Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Prepared for:
Prepared by
Prepared for:
The City of Ferndale Public Works
Ferndale, Washington
Prepared By:
Wilson Engineering, LLC
TABLE OF CONTENTS
TABLES
Table 1. Hardness and softening related parameters. .............................................................................4
Table 2. Major TDS components of groundwater at different points in time. ........................................6
Table 3. Comparison of hardness and corrosion potential characteristics ..............................................8
Table 4. Water chemistry testing results for December 21, 2012 and April 10, 2013. ............................9
Table 5. Disinfectants and DBPs in finished water. ................................................................................10
Table 6. Water usage existing and projected, assuming 2.78 percent annual growth .......................14
Table 7. Selectivity coefficients for a typical cation/anion exchange resins. .........................................16
Table 8. Benefits and disadvantages of Ion Exchange with sodium as regenerant. ..............................16
Table 9. Benefits and disadvantages of Ion Exchange with acid as regenerant. ....................................20
Table 10. Benefits and disadvantages of full deionization (cation and anion exchange). .....................20
Table 11. Potential Rejection .................................................................................................................22
Table 12. Benefits and disadvantages of nanofiltration.........................................................................24
Table 13. NF Unit Configuration and Characteristics .............................................................................28
Table 14. Capacity of Water Softening System for Maximum Day Demand ..........................................29
Table 15. Estimated Cost of Capital Improvements ...............................................................................31
Table 16. Estimated Cost of Operation and Maintenance .....................................................................33
FIGURES
Figure 1. Illustration of major TDS components of the Douglas Well at different points in time ...........6
Figure 2. Illustration of major TDS components of the Shop Well at different points in time ................7
Figure 3. Graphic of softening ion exchange resin and photos of dry and wet resin beads. .................15
Figure 4. Illustrations of Ion Exchange using acid as a regenerant ........................................................18
Figure 5. Schematic of Reverse Osmosis Membrane .............................................................................22
APPENDIX A Drawings
APPENDIX B Pilot Study Report and Data
APPENDIX C Water Quality Data
DRAFT i
ACRONYMS AND ABBREVIATIONS
DRAFT ii
WATER SOFTENING PROJECT REPORT CITY OF FERNDALE
1. Project Description
1.1 Problem Description
The City of Ferndale, in December 2011, converted from using the Nooksack River to groundwater
for its public water supply. The water supplied to the citizens of Ferndale from the groundwater
source is in compliance with Maximum Contaminant Levels for all primary contaminants. However,
since being put into service the well water has changed from being moderately hard to hard,
which is a nuisance and an aesthetic issue. Hardness was previously only about 70 mg/L (Nooksack
River as the source). In addition, the total dissolved solids (TDS) is near or over the 500 mg/L SMCL.
The existing treatment system consist only of manganese removal using greensand and disinfection
with hypochlorite. The existing treatment system cannot reduce hardness or TDS.
Identify and evaluate feasible options for softening the water (i.e., reducing hardness) and
implement the selected alternative. This evaluation examines fully the characteristics of the
groundwater and how those characteristics interact with existing and potential water treatment
methods. The goal is to identify and implement a water softening alternative that will result in a
water supply that has an acceptable level of hardness and meets all water quality standards. The
selected alternative should improve the overall quality of water supply; and it should not cause
any unwanted side effects.
Construction is expected to begin in early 2014 and be completed by August or September 2014.
The capital cost is estimated to be $1.8 to $2.1 million (including engineering). The Operations and
Maintenance cost is expected to be $65,000 to $85,000 per year (plus 500 hours of additional
operator time).
Financing will be existing City funds dedicated to the drinking water utility.
1.7 Treatment
Existing Treatment Capacity. The current WDOH approved capacity of the manganese treatment
system (two greensand filters in parallel configuration) is 2,170 GPM (3.12 MGD). This treatment
system has been effective at removing manganese, which would otherwise be well above the
Secondary Maximum Contaminant Level. It also removes the trace amounts of iron present, but
little else. The manganese treatment system has enough flow capacity to meet water demands
until 2028.
2. Planning
The City has an approved Water System Plan.
3. Water Quality
Review of water quality as it relates to the purpose of the proposed project, including results of raw-
and finished-water quality analyses. Include water quality information in the groundwater use
feasibility report and subsequent water sample results.
Hardness. The water chemistry parameters directly related to hardness (and to softening) are
shown in Table 1. Manganese is included here even though it is a relatively small contributor to
hardness, because it can have a very negative affect (fouling) on treatment systems. Of greatest
concern is the increase in hardness subsequent to the wells being placed into production. The
Douglas Well hardness has increased from 116 mg/L to 206 mg/L since the well was put online.
The Shop Well hardness has also increased, though to a lesser extent, from 92 mg/L to 125 mg/L.
The proportion of water withdrawn from the Shop Well was increased during the 2012 dry season,
from about 35 percent to about 65 percent of the total withdrawn. The resulting blended water
hardness was approximately 160 mg/L.
It appears that the hardness levels for both of these wells leveled off at the end of summer 2012
when the pumping rate decreased. The causes for the hardness increases and subsequent leveling
off of hardness are open to speculation. However, the increases are very likely due to entrainment
of brackish groundwater (relic seawater). Brackish groundwater has higher levels (relatively) of
chloride, sodium, bromide, calcium and magnesium.
Alkalinity. Alkalinity is an important water quality characteristic. Alkalinity is the capacity of water
for neutralizing an acid solution. For both wells, the alkalinity is essentially all in the form of
bicarbonate (HCO3-), excepting about 1 percent as carbonate (CO32-). It is the carbonate that forms
calcite scaling by combining with calcium. The alkalinity of both wells is substantially higher than
the hardness, which indicates that a portion of the alkalinity originates from sodium carbonate.
High alkalinity contributes to greater calcite scale deposition. High alkalinity is good for buffering
against pH changes. However, this buffering capacity is a disadvantage if the selected treatment
requires first lowering the pH.
Total Dissolved Solids (TDS). Table 2 shows TDS and conductivity and the major components of
TDS (other than hardness) in the well water at a few different points in time. The data prior to
2012 was obtained only after pumping the wells for a relatively short period of time (i.e., time
enough to flush adequately the wells for water quality testing and to calculate drawdown and
ultimate pumping capacity). Such tests will not always identify long-term trends from continuous
drawdown. Long-term trends can only be feasibly determined by putting a well into use.
Total Dissolved Solids (TDS) increased in both wells due to drawdown in 2012. The Douglas Well
(650 mg/L) is now above the Secondary MCL of 500 mg/L. The Shop Well (460 mg/L) is near the
SMCL. Because the Douglas Well groundwater has TDS exceeding the SMCL, it needs to be
blended with the Shop Well in order to supply water that is below the SMCL (assuming no
treatment to reduce TDS). The Shop Well to Douglas Well blend would need to be 4:1 at a
minimum, which is not feasible. The TDS increase during 2012 was due to increases in chloride
and sodium mostly, but also to increases in calcium, magnesium, sulfate, and bicarbonate.
Scattered brackish groundwater deposits are known to occur in the western Whatcom County and
especially near the mouth of the Nooksack River. A phenomenon called connate ascension occurs
when a freshwater aquifer overlies brackish connate water. Pumping may cause the connate
water to migrate upward and into the freshwater aquifer. This results in a brackish, briny
contamination similar to the effects of saltwater intrusion. Alternatively, pumping may draw in
higher salinity water laterally from adjacent brackish groundwater. Future changes in well water
salinity cannot be easily predicted. Therefore, to ensure continued use of the groundwater into
the future, any proposed treatment should be considered carefully to allow for adaptation to
changes in groundwater characteristics over time (i.e., if chloride and sodium and thus TDS
increase to unacceptable levels, will the treatment system mitigate for this or can additional
components be easily added to mitigate the TDS increase).
Chloride. After one year of production pumping, the chloride concentration has increased by
about 150 mg/L in the Douglas Well (from 56 mg/L to 218 mg/L) and by about 50 mg/L in the Shop
Well (from 33 mg/L to 91 mg/L). Chloride is a good measure of the overall increase in salinity
(chloride flows freely with the groundwater). In this case, chloride appears to be a direct indicator
of the amount of saline water (seawater in origin) in the groundwater. Seawater contains
19,000 mg/L chloride, so 190 mg/L indicates a 1.0 percent seawater content. Therefore, the
Douglas Well is currently at about 1.0 percent seawater and the Shop Well is at about 0.5 percent
seawater. Seawater contains 6,400 mg/L of hardness. Thus, 1.0 percent of seawater could add 64
mg/L of hardness.
Sodium. Sodium is relatively high in both wells (Douglas Well 137-146 mg/L and Shop Well
106-120 mg/L). Sodium in the Shop Well makes up a relatively larger proportion of the TDS than in
the Douglas Well. This is because the Shop Well has more sodium bicarbonate alkalinity than the
Douglas Well.
800
Carbonate
DOUGLAS WELL
Magnesium
700
Calcium
Sulfate
Chloride
600
Sodium
Concentration (mg/L)
400
300
200
100
0
2/23/1994 3/22/2007 12/22/2009 12/21/2012
Figure 1. Illustration of major TDS components of the Douglas Well at different points in time
(calcium, magnesium, and bicarbonate data not available for 1994 or 2007).
600
SHOP WELL
500
Carbonate
Magnesium
Concentration (mg/L)
400
Calcium
Sulfate
Chloride
300
Sodium
Total Diss. Solids
200
100
0
8/4/1999 3/22/2007 4/28/2010 12/21/2012
Figure 2. Illustration of major TDS components of the Shop Well at different points in time
(calcium, magnesium, and bicarbonate data not available for 1994 or 2007).
Hardness and Corrosion Relationship. The table below compares the hardness and corrosion
potential characteristics of the previous surface water supply versus the existing groundwater
supply. The surface water supply (from the Nooksack River) had very low alkalinity, moderate
hardness (about 60-70 mg/L), and neutral pH. The Langelier Index indicates that it was very
corrosive to pipes. Aside from pipe longevity issues, the health concerns of corrosion are
dissolution of lead and copper into the drinking water from lead solder and copper pipes,
respectively. However, Ferndales drinking water from the previous surface water supply was in
compliance with lead and copper limits. The current well water supply has very high alkalinity and
relatively high hardness and pH. The Langelier Index indicates that calcium carbonate scaling will
occur. The softening goal for the drinking water is a hardness of 70 mg/L. In addition, the
softening method used would ideally result in water that is neither corrosive nor scaling (Langelier
Index between -0.5 and 0.5).
Combined
Nooksack Combined Wells
River Wells Softened*
pH 7.2 8.2 7.9-8.2*
Full chemical testing Results. Table 4 gives the full chemical testing results (except for
disinfection) for samples collected on December 21, 2012 from the Douglas and Shop Wells and
from the blended well water (35 percent Douglas and 65 percent Shop) and the finished water
(treated and disinfected). Table 5 gives the results for disinfection chemicals and disinfection
byproducts (DBPs). These results are discussed in the following section.
Some of the chemical characteristics of the groundwater are directly relevant to softening, some
are indirectly related to softening, some cause problems for certain treatment methods, and
others are of concern for other reasons. The following four subsections discuss (1) primary
contaminants, (2) secondary contaminants, (3) disinfection and disinfection byproducts, and (4)
substances that foul treatment systems or interfere with treatment.
Table 4. Water chemistry testing results for December 21, 2012 and April 10, 2013.
12/21/2102 4/10/2013
Douglas Shop Blended Douglas Shop Method
Parameter Units Well Well Water Well Well MCL* Limit
Flow GPM 450 750 -- 280 580
pH S.U. 8.2 8.2 8.2 8.2 8.2
DO (field) mg/L 1.5 1.2 2.2 -- --
Chlorine (field) mg/L -- -- 1 -- --
Total Diss. Solids mg/L 652 454 556 650 464 500 20
Conductivity S/cm 1162 800 940 1199 814 10
Turbidity NTU 0.59 0.14 ND 0.33 0.69 0.1
Hardness mg CaCO3/L 206 125 152 195 119 3
Alkalinity mg CaCO3/L 228 244 237 230 257 1
Metals and other Cations
Aluminum mg/L ND ND ND ND ND 0.2 0.01
Barium mg/L 0.09 0.07 0.07 0.104 0.085 2.00 0.001
Boron mg/L 0.15 0.12 0.14 0.15 0.12 0.05
Calcium mg/L 43.6 25.9 31.8 41.4 24.9 0.5
Iron mg/L ND ND ND ND ND 0.3 0.05
Lead mg/L ND ND ND ND ND 0.015 0.0005
Magnesium mg/L 23.5 14.6 17.6 22.3 13.7 0.5
Manganese (Diss.) mg/L 0.222 0.102 ND 0.222 0.093 0.05 0.001
Manganese (Total) mg/L 0.226 0.106 0.011 -- -- 0.05 0.001
Potassium mg/L 6.2 6.1 5.5 5.5 5.4 0.5
Sodium mg/L 146 120 130 138 106 1
Strontium mg/L 0.29 0.18 0.22 0.266 0.163 0.05
Ammonia mg/L N 0.1 0.18 0 0.12 0.18 0.03
Anions
Arsenic mg/L 0.004 0.006 0.006 0.003 0.004 0.01 0.001
Bromide mg/L 0.88 0.35 0.36 0.88 0.36 0.05
Bicarbonate mg HCO3/L 276 295 287 278 311
Carbonate mg CO3/L 2 2 2 2 2 1.00
Bicarbonate mg CaCO3/L 226 242 235 228 255 1.00
Carbonate mg CaCO3/L 4 4 4 4 4 1.00
Chloride mg/L 218 91 149 221 88 250 1
Fluoride mg/L 0.36 0.25 0.28 0.28 0.19 4/2 0.1
Iodide mg/L ND ND ND -- -- 0.1
Phosphates mg/L P 0.06 0.17 0.13 0.07 0.17 0.01
Silica mg/L SiO2 22 20 20 16 19 0.05
Sulfate mg/L 26 40 35 24 32 250 0.2
* MCL = Maximum Contaminant Level
12/21/2012 1/15/2013
Finished Tap* Method
Parameter Units Water Water MCL Limit
Disinfectants**
Total Chlorine mg/L (as Cl2) ND 4 0.05
Free Chlorine Residual mg/L (as Cl2) ND 0.05
Total Bromine mg/L (as Cl2) 0.5 0.05
Free Bromine Residual mg/L (as Cl2) 0.5
Chloramines mg/L (as Cl2) 0.06 4
Chlorine+Bromine (field) mg/L (as Cl2) 1.0
DBPs
Bromodichloromethane g/L 2.7 3.6 0.5
Bromoform g/L 25.5 52 0.5
Chlorodibromomethane g/L 7.3 9.4 0.5
Chloroform g/L 1.1 1.3 0.5
Total Trihalomethanes g/L 36.6 66.2 80 0.5
Bromochloroacetic Acid g/L 1.1 1.2 0.3
Dibromoacetic Acid g/L 5.9 7.6 0.4
Dichloroacetic Acid g/L 1.0 ND 0.5
Monobromoacetic Acid g/L ND ND 0.5
Monochloroacetic Acid g/L ND ND 0.5
Trichloroacetic Acid g/L ND ND 0.5
Total HAA(5) g/L 6.9 7.6 60 0.5
Bromate mg/L ND 0.01 0.005
* The sample location is at the farthest point in the distribution system.
** The concentration of disinfectant was higher for the field measurement than for the lab test due
to decay of chlorine and/or bromine in transit to the lab.
ND Not detected.
Iron and manganese can foul membranes and cation exchange systems. Dissolved oxygen or
chlorine will convert Fe2+ (ferrous iron) into Fe3+ (ferric iron), which forms insoluble colloidal
hydroxide particles that may foul membranes or resins. The same process applies to
manganese, except that manganese is only slowly oxidized by dissolved oxygen.
Silica can foul membranes and ion exchange systems. Membranes will become fouled with
insoluble colloidal silica or silica gel when the concentrate stream silica content exceeds the
solubility of silica. The negatively charged form of silica will react with calcium, magnesium,
manganese, and especially with iron and aluminum to form insoluble silicates, which will
foul membranes.
Calcium carbonate or calcium sulfate precipitation can cause fouling of equipment.
Organic matter can cause growths on membranes in the absence of a disinfectant. Dissolved
organic matter will adhere to anion exchange resin and may be difficult to remove during
regeneration. Dissolved organic matter (TOC) is low in the groundwater (0.5-0.6 mg/L) but
still requires some additional maintenance for nanofiltration.
Iodide, nitrate, phosphates, silica, and sulfate can interfere with anion exchange treatment.
Only silica and sulfate are present at significant concentrations in the groundwater.
Iron and Manganese. The Douglas and Shop Wells both contain manganese and some iron, all in
the dissolved (unoxidized) state. Unoxidized iron and manganese do not cause significant
problems with either acid ion exchange or membrane treatment. It is important, then, not to
oxidize the raw well water (e.g., with chlorine or aeration) prior to ion exchange or membrane
treatment. That is, the well water should be pumped directly to the ion exchange or membrane
treatment.
Silica. The Douglas and Shop Wells both contain about 20 mg/L of reactive silica. 20 mg/L is well
below the solubility of silicic acid and therefore should not cause a problem for the cation
exchange process. This level of silica would be a concern for anion exchange, as it would compete
for exchange sites (i.e., increasing frequency of regeneration). A membrane treatment would
require anti-scalant measures be taken such as adding an anti-scalant chemical. The presence of
trivalent cations (Al and Fe) can greatly increase the scaling effect of silica.
Calcium. The Douglas and Shop Wells both contain calcium, which will cause calcium carbonate
scaling of membranes. A membrane treatment would require anti-scalant measures be taken such
as reducing pH and/or adding an anti-scalant chemical.
Water Rights. Water rights place limits on two different withdrawal rates. The City has the right to
withdraw a maximum 2,870 GPM (4.13 MGD) instantaneously and a maximum of 2,055 acre-feet
(1.872 MGD) per year. The Shop Well has the more limited water rights of the two source wells:
870 GPM instantaneous and 440 acre-feet (0.39 MGD) per year. The volume pumped from the
Shop Well exceeded its annual water rights limit during 2012 because operation of the Douglas
Well was kept to a minimum because it had harder water.
The Douglas Well showed a greater increase in hardness and TDS (mainly chloride and sodium in
addition to hardness) as a result of well drawdown than did the Shop Well. Based on the results of
the first year of operation, further increasing the pumping rate from the Douglas Well may further
increase the levels of hardness and TDS in the pumped water.
Water Demand. Table 6 shows the existing and projected water demand (assuming 2.78 percent
annual growth). Note that the volume pumped from the wells is 3-5 percent higher than the
demand due to water lost during the greensand filters backwashing process. The demands shown
in Table 6 also includes the distribution system leakage rate, which is relatively low at only 5
percent or less. The existing and future peak day demands are well below the instantaneous water
rights limit. The capacity of the greensand filters is more than adequate for the existing demand.
The limitation on annual withdrawal quantity would potentially be an issue circa 2032. The 2012
population and number of connections are 11,080 and 5,358, respectively. The demand from the
buildout population of 25,000 could exceed the annual water right limit. Therefore, treatment
processes that waste large amounts of water may not be preferable unless more water rights can
be obtained.
This is sufficient to meet Maximum Day Demand through approximately 2017 and Maximum
Month Demand through 2031.
The final design flow used for the selected alternative is:
This is sufficient to meet Maximum Day Demand through approximately 2019 and Maximum
Month Demand through 2034.
Table 6. Water usage existing and projected, assuming 2.78 percent annual growth
Well drawdown has resulted in increased salinity (and hardness) for the two Ferndale wells.
However, the effect on the Douglas Well has been greater as will be discussed in the next section.
It should be noted that the Douglas Well screened intake is 75 feet lower in elevation than the
Shop Well screened intake. Because salinity often increases with depth, the disparity between the
two wells is not atypical.
5. Analysis of Alternatives
A comparison of alternative solutions and the rationale for selecting a proposed alternative is
required for all types of projects (WAC 246-290-110(4)(c)). The comparison will include life-cycle
cost evaluations (initial costs plus on-going operations and maintenance costs).
This analysis will include nanofiltration, ion exchange, lime softening, and source blending.
This evaluation examines the benefits and disadvantages of several potential water softening
alternatives. The effectiveness of any softening system is directly affected by the water chemistry
and flow requirements of the water supply. The issues and water characteristics considered in this
evaluation of softening treatment alternatives include:
Figure 3. Graphic of softening ion exchange resin and photos of dry and wet resin beads.
Regeneration. Eventually, after continued use, a point is reached when very few sodium or
potassium ions remain on the resin, thus no more calcium or magnesium ions can be removed
from the incoming water. The resin at this point is said to be exhausted, and must be recharged or
regenerated. Regeneration is performed by flushing the media with a large quantity of salt
(sodium chloride) brine (which must be disposed of in the sewer system typically).
Table 7. Selectivity coefficients for a typical cation/anion exchange resins.
Selectivity Selectivity
Cation Coefficient Anion Coefficient
Li+ 0.76 HSiO3 <1
H+ 1.00 F 0.2
Na+ 1.56 H2PO4 0.5
NH4+ 2.01 HCO3 0.8
Mn2+ 2.07 OH 1.0
K+ 2.28 Cl 1.5
Mg2+ 2.59 Br 3.5
Ca2+ 4.06 NO3 5.1
Sr2+ 5.13 HSO4 9.4
Ba2+ 9.06 I 11
Potassium chloride can be used instead of sodium, if sodium is a concern; however, the cost of the
salt is nearly double. In addition, the salinity will still increase in drinking water and the amount of
chloride brine discharged to waste is still very high. The table above shows that potassium has a
selectivity coefficient that is not much less than magnesiums, thereby making potassium much
less efficient than sodium at removing magnesium. Note also in the above table that strontium
and barium compete strongly for negative sites and will reduce the resin softening efficiency.
Benefits Disadvantages
Flexible can easily increase number of Requires discharge of large quantities of salt
treatment units or can vary the volume brine to the sewer (adds about 300 mg/L)
treated to adapt to changing hardness chloride to wastewater)
concentrations
5.1.3 Implementation
The water would be split after being treated by the greensand filters. About half would go to
the ion exchange system and half would bypass it. The softened water, containing near zero
hardness, would be combined with the bypassed water stream to achieve an overall 50
percent reduction in hardness. The operator would be able to control the flow rates to each
of the parallel systems in order to achieve the desired final hardness level. Incoming
hardness would be monitored daily to adjust the proportions of the two flow streams.
Two cation exchange tanks would give 1600 GPM total flow with no backup unit. Three
cation exchange tanks would give 2500 GPM total flow with no backup unit or 1600 with
one backup unit. Each tank is 9 feet in diameter. The regeneration equipment would require
a space of about 20 feet by 12 feet. The booster pumps would require a space of about 12
feet by 10 feet. The ancillary equipment would not require much additional room.
These units would be installed downstream of the greensand units. This is necessary
because manganese and iron will foul the resin. Because of this arrangement, the feedwater
to the IEX will have to be dechlorinated and then repressurized with booster pumps.
The amount of water wasted to produce about 1600 GPM or about 2.3 MGD water
(assuming 3% waste from the greensand) would be about 50 GPM (about 4%).
The capital cost is about on par with the cost of ion exchange using acid as a regenerant. See
Cost Estimate section.
The O&M costs are higher than for ion exchange using acid as a regenerant and about the
same as for nanofiltration. See Cost Estimate section.
Benefits Disadvantages
Reduces TDS (by about 15%), which improves Does not remove anions (chloride, sulfate,
taste bromide)
Less chlorine use (about 30% less), which Does not remove DBP precursors
reduces DBPs by 10-20%
Can remove sodium (however, this would The elevated well water sodium concentration
increase cost and complexity of the system) increases frequency of regeneration somewhat
Table 10. Benefits and disadvantages of full deionization (cation and anion exchange).
Removes anions (chloride, sulfate, bromide) Increased capital and O&M cost
Degasifier may be reduced in size or eliminated Silica fouling may reduce life of resin
Removes less alkalinity than softening only. Requires handling/disposal of strong acid base
Two cation exchange tanks would give 1600 GPM total flow with no backup unit. Three
cation exchange tanks would give 2500 GPM total flow with no backup unit or 1600 with
one backup unit. Each tank is 9 feet in diameter and the degasifer would require about the
same amount of room. The acid tank and regeneration equipment would require about
20 feet by 12 feet. The ancillary equipment would not require much additional room.
These units would be installed upstream of the greensand units and prior to chlorination.
Because of this arrangement, the feedwater to the IEX will not have to be dechlorinated or
repressurized with booster pumps.
The amount of water wasted to produce about 1600 GPM or about 2.3 MGD (assuming 3%
waste from the greensand) would be about 50 GPM (about 3-4%).
DBPs would likely be reduced by 10-20%. Total chlorine usage would be reduced by about
10-20%.
The capital cost is about on par with the cost of ion exchange using sodium as a regenerant.
However, the operating cost is less than for ion exchange using sodium as a regenerant or
nanofiltration. See Cost Estimate section.
5.3 Nanofiltration
5.3.1 General Process Description
Reverse Osmosis (RO) systems are becoming more commonly used for water treatment. Typical
RO units are high-pressure (250 psi or more) membrane-based systems for removing essentially all
dissolved materials and are excellent for desalination of seawater. However, the volume of
concentrate that has to be waste is quite high 30-50 percent. Low-Pressure RO and
nanofiltration (NF) require pressures as low as only 70 psi and are very effective in removing
hardness and other relatively large dissolved substances (see Table 11).
limit to recovery rate. Based on the relatively good quality of the well water, a recovery rate of
87.5% or more should be achievable if an appropriate anti-scalant is added to the feedwater.
Benefits Disadvantages
Removes 90% to 95% of hardness More water wasted (and more wastewater
(including manganese). Alkalinity removal to WWTP). About 7% to 10% waste overall.
does not affect WWTP.
Reduces TDS (by about 40-50% overall), More susceptible to fouling due to the
which improves taste. concentrated nature of reject water
Less chlorine use (about 40-50% less) and Some alkalinity may be lost due to pH
removes DBP precursors, which reduces adjustment needed to prevent scaling
DBPs by 20-50% overall and improves taste.
Two nanofiltration skids (and greensand) would give 1600 GPM total flow with no backup
unit. Three nanofiltration skids would give 2400 GPM total flow with no backup unit or 1600
with one backup unit. Each skid would contain virtually all of the equipment and would
measure about 23 feet by 8 feet. The cleaning/dosing solution tanks and equipment would
require about 16 feet by 8 feet.
The amount of water produced using 1750 GPM of well flow would be about 1600 GPM or
2.3 MGD. The total waste (assuming 3% waste from the greensand) would be about
140 GPM (about 8%).
These units would be installed in parallel with the greensand units and prior to chlorination.
Because of this arrangement, the feedwater to the NF will not have to be dechlorinated.
Feedwater will have to be conditioned with anti-scalants prior to introduction into the
membrane units. The feedwater to the membranes will have to be pressurized to about 70-
140 PSI with booster pumps.
The DBPs would likely be reduced by about 40-50%. Total chlorine usage would be reduced
by about 30-40%.
The capital cost of nanofiltration is about 60 percent higher than the cost of ion exchange
using acid as a regenerant. See Cost Estimate section.
The O&M cost of nanofiltration is about 10 percent less than for ion exchange using acid as
a regenerant. See Cost Estimate section.
Nanofiltration (or low-pressure reverse osmosis) is a feasible option for softening water from the
Citys wells. The total amount of water wasted can be limited to about 8 percent, including the
amount wasted in the greensand filters, by using a three-stage system. Nanofiltration has many
positive benefits: reductions in TDS, arsenic, organic chemicals, DBP precursors, and heavy metals.
At the target recovery of about 87.5 percent, membrane fouling is a concern and will need to be
well controlled with the addition of anti-scalant.
All other treatment methods are deemed infeasible or, as in the case of sodium ion exchange,
undesirable.
The design flow of 1600 GPM (for all alternatives) assumes 160 mg/L hardness and that 50 percent
of this flow bypasses softening. This design flow is sufficient to meet Maximum Day Demand
through 2017 or Peak Month Demand through 2031, while keeping hardness at 70 mg/L. If well
water hardness rises, then the produced water hardness will also rise. For example, at 175 mg/L
hardness and a flow of 1600 GPM, treated water hardness would to increase to 87 mg/L (which is
generally still an acceptable level) during the higher flow periods. Alternatively, the proportion of
softened water could be increased by increasing the softening system design flow.
Chlorination of the water before and after manganese treatment produces disinfection byproducts
primarily THMs, which need to be kept below 0.080 mg/L for compliance. Nanofiltration would
reduce THMs by as much as 50 percent. Cation exchange will only partially reduce THMs. Changes to
the greensand filtration system and disinfection system can be implemented in the future if
necessary to further control disinfection byproducts.
The well pumps will supply water to the system at a pressure of approximately 28 psi. This is enough
pressure for prefiltering and to charge the nanofiltration pressure pumps. The pressure drop across
the prefilter is 0 psi. The internal flows and pressures for a single 500-GPM NF unit are shown in
Table 13. The system feed pressure will be nominally 135 psi at 560 GPM; head loss through each
stage will result in a final pressure of 85 psi for the waste brine discharge. Permeate will flow by
gravity (not under pressure) to the chlorine contact chamber inlet piping, which is located just
beneath the floor of the Water Treatment Plant.
The piping to, from, and within the NF system will have pipe friction pressure losses close to
zero. Pressure loss is almost exclusively within the membrane elements - approximately 50 PSI
through 3 stages.
The maximum future flows to and from the NF system are 1,680 GPM and 1,500 GPM,
respectively, for which 10-inch piping is sufficient to keep head losses below 1 PSI.
The maximum flows to and from each NF unit are 560 GPM and 500 GPM, respectively, for
which 6-inch piping is sufficient to keep head losses below 1 PSI.
The maximum flow to each pressure vessel is 55 GPM, respectively, for which 1.5-inch or 2-inch
piping is sufficient to keep head losses below 1 PSI.
The Clean in-Place (System) will require a 380 GPM pump. At 65 PSI the piping would be 6-inch
diameter. At 70 PSI the piping would be 4-inch diameter. The waste discharge piping would be
6-inch diameter.
A permeate flush system will require a small pump and controls to direct permeate into the
pressure vessels at the onset of any shutdown periods.
Lead element
Feed flow GPM 51.1 43.9 34.7
Product flow GPM 5.7 4.5 3.0
Product TDS mg/L 5.3 22.4 88.9
Flux rate GFD 18.7 14.7 9.8
Last element
Product Flow GPM 4.6 3.2 1.7
Product TDS mg/L 19 75 251
Brine/Product Ratio ratio 4.3 6.5 11.9
Brine Flow (61 GPM total GPM 19.9 20.8 20.6
Net Driving Pressure PSI 100 67 36
Table 14. Capacity of Water Softening System for Maximum Day Demand
b. CIP Pump: 1 x 25 HP
2. 1-phase 110V power:
a. Valves, instrumentation, lighting
b. Anti-scalant pumps (2; plus 1 future)
c. CIP Tank heater
3. PLC for fully automatic operation (e.g., Allen Bradley MicroLogix with Panelview 1000
HMI)
4. MCC Panel
5. SCADA integration
8. Legal Considerations
There are no legal considerations for this project. All work will consist of modifications and additions
to the existing water treatment plant and all work will be inside the existing building or on the
building walls. This project does not increase capacity beyond that already planned for in the Citys
Water System Plan. Wastewater discharge will be to the adjacent City Wastewater Treatment Plant
and the discharge will be in compliance with all pretreatment requirements of the State and the
City.
Weekly Tasks
1. Review operating conditions chart trends for problems and to see if cleaning is needed.
2. Chemical In-Place (CIP) Cleaning System
a. Inspect storage tanks, fittings and gaskets for wear and tear.
Monthly Tasks
1. Reorder anti-scalant and cleaning chemicals, as necessary.
2. Perform chemical cleaning of softening system if needed.
Semi-Annual Tasks
1. Chemical Feed
Annual Tasks
1. Change the oil in all metering pumps.
2. Change the valve cartridge assembly on the sodium hypochlorite metering pump.
3. Inspect the metering pumps and replace if necessary
4. Check instrument calibrations (flow, pH, Conductivity)
A total of 53 points classifies this plant as Class II, which has a point range between 31 and 55 points.
This plant will require a Water Distribution Manager 2 (WDM2) and a Water Treatment Plant
Operator 2 (WTPO2) to supervise daily operations. The Citys current staff is certified to operate this
plant.
Prior to start-up, the system will be disinfected and pressure tested per DOH and AWWA standards.
Prior to start-up, the Contractor will provide operations and maintenance manuals for all
equipment.
Prior to any discharge of softened water to the chlorine contact chamber (and hence to the
distribution system), the treated water will be tested for compliance with water quality standards.
Start-up services will be provided by the Contractor and their certified and trained technicians, with
assistance by the Engineer and Water Treatment Plant operators. The proper functioning of all
components and the performance of each unit will be verified prior to discharge to the chlorine
contact chamber.
Treatment Plant Operators already have three months of experience operating a 1/16 scale pilot
system.
If any single component of the water softening system fails, the facility will still meet primary
drinking water standards. If one of the two (or three in the future) NF units goes offline, water
discharged to the distribution system will still be in compliance with all drinking water standards.
The system has two identical units operating in parallel. If one is offline, the treatment plant can still
meet the ADD flow demand for the MDD flow demand, in this situation, the hardness and TDS
would become elevated but the produced water would still be potable.
Brine Pressure Alarm This alarm indicates that the brine pressure
is too low or too high.
High or low pH indicates a malfunction. The
High or Low pH Alarm, shut down City will need to troubleshoot the problem
before the system can start up again.
During Chemical-In-Place cleaning (with acid), the system is isolated from the distribution system
with isolation valves. In addition, continuous reading pH meter with alarm will alert the operator
(24 hours a day) if the pH of the NF system or the finished water deviates from acceptable limits and
will automatically shut down the system. Turbidity and Conductivity meters will alert the operator of
any breakthrough of concentrate into the permeate stream.
A taste test was performed by a panel from Seattle Public Utilities on the unsoftened water and the
unsoftened water blended with softened water. The blended water had a superior taste compared
to the unsoftened water. The panel subjectively rated the blended water as I am sure that I could
accept this water as my everyday drinking water.
15. Summary
Six alternatives were considered and are listed below. The last three listed alternatives were
rejected as infeasible due to high cost. The first three were evaluated in detail.
Ion Exchange (IEX) (sodium regenerant) [the process most typically used]
Ion Exchange (IEX) (mineral acid regenerant)
Nanofiltration (NF)
Lime softening (a common process for larger water systems)
High rate solids contact clarifiers (a variant of lime softening)
Blending (with a soft water source)
The design flow used for the alternatives evaluation (1600 GPM total flow to distribution system)
meets Maximum Day Demand through 2017 (i.e., 9 years of 2.78% annual population growth) or
Peak Month Demand for 18 years (through 2030). The groundwater chemistry is compatible with
any of the three softening methods evaluated.
Ion exchange (IEX) is a feasible alternative for softening. The typical method of regenerating IEX
resin with sodium is unacceptable, because it increases sodium and total dissolved solids (TDS)
+
concentration in the water. Using mineral acid (H ) for regeneration is feasible because it adds no
sodium and because the groundwater is alkaline enough to buffer the acid added by treatment.
Nanofiltration (or low-pressure reverse osmosis) is a feasible alternative for softening. NF is identical
to reverse osmosis (RO), except that it does not remove as much sodium, chloride, or other small
molecules. However, NF wastes less water and uses much less power than high pressure RO. A
three-stage NF can recover up to 90 percent of the feedwater stream, which makes it a feasible
choice, though somewhat more costly to build and operate than a one- or two-stage system.
Both IEX and NF will effectively soften the water. For either of these alternatives, about half or more
of the incoming well water would be softened. The other half would bypass softening and be
treated only for manganese. These two flow streams would then be combined, resulting in a
hardness of about 70-80 mg/L. The advantages of IEX and NF are compared in the table below.
Reduces TDS by about 15%, which may Reduces TDS by about 50% overall, which
improve taste slightly. improves taste.
Less chlorine use (about 30% less), which Less chlorine use (about 50% less) and removes
reduces DBPs by 10-20% and improves taste. DBP precursors, which reduces DBPs by 50% and
improves taste.
System fouling minimal (acid regeneration Does not require handling large amounts of
removes iron and manganese). strong acid.
The estimated capital costs are shown below. NF has higher capital costs but lower operating costs.
NF addresses other water quality concerns such as TDS, chloride, DBPs, and arsenic, any of which
could become a compliance issue in the future.
The estimated O&M costs are shown below. These costs include primarily chemicals, power, minor
repairs, and funding replacement of resin or membranes. Costs for staff labor are not included
below.
Nanofiltration (NF) is the recommended softening system because it provides a good softening
system, the best overall treatment, the best improvement in taste, can better address potential
future compliance issues, and requires less operator effort.
The pilot test of the NF softening system operated for 92 days with no decrease in performance.
This is very good indication that the system will work well with a low level of maintenance over
time.
The water softening system will be installed in the existing Water Treatment Plant building. The
building has enough room for three 500-GPM units and associated equipment and piping. The
planned total of three 500-GPM units will allow for at least 2,550 GPM of capacity. A fourth unit
could be installed as well. The water treatment plant does not have space for a fifth unit.
Construction bid documents will be prepared in the fall of 2013, Construction of the system (with 2
NF units) will begin in early 2014 and be completed end of the summer of 2014. The third NF unit
will likely be installed in the year 2018 or later.
APPENDIX A
DRAWINGS
WATER SOFTENING PROJECT REPORT CITY OF FERNDALE
APPENDIX B
Phone 800-240-3330
Phone 952-448-4884
Fax 952-448-4886
Web WIGEN.COM
August 2013
PREPARED FOR:
WILSON ENGINEERING
Phone 800-240-3330
Phone 952-448-4884
Fax 952-448-4886
Web WIGEN.COM
TABLE OF CONTENTS
PAGE
SUMMARY ....................................................................................................... 1
MEMBRANE SKID CONFIGURATION .................................................................. 3
RO PILOT SYSTEM RESULTS ............................................................................ 6
RECOMMENDATIONS ....................................................................................... 16
APPENDICES
1. RO Projections
2. Laboratory Data
3. Log Data
4. Autopsy Report (In final report)
The pilot plant design was based on water analys es provided by Wilson Engineering shown in
Table 1. The pilot RO system was designed to closely simulate the flux rate, crossflow,
individual element recovery and overall system recovery of a full scale 3-stage system. The
preliminary water quality data used for perfor mance projections for the pilot system using
Toray TMG10 membranes is shown in Table 1.
The system was started up by Wigen Water Technologies personnel between June 7-14 and
City of Ferndale operators were trained on operation and data collection. The pilot system was
operated using a dedicated supply from one of t he Citys wells. The Citys operators recorded
all data and collected RO feed, RO perm eate, and RO concentrate water sam ples throughout
the pilot study. Water samples were analyzed by Edge Analytical Laboratories. Data was also
logged automatically by the RO PLC. The pilot plant was started up operating at a recovery of
87% as shown in the pr ojection in Appendix 1. Due to silic a levels in the f eed water and the
high target recovery rate, a silica specific anti-scalant, Vitec 4000 (Avi sta Technologies, Inc.),
was selected to achieve the targeted recovery rate.
On start-up, due to inco rrect calibration of the concentrate flow meter, the con centrate flow
was lower than expected which resulted in the pilot being operated at a recovery well above
87%. At this recovery the pressure drop quickly built up across the third stage due to scaling on
the membranes. A CIP was perform ed on the membranes on June 14 th, and the concentrate
flow was corrected to provide the desired 87% recovery. The CIP consisted of a single acid
clean of the m embranes at pH 2. Following this cleaning the interstage pressured returned to
start-up levels indicating the scaling had been removed. Very stable permeate flows and
conductivity levels through the pilot period demonstrate that there has been no fouling and that
the pilot design is satisfactory for scale-up for the full-scale system.
1
PreliminaryROPilotStudyReport
CityofFerndale,WA
SUMMARY
Table 1: Projected Feed Water Quality
Parameter Value
Fluoride, mg/l 0.28
Bicarbonate, mg/l as CaCO3 275
Sodium, mg/l 132
Silica, mg/l as SiO2 21
Potassium, mg/l 6
pH 8.2
Sulfate, mg/l 35
Magnesium, mg/l 18.5
Total Dissolved Solids (TDS), mg/l 657
Chloride, mg/l 133
Calcium, mg/l 33.3
Boron, mg/l 0
Strontium, mg/L 0.23
Barium, mg/l 0.08
Ammonia Nitrogen, mg/l as N 0.14
Nitrate Nitrogen, mg/l as N 0
Iron, mg/l NA
Manganese, mg/l NA
2
PreliminaryROPilotStudyReport
CityofFerndale,WA
MEMBRANE SKID CONFIGURATION
The membranes were configured in a 3-3-2- 2-1-1, 3-long array in order to keep the
equipment size to a minimum. This banking configuration gives the same results as a
3-2-1, 6-long array. A 6-long array will be used in the full scale equ ipment. Many
systems are piloted on a much smaller scale and sometimes require that the
concentrate be recycled in order to maintain adequate cross flow throughout the
system. This can change the feed water quality significantly depending on how much
concentrate is recycled, m aking it difficult to predict how the full scale system will
operate. Concentrate recycle was not used during this study so the results should be
very similar to what will be achieved by the full scale system. The results should also
3
PreliminaryROPilotStudyReport
CityofFerndale,WA
MEMBRANE SKID CONFIGURATION
offer an accurate representation of the need for m embrane cleanings and cartridge
filter replacements.
RO PRETREATMENT CHEMICALS
The RO pilot system included a chemical metering pump for the addition of
antiscalant. Vitec 4000, a product of Avista Technologies, was used as the
antiscalant. This antiscalant is designed to inhibit silica scaling at higher
concentrations than typical antiscalants which was necessary to achieve the high
recovery rate of 87%. Antiscalants are surface-active materials that interfere with
precipitation reactions using three mechanisms; threshold inhibition, crystal
modification and dispersion. As crystals begin to form at the subm icroscopic level,
negative groups located on the antiscalant molecule attack the positive charges on
scale nuclei interrupting the electronic balance that is necessary to propagate growth
of the crystal. The resulting effect is that precipitants that would normally fall out of
solution and deposit on the membrane surface stay in solution and are removed in the
concentrate stream.
Because of the small amounts of antiscalant needed for the low flows involved in this
pilot study, a 10% solution was used so that a reasonable frequency could be set on
the chemical pump. A 3 mg/l dose of Vitec 4000 was used in this study.
The chemical pump was contro lled automatically by th e RO PLC. When the inle t
valve opened, the pump was turned on and began pumping the prescribed dosage into
the RO feed stream just prior to the cartridge filter.
INSTRUMENTATION
The instrumentation provided on the pilot skid is identical to the full-scale units.
The pilot RO was supp lied with a variab le speed drive on the high pressure pump.
The concentrate flow was adjus ted manually by the operator using the concentrate
throttling valve. Perm eate flow was incr eased/decreased automatically by the high
pressure pump VFD and the concentrate flow was then adjusted using the concentrate
throttling valve. On th e full scale units, a VFD will m odulate the pu mp speed to
control the output pressure to achieve the desired permeate flow. The concentrate
valve can be automated or adjusted manually to achieve the desired recovery of 87%
on the full scale system.
A CIP system consisting of a pum p, 65 gallon tank and 5kW heat er was provided to
test the effectiven ess of clean ing the RO membranes. The pump was designed to
produce 30 gpm at 60 psig. The skid was provided with valves that enable each stage
of the RO to be cleaned independently at a flow of 10 gpm per vessel using either
high or low pH cleaners.
PROJECTED PERFORMANCE
5
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
RO RECORDED DATA
The following table contains a statistical analysis of the RO perform ance data that was recorded
during the pilot study from June 14 to Augus t 11, 2013. The normalized permeate flow and
conductivity was calcu lated based on startup conditions. The formulas used to normalize the
data will be discussed in the next section.
6
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
Grab samples were taken from the RO feed, permeate and Concentrate streams regularly
and sent to Edge Analytical Laboratories for analysis. The range of the laboratory
results taken up to July 9, 2013 are reported in Table 2 below for all stream s. This data
will be interpreted and discussed in the analysis section of this report. All of the
laboratory test results are included in Appendix 2.
Analyte ConcentrationRange
Feed Permeate Concentrate
pH
Alkalinity(mg/lasCaCO3) 244251 3.217.19 21052155
Hardness(mg/lasCaCO3) 145200.1 ND0.1 11901466.5
Sodium(mg/l) 2.63.0
Bromide(mg/l) 0.0130.014
AmmoniaNitrogen,mg/lasN ND
Sulfate(mg/l) 28 ND0.03 239248
Chloride(mg/l) 148167 2.32.5 2041374
Silica(mg/lasSiO2)
Arsenic(mg/l) 0.03 ND0.0003 0.0230.025
Iron(mg/l) ND0.01 ND 0.06
Manganese(mg/l) ND
TDS(mg/l) 551568 ND 47094819
7
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
DATA NORMALIZATION
The data recorded by the operators is considered raw data. Before any meaningful analysis of
the ROs perform ance can be conducted, the data must first be normalized. Perm eate flow
through a mem brane is affected by water temperature, applied pressure and osmotic pressure.
Normalization of the data to startup conditions must be perform ed to compensate for day to day
changes of these parameters. After normalization, meaningful trends can be spotted in the
operation of the RO system.
For the full scale system, daily op erating data w ill be recorded directly into a no rmalization
spreadsheet that will plo t the normalized trends and allow the operators to m ake determinations
about potential problems and/or the need for m embrane cleaning. Generally, when norm alized
permeate flow decreases by 15% or normalized permeate conductivity increases by 15%, a
cleaning is needed.
The equations used to norm alize the data are discussed on the following tables. A spreadsheet
provided with the full scale equipment will calc ulate the v alues automatically. To understand
what these numbers mean in regard to the pe rformance of the RO sy stem, a good understanding
of the theory behind normalization is required and is outlined on the following pages.
8
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
EQUATION:
Qs = Pfs Ps Pps - Pfcs - fcs TCFs Qo
Pfo Po Ppo - Pfco - fco TCFo
fc =((0.0117 x Cfc) - 34 / 14.23 X (T + 320)/345)/.0689 (psi) For Cfc > 20000 mg/l
fc = (Cfc x (T+320) / 491000)/.0689 (psi) For Cfc < 20000 mg/l
Cfc = Concentration of the feed-concentrate
9
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
Assume both values are the same. The source water and the machine recovery
don't tend to change enough to affect normalization and therefore = 1
EQUATION:
Pfo Po Ppo - fcs- po Cfcs
Cps = Cpo
Pfs Ps Pps - fcs - ps Cfco
Assume that the osmotic pressure of the
permeate stream is neglible and therefore = 0
10
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
DATA ANALYSIS
All the data for this pilot study was compiled in a single Excel file , and is included in Appendix 2
(to be included in final report).
The pressure drop pre-filter throughout the pilot study was very consistent. The differential
pressure across the filter, at an overall feed flow of 34.48 gpm, ranged between from 0 to 6 psi
over the course of the pilot study. This filter was replaced on August 7, representing a
replacement frequency of approximately 2 months.
30.00
20.00
Normalized Permeate Flow
Linear (Normalized Permeate Flow)
10.00
0.00
05-Jun-13 15-Jun-13 25-Jun-13 05-Jul-13 15-Jul-13 25-Jul-13 04-Aug-13 14-Aug-13
Date
The system was designed to be operated using a permeate flow of 30 gpm and a concentrate flow
of 4.48 gpm to achieve a recovery of 87%. Actual perm eate flows ranged from 30 gpm to 32
gpm, and a veraged 31 gpm during the study. Concentrate flows ranged from 4.7 gpm to 4.9
gpm, and averaged 4.8 gpm . Above is a graph of the process flows dur ing the pilot study up to
August 11. So far, all flows have remained consistent throughout the study.
11
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
Differential pressure drop across the membranes is an indication of foulants that may be blocking
the feed sp acers within the m embranes. Over the course of the study, differential pressures
across the 18 membranes in series ranged from 50 to 61 psid or approximately 2.8 3.1 psid per
membrane on average. This is well within normal manufacturer specifications for 40
membrane elements. Normal operating pressure drop was approxim ately 58 psid and remained
stable throughout the pilot. No indication of fouling is present.
60.0
50.0
40.0
30.0
Machine Pressure Drop (Psi)
20.0 Linear (Machine Pressure Drop (Psi))
10.0
0.0
05-Jun-13 15-Jun-13 25-Jun-13 05-Jul-13 15-Jul-13 25-Jul-13 04-Aug-13 14-Aug-13
Date
12
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
4.00
2.00
0.00
05-Jun-13 15-Jun-13 25-Jun-13 05-Jul-13 15-Jul-13 25-Jul-13 04-Aug-13 14-Aug-13
Date
The graph above shows normalized and actual permeate conductivity over time. The normalized
permeate conductivity is stable through the pilot study period.
13
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
SRSystem
C f Cp
x100
Cf
C f Feed Concentrat ion
C p Permeate Concentrat ion
Actual membrane rejection is calculated differently since, as the water travels through the
machine, the feed concentration changes. To calculate actual m embrane salt rejection, the
following equations are used:
SRmembrane
C avg Cp
x100
Cavg
Cavg
C f Cc
2
C f Feed Concentrat ion
C c Concentrat e Concentrat ion
C p Permeate Concentrat ion
Actual Membrane Rejection (SR membrane) varied from 99.59% to 99.52% with an average of
99.56%.
Manganese is present in the feed water at around 0.2 m g/L. Feed water concentrations have not
been measured to date, but som e measurements will be taken ove r the latter part of the study.
The Manganese level in the RO Permeate was measured and was below detection.
TDS, hardness and sulfate levels were also reduced to below detection.
Chloride and sodium levels were reduced to the 2-3 mg/L range.
14
PreliminaryROPilotStudyReport
CityofFerndale,WA
RO PILOT SYSTEM RESULTS
MEMBRANE CIP
Wigens technician conducted a single Acid CIP of the entire RO skid on June 14 th, 2013
following a buildup in pressure in the third stage due to the recovery being set too high. This
cleaning was conducted at a pH of 2 (starting pH) and circulated for 30 minutes, then stood idle
for 30 minutes with a final circulation of another 30 minutes.
Following the CIP the third stage pressure returned to start-up levels indicating the build-up of
scale had been successfully rem oved. Following the CIP the antiscalant dose was also increased
from 2 to 3 mg/L to minimize the risk of further scaling.
MEMBRANE AUTOPSY
At the conclusion of the pilot study following a CIP an autopsy shall be conducted on the last
(tail) element in the final stage of the pilot plant to determine if there are any foulants remaining
on the membrane.
The full autopsy report shall be included in the final pilot report as Appendix 4 - Autopsy.
15
PreliminaryROPilotStudyReport
CityofFerndale,WA
RECOMMENDATIONS
In summary, the piloted flux rate of 13.95 GFD a nd recovery of 87% al ong with the chem ical
dosing used in the pilot study has allowed for effective application of reverse osmosis on the
Ferndale well water supply. Overall m embrane rejection was good and reduced dissolved solids
by 98.8% on average. No increase in feed pressure was noted, and no change in permeate quality
over the pilot study has been observed. .
Prefilter change outs sh ould not ne ed to occur more than once every 6-8 weeks based on the
performance during the pilot study. No membrane fouling has been observed using a 3.0 mg/l
dose of Vitec 4000 antiscalant over the course of the pilot study.
The full scale system would be designed to produce approximately 400 gpm of RO permeate per
skid. At 87% recovery, 460 gpm would need to be feed to each RO skid.
The RO would likely be designed with a banking of 9-6-3:6L at a flux of 13.33 GFD using Toray
TMH20A-400 membranes. Assum ing the system operates at the piloted 3 mg/l antiscalant
injection rate, the chemical consumption would be as follows:
Vitec 4000 Antiscalant 28.38 lb / m illion gallons of permeate water produced. Market price
for Vitec 4000 is $4.12 / lb, when purchased in 2500 lb tote bins. Total cost per year for
Antiscalant is $24,581 / year, assuming one RO skid is running 100% of the time.
Assuming a cost of $0.08/kW-hr, electrical consumption would be approxim ately $115 / million
gallons of permeate water produced, or $24,193/year at 100% production rate of one RO skid.
This is the cost of operating the high pressure RO pump only.
Cartridge filter change o ut frequency should not exceed once every 42 days so the annual cost
would be approximately $1,000/year
16
PreliminaryROPilotStudyReport
CityofFerndale,WA
APPENDIX 1 RO PROJECTIONS
17
PreliminaryROPilotStudyReport
CityofFerndale,WA
)%'.##'$'+'(%(#%(((.()#($@175
&B?:53D>1=55B>41<5,D?B1I)#&9<?D&B?:53D9?>1C51D5
&B?:53DD9D<5
&B?7+5BC9?> *@41D5
E>8<+5BC9?>
*@41D5
#1I
'%(.()#)
554G1D5BDI@5B13;9C8,5<<?B(?6D5>54)1@G1D5B
554D5=@5B1DEB5
457 45C97>
'%6554@B5CCEB5
@C916D5BI51BC
)?D1<4966 @B5CCEB5
@C965542B9>59>3< 4&=1>96?<4C@9@5
B9>5@B5CCEB5
@C91D5H9D6B?=<1CD21>;
5C97>@5B9?4 I51BC
?E<9>713D?B 16D5BI51BC
(1<D@1CC1759>3B51C5 16D5BI51BC@5BI51B
(9<D45>C9DI9>45H 9>'%6554(0
(ICD5=B53?F5BI
(ICD5=65546<?G @= @4
&B?4E3D6<?G @=
@4
F5B175CICD5=6<EH =
4
<
=
8
B9>56<?G @= @4
554C1<9>9DI =7
<1C9?>
&B?4E3DC1<9>9DI =7
<1C9?>16D5BI51BC
B9>5C1<9>9DI =7
<1C9?>
&B?:53D>1=55B>41<5,D?B1I)#&9<?D&B?:53D9?>1C51D5
&B?:53DD9D<5
,)'$".((%B97)B51D54?>3 H@
1C=7
<9?>65546554B9>5&5B=51D5
1<39E=
#17>5C9E=
(?49E=
&?D1CC9E=
==?>9E=
1B9E=
(DB?>D9E=
8<?B945
(E<61D5
931B2?>1D5
1B2?>1D5
<E?B945
(9<931
B55%
)?D1<1C9?>
)?D1<1C=5A
<
E(
3=
%
E(
3=
@+1<E5
"1>75<95B>45H
(D9661F9C>45H
&')')#$)
#51CEB51719>CD1%C31<5$?>5
(%1>4%(31<59>8929D?B$?>5
(1DEB1D9?>9>45H9>2B9>5 =1H (D9661F9C>45H
'%+'."#)(
6?B&0B9>5(1DEB1D9?>#1H '53?F
1(%
1(%
(B(%
1
(9<931
&%())')#$)
#51CEB51719>CD@5B=51D5$?>5
9C3<19=5B
)85@B?7B1=9C9>D5>454D?25EC542I@5BC?>C81F9>7D538>931<C;9<<1DD859B?G>49C3B5D9?>
1>4B9C; )85@B?:53D9?>C?2D19>54G9D8D85@B?7B1=1B5D855H@53D54CICD5=@5B6?B=1>35
21C54?>D851F5B175>?=9>1<5<5=5>D@5B6?B=1>351>41B5>?D1ED?=1D931<<I7E1B1>D554
)?B1IC81<<>?D25<912<56?B1>I5BB?B?B=9C31<3E<1D9?>9>D85@B?7B1=
)85?2D19>54B5CE<DC31>>?D25EC54D?B19C51>I3<19=6?B<9129<9DI?BG1BB1>DI
D9CD85EC5BCB5C@?>C929<9DID?=1;5@B?F9C9?>C1719>CD6?E<9>7C31<9>71>4385=931<
1DD13;CD?133?E>D6?B@9@9>71>4F1<F5@B5CCEB5<?CC5C6554@E=@CE3D9?>@B5CCEB51>4
@5B=51D5213;@B5CCEB5 ?BAE5CD9?>C@<51C53?>D13DEC
)?B1I>4ECDB95C>3 '%#5=2B1>5&B?4E3DC5@D
#981=138?=5*B1I1CE8921
1@1>
)"-
)?B1I#5=2B1>5*(>3
)81D385B?EBD&?G1I
*(
)"
-
)?B1I#5=2B1>5EB?@5
B125>13;5BCDB1CC5& % ?H
#E>385>CD59>(G9DJ5B<1>4
)"-
)?B1IC91&D5 "D4
)"
-
&BE45>D91<)?G5B539<(DB55D(9>71@?B5
)?B1I<E5CD1B#5=2B1>5? "D4
)5<
1H
/?>5)91>J8E9B@?BD>4ECDB91</?>559:9>789>1
*'"8DD@
)&*0*+$+
776I3F7DFKB7D35=;E:.7>>AD*A8F7@76+3BI3F7D
776F7?B7D3FGD7
679 67E;9@
)&8776BD7EEGD7
BE;38F7DK73DE
+AF3>6;88 BD7EEGD7 BE;87764D;@7;@5> 6'?3@;8A>6EB;B7
D;@7BD7EEGD7 BE;3F7J;F8DA?>3EF43@=
7E;9@B7D;A6 K73DE
AG>;@935FAD 38F7DK73DE
*3>FB3EE397;@5D73E7 38F7DK73DEB7DK73D
*;>F67@E;FK;@67J ;@)&8776* 2
*KEF7?D75AH7DK
*KEF7?87768>AI B?
B6
'DA6G5F8>AI B? B6
H7D397EKEF7?8>GJ ?
6
>
?
:
D;@78>AI B? B6
776E3>;@;FK ?9
>3E;A@
'DA6G5FE3>;@;FK
?9
>3E;A@38F7DK73DE
D;@7E3>;@;FK
?9
>3E;A@
')+)+$%+
$73EGD7393;@EF3&E53>7%A@7
*&3@6&*53>7;@:;4;FAD%A@7
*3FGD3F;A@;@67J;@4D;@7 ?3J *F;883H;E @67J
)&-)0# $ +*
8AD '2D;@7*3FGD3F;A@$3J )75AH
3*&
3*&
*D*&
3
*;>;53
'&*++)+$%+
$73EGD7393;@EFB7D?73F7%A@7
+&)0$$)%)-)*&*$&* **0*+$* %B397
% - ,##$%++
3@=>7?7@F%A
7768>AIB?
>7?7@F+$'BE;
7>F3BBE;
776&E?AF;5BE;
$7?4 &E?AF;5BE;
D;@7&E?AF;5BE;
776+*BB?
D;@78>AIB?
D;@7+*BB?
'7D?73F78>AIB?
'7D?73F7+*BB?
'7D?73F7G*
BDA8;>7G*
BDA8;>7G*
)75AH7DK
)7<75F;A@%3>
'3EE
3@=
>+%A
('?
6
'2&E?
'2D;H
%3>)7<
+AD3K$7?4D3@7,* @5
+:3F5:7DAGDF'AI3K
,*
+#
/
+AD3K$7?4D3@7GDAB7
D347@35=7DEFD3EE7' & AJ
$G@5:7@EF7;@*I;FL7D>3@6
+#/
+AD3KE;3'F7 #F6
+#
/
'DG67@F;3>+AI7D75;>*FD77F*;@93BAD7
+AD3K>G7EF3D$7?4D3@7A #F6
+7>
3J
1A@7+;3@L:G;DBADF @6GEFD;3>1A@77;<;@9:;@3
,)#:FFB
PreliminaryROPilotStudyReport
CityofFerndale,WA
APPENDIX 3 LOG DATA
PreliminaryROPilotStudyReport
CityofFerndale,WA
APPENDIX 4 AUTOPSY
PreliminaryROPilotStudyReport
CityofFerndale,WA
WATER SOFTENING PROJECT REPORT CITY OF FERNDALE
APPENDIX C
Permeate ALKALINITY mgCaCO3/L 4.3 3.21 7.19 4.96 4.46 3.57 5.47 5.27 6.17 6.91 5.68 5.2 24% 2
Permeate AMMONIAN mg/L ND ND ND 0.14 0.06 0.01 0.07 94% 0.03 0.0112
Permeate ARSENIC mg/L ND 0.0003 ND 0.0003 0.0003 0.0003 0.0005 0.0003 0.0004 0.0004 0.0004 0.00036 20% 0.001 0.00003
Permeate BROMIDE mg/L 0.013 0.014 0.011 0.012 0.013 0.012 0.013 8% 0.005 0.0002
Permeate CHLORIDE mg/L 2.3 2.3 2.5 2.7 2.7 3.1 2.6 2.9 3 2.9 3 10% 0.1 0.014
Permeate HARDNESS mgCaCO3/L ND ND 0.1 0.3 0.3 0.1 0.08 0.09 ND ND 0.2 0.17 59% 3.3 0.055
Permeate HYDROGENSULFIDE mg/L ND ND ND ND ND 0.1 0.1
Permeate IRON mg/L ND ND ND 0.006 ND 0.005 ND ND ND ND ND ND 0.05 0.004
Permeate MANGANESE mg/L ND ND 0.0005 ND 0.008 0.004 125% 0.005 0.0012
Permeate SILICA mg/L 0.34 0.34 0.05 0.002
Permeate SODIUM mg/L 2.6 3 4.1 2.8 3 4.1 3.3 20% 1 0.03
Permeate SULFATE mg/L ND 0.03 ND ND ND 0.1 0.03 0.06 ND 0.06 0.1 51% 0.2 0.016
Permeate TOTALDISSOLVEDSOLIDS mg/L ND ND ND ND 11 13 14 14 4 10 ND 11 34% 10
Concentrate ALKALINITY mgCaCO3/L 2155 2120 2105 2095 2080 2080 2035 2040 1990 2015 1980 2063 3% 2
Concentrate ARSENIC mg/L 0.025 0.025 0.023 0.025 0.024 0.025 0.032 0.032 0.031 0.023 0.023 0.026 14% 0.001 0.00003
Concentrate CHLORIDE mg/L 1218 1204 1374 1362 1360 1401 1397 1445 1501 1535 1527 1393 8% 10 1.4
Concentrate HARDNESS mgCaCO3/L 1190 1467 1648 1594 1534 1593 1616 1529 1632 1636 1533 1543 8% 3.3 0.05
Concentrate HYDROGENSULFIDE mg/L ND ND ND ND ND 0.1 0.1
Concentrate IRON mg/L 0.06 0.06 0.06 0.05 0.05 0.1 0.04 0.05 0.05 0.06 0.053 0.06 27% 0.05 0.004
Concentrate SULFATE mg/L 243 239 248 243 243 237 238 238 232 242 241 240 2% 0.2 0.02
Concentrate TOTALDISSOLVEDSOLIDS mg/L 4718 4709 4819 4883 4780 4820 4824 4844 4956 4880 4872 4828 2% 10
NDNotDetected
PQLPracticalQuantitationLimit
MDLMethodDetectionLimit
PILOT STUDY FIELD DATA
Cfc(uS)Conc of Feed-Concentrate
Temperature Correction Factor
(Psi)Machine Pressure Drop
(psi)Prefilter Pressure Drop
(psi)Interstage #1 Pressure
(psi)Interstage #2 Pressure
(psi)Permeate Pressure
(Qp+Qc)Recovery Qp/
(psi)Primary Pressure
(psi)Prefilter Pressure
Passage Cp/Cavg %
Conc Flow (gpm) Qc