications, the ULR process analyser was found to be producing accurate results, comparable to the reference grab sample analysis. Simple data evaluation showed that, based on
the analyser readings, dosage of dechlorinating agent (e.g., MBS in this case) could be safely
reduced and later eliminated without compromising the quality of the operations and risk to increase biofouling of the membranes. Solely chemical cost savings can potentially return all investments in the analyser at this facility in three to five years. However, once other direct and indirect savings (e.g., reduction in CIP frequency, associated labour and chemicals, extended membrane life, reduction in production losses, etc.) are factored in, the ROI period becomes shorter and more appealing. The instrument was left running at this facility
for a long-term evaluation and after over a year- long test, more observations were collected. For example, the analyser responded to a recent event related to a GAC tank failure (Figure 6). The first pass RO feed comprised the combined effluent from all carbon beds (GAC tanks). Two out of four carbon beds each account for ~20 per cent of the total flow and the other two for ~30 per cent each. Sodium metabisulfite (if online) is injected downstream of the carbon beds and upstream of the RO membranes. The event presented in the graph (Figure 6) happened after the MBS feed stopped on 6 June 2021. It was discovered that one GAC tank’s effluent was bringing 150ppb of chlorine to the combined sample and another – 80ppb at ~50 per cent of total flow. This contribution was immediately detected and recorded by the analyser and once the media in the bad GAC tank was replaced (9 July 2021), the chlorine concentration came down to the desired level of < 30ppb as the grab sample analysis confirmed at 14:58 on 9 July2021 (Figure 6). Thus, the new analyser helped to point in the
right direction for troubleshooting GAC media that can be exhausted, or the tanks can develop channels inside carbon granules where chlorine may pass through. This is another potential benefit of the new instrument, especially when its outputs are connected to the facility’s SCADA system, or DCS, and the readings are used for decision support, if not for dechlorination control.
CONCLUSION This case study demonstrates the value of highly accurate direct chlorine measurements at minimal maintenance efforts and supports all chemical and labour cost savings elucidated by the instrument projecting ROI in approximately two years.
Figure 6 - GAC tank event. The flow through the analyser was also indicative of the increasingly intermittent operation, which did not affect the instrument’s performance.
Instrumentation Monthly April 2022 Analog Devices
www.analog.com 41
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78
