Title: Mineral scale monitoring for reverse osmosis desalination via real-time membrane surface image analysis
Journal: Desalination
Authors: Alex R. Bartmana, Eric Lystera, Robert Rallob, Panagiotis D. Christofidesa and Yoram Cohena
Institute:
a Department of Chemical and Biomolecular Engineering, University of California, Los Angeles, 420 Westwood Plaza, Chemical Engineering Offices, Boelter Hall 5531, Los Angeles, CA 90095, USA
b Department d'Enginyeria Informatica i Matematiques, Universitat Rovira i Virgili, Tarragona, Catalunya, Spain
The original and creativity of paper: This paper developed a real-time approach as well as online membrane surface image analysis software to detect mineral scaling on RO membrane. Moreover, this approach can monitor crystal size and number density.
Summary
In this study, an ex-situ direct observation membrane monitor (MeMo) was developed for detecting of mineral scale formation as real-time system. The purpose of such monitoring is to provide quantitative information in order to appropriately initiate system cleaning/scale dissolution. Mineral scale in the MeMo system is monitored by comparison of consecutive images of the membrane surface for the purpose of determining the evolution of the fractional coverage by mineral salt crystals and the corresponding crystal count in the monitored region. Figure 1 shows the setting of membrane monitor MeMo whereas Figure 2 shows schematic of the membrane monitor (MeMo) testing arrangement.
Fig. 1. Membrane monitor (MeMo) cell: 1) Incident light for crystal detection is parallel to the membrane surface, 2) feed water stream entry, 3) membrane coupon underneath transparent top, and 4) example of imaged portion of the membrane surface (size and position can be adjusted).
Fig. 2. Schematic of the membrane monitor (MeMo) testing arrangement.
During mineral scaling experiments, membrane surface images were automatically captured (at a time interval of 15 min) and stored on the MeMo data acquisition computer at a prescribed interval. After that, captured images were analyzed using imaging processing software developed specifically for the MeMo system. The membrane surface image analysis (MSIA) software consists of the following components; image pre-processing/algorithm initialization, image subtraction and smoothing, edge detection and hysteresis thresholding, crystal confirmation and crystal count/area calculation, and data output. A description of the image analysis approach is provided in Fig 3.
Fig. 3. Flowchart of image analysis algorithm with representative example image outputs for selected algorithm steps: a) original camera image saved to MeMo computer disk, b) image resulting from subtraction of two most recently captured images (black pixels represent little to no change in pixel value and white denotes large changes when compared to the previous image), c) subtracted image after image filtering and edge detection, and d) final cumulative scaling image after morphological transforms.
Real-time images of the membrane surface in the MeMo membrane channel were successful analyzed online to detect the onset of mineral crystals as well as to monitor the evolution of the fractional coverage by mineral salt crystals and crystal count. With the present approach, the results are very meaningful for cleaning protocol setup. According to the approach, when mineral scale coverage reaches a prescribed threshold, a control signal can be sent to the plant control system in order to initiate any number of cleaning protocols.
Application & further study: Discovery technique is very meaningful for fouling mechanism investigation.
By Monruedee Moonkhum
Email: moon@gist.ac.kr
