Developing a deterministic model for predicting cleaning frequency due to inorganic scaling on reverse osmosis membranes

Project ID: 792
Principal Investigator: Frank Leitz
Research Topic: Desalination and Water Treatment
Priority Area Assignments: 2013 (Advanced Water Treatment), 2014 (Advanced Water Treatment)
Funded Fiscal Years: 2013 and 2014
Keywords: water treatment, reverse osmosis, membrane fouling, modeling, membrane cleaning, economic analysis

Research Question

Reverse osmosis is the most widely used brackish groundwater desalination technology in the United States. One of the major limitations of reverse osmosis membranes is inorganic scaling, which causes deterioration of membrane performance. During operation, reverse osmosis membranes are chemically cleaned on a regular interval to restore membrane performance. Cleaning frequency varies depending on feed water quality, and operating parameters, including feed water recovery, operating pressure, and temperature.

Designers, engineers, and membrane operators frequently use commercially available membrane models to predict the performance of reverse osmosis membranes (e.g., rejection of ions and permeate flux rate). However, these models cannot predict cleaning frequency for reverse osmosis systems. The lack of commercial information available for assessing fouling propensity and cleaning frequency limits the ability to predict plant operational costs.

Although not readily available in commercial models, research on fouling mechanisms and predictive equations is available. Notably, the silt density index (SDI) and models related to concentration polarization and saturation indices have been used to assess membrane fouling. Water treatment plants often have limited information on the required inputs for these models. The goal of this research is to gather water chemistry and membrane operation data from membrane plants to determine cleaning requirements based on projected fouling curves to answer the research question:

Can a mathematical model be developed that will predict cleaning frequency for reverse osmosis systems in brackish groundwater desalination plants in order to more fully predict the operational costs associated with advanced water treatment plants providing water for municipal use?

Need and Benefit

There is a need to further understand and characterize reverse osmosis desalination as it is the most widely used brackish groundwater desalination technology in the United States. Brackish groundwater has natural challenges related to high mineral content with the tendency to form sparingly soluble salts. Formation of these salts on the membrane surface is referred to as inorganic scaling. Scaling causes decreased flux, increased operating pressure, and increased plant downtime due to cleaning requirements. Chemical cleaning is performed to restore membrane performance with cleaning frequency depending on feed water quality and plant operating parameters.

Commercial membrane models currently do not predict cleaning frequency for reverse osmosis systems. Commercial information available for assessing fouling propensity and cleaning frequency relies on an as needed basis or rule of thumb frequency (annual, biannual, quarterly, etc.). Evaluation of frequency on an as needed basis generally consists of operating indicators such as flux decline or increasing trans-membrane pressure. This approach is generally characterized as a reactionary response, but there is value in determining a predictive model to accomplish stable operation. Stable operation can be characterized as a very low decrease of the mass transfer coefficient and a low increase in differential pressure, resulting in a low cleaning frequency of the membrane units.

There are a number of benefits related to predicting cleaning frequency. A few of these benefits are summarized below:

-Projections allow for optimization of operational parameters to reduce the overall frequency of cleaning and potential for irreversible fouling,
-A model that takes into account feed water chemistry as an indicator to fouling is applicable for facilities across the western US, and
-By assessing the fundamental mechanisms for fouling the model can reduced into equations that can be included in a cost model.
-Development of a more robust cost model that includes cleaning frequency will assist water utilities during the planning stage to better understand the potential costs of brackish groundwater desalination.

The state of Texas and Texas Water Development Board addressed the need for this project during the Reclamation Research Jam. With 32 brackish groundwater desalination plants statewide, the needs of Texas facilities are similar to water utility needs nationwide. Although the current total capacity of Texas brackish groundwater desalination plants 70 mgd, according to the 2011 Texas State Water Plan, an additional 160 mgd (180,000 acre-feet per year) of brackish water is needed by 2060 to meet future water supply needs. An explicit description of foundation for this research is provided in the Texas Water Development Report entitled "Desalination Database Updates for Texas". In the document facilities were solicited for information regarding a number of plant factors including reverse osmosis membrane cleaning. Desalination facilities in Texas reported inorganic scaling as one of the most predominant operational problems experienced at the plant.

Contributing Partners


Research Products

Bureau of Reclamation Review

The following documents were reviewed by experts in fields relating to this project's study and findings. The results were determined to be achieved using valid means.

Survey Form for Collecting Reverse Osmosis Plant Operation Data (interim, PDF, 72KB)
By Dr. Katharine Dahm, Dr. Saqib Shirazi, Anna Hoag, Andrew Tiffenbach and Dr. Katie Guerra
Report completed on May 07, 2013

Survey Form for Collecting Reverse Osmosis Plant Operation Data

1. General Information
2. Plant Information
3. Source Water Information
4. Plant Operation Information
Keywords: survey, water quality, membrane data, cleaning procedures

Last Updated: June 29, 2015