The Sagra drinking water treatment plant (DWTP)

The Sagra East DWTP treats the water that comes from the Almoguera dam to the El Algodor grouping of municipalities at the diversion point of the Noblejas DWTP which crosses the Jarama river in the municipal area of Aranjuez (Madrid).

Silt density index (SDI)

The SDI, or silt density index, expresses the concentration of colloids in the water, allowing us to categorise the quantity of soiling elements entering RO systems. There is no direct correlation between SDI and turbidity; however, in practical terms, it has been observed that feed waters with a turbidity of less than 1 nephelometric turbidity unit (NTU) or an SDI less than 5, have a low potential for RO membrane fouling. This is an empirical test in which we must keep the pressure and temperature constant over the 15 minutes that it usually lasts. The metering device includes a system whereby water is passed under pressure through a 0.45-micron Millipore membrane (or similar) while the time taken for the passage of a certain amount of water is recorded.

When we talk about water treatments, regardless of whether the origin of the water is marine, continental, or treated waste, we cannot ignore the importance of the potential problems caused by the presence of colloidal particles. These are largely responsible for the colour, turbidity, taste, and odour of water. Moreover, their small size (between 10−3 and 1 μm) means that they cannot be naturally filtered out. The origin of colloids is remarkably diverse:

  • Microorganisms (bacteria, viruses, and algae).
  • Minerals (colloidal clays, silts, metal salts, and silica, among other compounds).
  • Organic (such as surfactants, colourants, and humic or fulvic acids).

Without proper pre-treatment systems, these small, non-dissolved particles present in feed water will find their way to reverse osmosis (RO) membranes, settling on their surface and thus causing a fouling problem.

As a consequence of this fouling (as well as other factors such as age and temperature, etc.), membranes lose their most important characteristics over their normal operating period, especially in terms of the flow rate and the selectivity to salts. Therefore, depending on the membrane operating conditions, there may be a decrease in the flow rate of up to 10–12% per year. Thus, the passage of salts through the membrane increases by around 10–15% per year. This causes an increase in the feed pressure, a reduction in flow, and a worsening of the quality of the permeate (higher conductivity). Depending on the type of fouling, recovering a contaminated membrane through cleaning is often complex and so it is always better to stop the passage of the causative elements with adequate physical and chemical treatment before feeding the water into the trains.

Colloidal particles have a negative electrical charge in natural waters which results in a repulsion phenomenon that prevents them from joining to form larger particles. Therefore, we must neutralise these negative charges by adding positive charges that destabilise them so that the forces of attraction exceed those of repulsion. This leads to agglomeration of the particles—a phenomenon known as coagulation.

The chemical reaction of coagulation occurs in three stages:

  • Neutralisation of the negative charges of the colloidal particles by adding a coagulant (chemical reagent) with positive ions.
  • Coagulant reaction and the formation of positively charged colloidal hydrated oxide flocs which attract negatively charged colloidal impurities.
  • Adsorption (linked bonding or surface bonding) of impurities by the flocs.

When the reagent is solubilised in water it releases positive ions with a sufficient charge density to attract the colloidal particles and neutralise their charge.

The masses formed by the aggregation of destabilised colloidal particles (flocs) have a gelatinous texture and have dimensions in the order of tens to hundreds of microns. These flocs have an appreciable sedimentation rate and can therefore be separated from the water by decantation.

The type and amount of coagulant, pH, stirring speed of the mixture, coagulation period, and water temperature are just a few of the many factors that influence the coagulation process.

Experience with CHEMIFLOC® PA 47 at the Sagra drinking water treatment plant

Antecedentes

In 2012, we began to work with the Sagra drinking water treatment plant (DWTP) to search for the most appropriate treatment(s) to reduce high silt density index (SDI) values in water containing high quantities of sulphides.

Sulphide ions are reduced forms of sulphur that can appear under anaerobic conditions as the result of the action of sulfate-reducing bacteria. These bacteria (which are usually anaerobic) use sulfate as an energy source, reducing it to sulphur in the process. In the presence of sulphur, iron precipitates and forms Iron(II) sulphide (FeS). This is why ferric chloride is commonly used as a coagulant in these plants.

The Sagra facility had a pre-zonation process in place followed by a coagulation chamber in which ferric chloride (FeCl3) was dosed, a flocculation chamber, a lamellar decanter, sand filters, and finally a RO plant.

Although FeCl3 has the advantage of forming a large floc that decants immediately, it is also associated with a series of disadvantages such contributing colour to the water, the need for pH adjustment, and most importantly, its efficacy requires very high doses which translates into high plant costs.

Since the start of the Sagra plant operation in 2011, and especially during the summer period, the SDI values exceeded 3.2 with peaks of up to 4.

At Chemipol we are committed to looking for the best alternative(s) for proper plant functioning and consequently, we also aim to optimise the RO membranes, always within the current regulations cited in Order SAS/1915/2009 from 8 July on substances for the treatment of water intended for the production of water for human consumption published by the Spanish Ministry of Health and Social Policy.

Laboratory jar tests

Based on the information obtained while we studied the situation at the Sagra DWTP before starting the project and considering also that polyacrylamides could not be used in this installation, we identified and preselected a group of coagulants which would be used in the next phase of testing in the laboratory.

Chemical coagulation

According to Nelson L. Nemerow, chemical coagulation can be defined as a “unit process to cause the coalescence or aggregation of suspended non-sedimentable material and colloid particles in wastewater” (Nemerow, 1977). This process reduces the forces that repel colloids, forming heavier particles that precipitate through a process known as sedimentation.

According to Francis Edeline, coagulation is the “chemical process by which a coagulant (regardless of whether it is a chemical or natural substance) is added to water with the aim of destroying the stability of colloids and promoting their aggregation.”

Technical specifications

CHEMIFLOC® PA 47 is an organic coagulant based on polyDADMAC (polydiallyldimethylammonium chloride). It is not a dangerous product and is also free of Fe (III) and Al (III) metal ions. It is a liquid with a colourless to yellow appearance.

In the laboratory we designed a series of comparative jar tests to simulate the clarification process that occurs in the plant, always keeping in mind that the results do not depend only on the properties of the water to be treated, but also on the construction characteristics of each plant (such as, for example, the shape and dimensions of the decanters or the method used to add coagulants). The jar tests consisted of adding increasing doses of coagulant (measured in parts per million, or ppm) to a series of glass jars containing equal aliquots of the DWTP water. After the appropriate stirring period we then observed the characteristics of the coagulate that had formed, including its physical and chemical properties, floc sizes, and sedimentation time. We also noted the turbidity in each sample after the addition of the coagulant.

Our results indicated that from among all the coagulants we studied, the addition of our CHEMIFLOC® PA 47 in very small quantities improved the efficiency of the FeCl3, allowing a reduction in the 40-ppm dose that had been used by the plant up until then.

With this preliminary conclusion, we proceeded to the testing phase in the plant.

Tests at the drinking water treatment plant

Addition point

We decided to add the CHEMIFLOC® PA 47 directly into the flocculation chamber because the required dose was exceedingly small. This allowed us to avoid making a proposal that would entail large modifications to the plant installations. Only the automatic polyelectrolyte preparer—which comprised three tanks, the first two with stirrers—was modified by adding a pump connected to the signal of the solid polyelectrolyte dosing hopper. A timer was integrated into this connection, which made it easier to dose the coagulant in order to achieve a preparation between 0.5 and 1.0%.

Plant test results

Which changes were observed at the Sagra DWTP once our proposal was implemented?

Firstly, by adding only 0.8 ppm of CHEMIFLOC® PA 47 to the process in the flocculation chamber, the consumption of FeCl3 was reduced from 40 ppm to 25 ppm. Figure 1 shows the evolution of the dosage of both products.

Given that contact of ferric chloride with water releases heat and is highly corrosive, the corrosion of the facilities are reduced in a system that uses less FeCl3 and therefore, this change favours good maintenance of the plant and its facilities and deposits. However, it is difficult to quantify the effect of reducing the corrosion at the Sagra DWTP associated with the presence of lower amounts of FeCl3.

Nonetheless, we can clearly measure and calculate the reduction in sludge treatment costs associated with the reduction in FeCl3 doses. Figure 2 shows the tons of dehydrated sludge produced in 2012 (when the project started). In 2013 and 2014 25% less dehydrated sludge was produced compared to 2012, and in 2015 20% less was generated relative to the previous year. Thus, the accumulated difference in the production of dehydrated sludge in the fourth year was 40% lower.

The CHEMIFLOC® PA 47 dose also led to measurable improvements in the function of the RO and installed membranes. The SDI was considerably reduced, a sign of that fewer soiling elements were reaching the membranes. Indeed, the number of annual rack cleanings went from 8 in 2012 to 5 in 2015. This resulted in savings in terms of cleaning products, energy, and work time, and above all, increased the useful lifespan of the membranes.

Although the SDI is an indicative measure rather than a determining parameter, it is especially useful for comparing results at the end of each stage of a pre-treatment, thus allowing us to assess the effectiveness of any changes made to the system. Thus, as shown in Figure 4, it was evident that inclusion of CHEMIFLOC® PA 47 in the Sagra DWTP processes produced a substantial improvement, resulting in an SDI of less than 2. Given that the membrane manufacturers recommend keeping this value below 3, the Sagra DWTP had therefore obtained a considerable margin for error that even allowed them to reduce the cost of ozonation.

Conclusions

Almost 10 years after this project began, the process remains optimised with the use of CHEMIFLOC® PA 47. Here we summarise this optimisation in the following figures:

  • The addition of only 0.8 ppm of CHEMIFLOC® PA 47 reduced the FeCl3 consumption by 15 ppm.
  • The amount of dehydrated sludge produced fell from about 165 tonnes/year to less than 100 tonnes/year.
  • The reduction in soiling elements reaching the membranes meant that the plant went from 8 cleanings per year to only 5.
  • The useful lifespan of the RO membranes was extended to 12 years, and as of today, they still continue to efficiently filter the water feed at the plant.
  • Problems resulting from corrosion at the facilities were also reduced.

These results showed significant and measurable improvements in the plant’s functioning and unequivocally indicate that it is worth investing in the search for alternatives and improvements in RO pre-treatment(s). Conventional solutions can also be revised to help find means to optimally conserve the plant’s equipment and membranes.

From all of us at Chemipol, we encourage you to count on us to help you with this process.