Engineering Design and Analysis (E 10)

Household Water Treatment in Less Developed Countries

Introduction

According to the United Nations, over 1 billion people do not have access to safe drinking water. As a result of disease-causing waterborne pathogens in their water, most of these people suffer from outbreaks of waterborne diseases such as cholera, amoebic dysentery and hepatitis, and each year over 2 million people die from various waterborne diseases. While unsafe drinking water is not the sole cause of waterborne diseases, it is clear that treatment of drinking water supplies can lead to a significant decrease in the incidence of waterborne disease.

During the past fifty years, most efforts to improve the quality of drinking water in less developed countries has focused on the construction of large drinking water treatment plants capable of providing water for hundreds of thousands of people. While this approach often is the most cost effective means improving drinking water quality in densely populated areas, it is less effective in rural communities or in cities where water distribution infrastructure is lacking. To serve these people, engineers have begun to develop water treatment systems for use in individual households. The design of household water treatment systems and the implementation of programs that use these systems provide good examples of the many challenges engineers have to overcome as they solve important problems facing society. This module will provide you with a better understanding of water treatment, the tradeoffs inherent in designing systems that employ technologies with different levels of sophistication and the need to consider local culture and available infrastructure when trying to solve problems in less developed countries.

Your team will be provided with information on the physical, chemical and biological characteristics of a water source that your household system must be able to treat. During the first three weeks of the module, you will evaluate design parameters for home water treatment systems. Each system relies upon a different approach to inactivate or remove disease-carrying organisms from drinking water and the design of each system will depend on the characteristics of the water that is being treated. During the fourth week, you will compare the costs of the systems and evaluate the barriers to successful implementation of each system. Finally, you will present a conceptual design of a household water treatment system for your water to the class during our final meeting.

Background

Why household water treatment?

The US National Academy of Engineering ranked the development of water treatment and distribution systems as the fourth most important engineering achievement of the 20^th century (after electrification, automobile and airplanes; see http://www.greatachievements.org/). Prior to the widespread construction of water treatment systems, diseases such as cholera, dysentery and typhoid fever were common in the United States. Outbreaks of these diseases led to death or debilitating illness (e.g., typhoid fever claimed the lives of 25 out of every 10,000 people each year). In the 21^st century, we take the safety of our tap water for granted and outbreaks of waterborne diseases are a rarity.

The remarkable public health benefits of water treatment have been slow in coming to less developed countries. In many countries, the current situation is no different from the conditions in the United States one hundred years ago: widespread outbreaks of diseases and inadequate water treatment are commonplace in much of Sub-Saharan Africa and Southeast Asia. In other countries, affluent consumers have access to high quality, safe drinking water while the poor either drink unsafe water or spend a considerable portion of their income to purchase water from vendors selling treated water from trucks or street vendors. In recognition of the need to provide safe drinking water, many countries and international aid agencies have invested in large-scale water treatment plants. More recently, private companies have started to build and operate large drinking water treatment systems in developing countries. Unfortunately, these centralized water treatment systems have had mixed records for delivering water to the urban poor and rural farmers who often live in areas that are not served by modern treatment plants.

During the past decade, some engineers have started to develop new approaches for providing people with safe water by designing small-scale water treatment systems. Simple, inexpensive systems have been designed that have the capacity to treat water for a small village, individual wells or households. The design of household treatment systems is particularly challenging because the systems require individuals with little experience in water treatment or modern technology to become the operators of treatment systems that need to function without failure for months or years between major repairs or installation of replacement parts.

How do household water treatment systems work?

Water treatment systems function by inactivating (i.e., killing) pathogens or by removing them from drinking water by physical processes (e.g., filtration). Many different types of disease-causing organisms are potentially present in drinking water sources and entire courses are taught at Cal on the biology of these organisms. For this module, we will focus our attention on three types organisms that often are responsible for disease outbreaks associated with unsafe water: viruses, bacteria and protozoa. We will consider a representative organism from each group to illustrate how the size of the organism and their resistance to treatment dictates the design of water treatment systems. The organisms that we will consider are:

Rotavirus: In less developed countries, diarrhea is one of the leading causes of death in children who are malnourished and otherwise stressed by other diseases. Rotavirus is believed to be responsible for 30-50% of the cases of severe diarrhea in humans. As you may have surmised, rotavirus is a virus, which is a small (~80 nm) and simple organism, consisting mainly of a protein coat and RNA capable of infecting a host. People infected with rotavirus excrete approximately 100 billion viruses per day in their feces. For comparison, exposure of a person to a dose as low as 10 viruses can result in illness. Thus, rotavirus can easily be transmitted by poor hygeine (e.g., not washing hands) or by consumption of contaminated drinking water. The small size of rotavirus and most other viruses make them relatively easy to inactivate but difficult to remove from water by physical treatment processes.

Vibrio cholerae: These organisms are responsible for outbreaks of cholera, a disease that infects over 5 million people each year, resulting in severe illness and hundreds of thousands of deaths. V. cholera is a bacterium with a diameter of approximately 1 mm. Unlike viruses, bacteria are capable of surviving and multiplying outside of their human hosts. Because they live and reproduce outside of their hosts, outbreaks of these organisms are difficult to control. Compared to viruses, much larger numbers of these organisms are needed to cause illness and therefore, it is usually easier to design a water treatment system to protect people from cholera outbreaks.

Giardia lamblia: This organism is a protozoa that causes giardiasis, a disease that is endemic worldwide, causing diarrhea and intestinal pain in over 25 million people per year. Giardia also is a concern in the United States, especially in communities that use unfiltered surface water (e.g., San Francisco, New York). Giardia exists in two forms: as trophozoites (which is the form that lives inside of its human host) and as cysts (which is the form that is excreted). Infections can occur after ingestion of as few as 1 cyst. Therefore, water treatment systems must be able to remove the cysts, which are approximately 10 mm in diameter and are protected with a thick, protein-containing outer coating that protects the organisms.

Table 1: Summary of information on waterborne pathogens of concern. L_Cl and L_UV are defined on pages 4 and 9.

Pathogen	Minimum Infective Dose (#)	Diameter (mm)	LCl (L/mg min)	LUV (cm^2/mW sec)
Rotavirus	10	0.08	2.3	0.12
V. cholera	1,000,000	1.0	140	0.30
G. lamblia	1	10	0.070	0.23

The design of household water treatment systems is dictated by the characteristics of the water to be treated. To illustrate how these differences affect the design of water treatment systems, each team of students will evaluate household water treatment systems for a specific type of water, indicated in Table 2. The situations that you will consider in when you design a water treatment system in the final part of the module also are described in Table 2 with html links to related sites on the course homepage. Your instructor will assign you one of these sources during the first class meeting. The different water characteristics and their relevance to treatment system design are summarized below:

Pathogen concentration and infective dose: These two characteristics will determine the extent of treatment that is needed to prevent infection. For example, if a water sample contains a million rotaviruses per liter and the minimum infective dose is 10 viruses, the treated water would have to contain less than 5 viruses per liter, assuming that the user consumes 2 liters of water per day. This would require that the system remove over 99.999% of the viruses in the water. When designing water treatment systems, engineers often include safety factors to account for possible variations in the concentrations of pathogens in the water or less-than-optimal performance of treatment systems. As you design your system, you will need to weigh the safety factor against the costs of achieving the lower-than-needed concentrations of pathogens and other design factors.

Total suspended solids: Water normally contains suspended particles (e.g., silt, clay, bacteria). These suspended particles can affect water treatment processes in several ways. For treatment processes that inactivate pathogens through exposure to chemical disinfectants such as chlorine, the suspended particles can protect the organisms from the disinfectant. For treatment processes that remove the pathogens by physical processes, suspended particles can increase the efficiency of the process as the pathogens associated with large particles are easier to remove than free-living organisms. Of course, suspended particles also can complicate treatment by clogging filters and depositing on surfaces. The amount of particles present in water depends upon the source of the water and typically varies over two orders of magnitude: 100 mg/L of suspended solids is typical of a muddy river while 1 mg/L is typical of shallow groundwater.

Other water quality parameters: A number of other factors including pH, temperature and dissolved light-absorbing substances could affect the treatment process. Although we will not explore these factors during this module they are still important to the design of treatment systems.

Table 2: Characteristics of waters to be considered for household water treatment.

#	Name	Pathogen	Log (Initial Conc’n., #/L)	Suspended Solids (mg/L)	a (cm^-1)
1	Las Animas, Mexico^1	Rotavirus	1 x 10^6	100	0.5
2	Las Animas, Mexico^1	V. cholera	1 x 10^8	100	0.5
3	Las Animas, Mexico^1	G. lamblia	1 x 10^3	100	0.5
4	Kakuma Refugee Camp^2	Rotavirus	1 x 10^7	10	0.1
5	Kakuma Refugee Camp^2	V. cholera	1 x 10^9	10	0.1
6	Kakuma Refugee Camp^2	G. lamblia	1 x 10^4	10	0.1
7	Borgne. Haiti	Rotavirus	1 x 10^5	1	0.05

¹ This village is located in a rural region of Baja California Sur, Mexico where surface water from a small river serves as the main water supply. The water is pumped up to the houses by a gasoline-powered pump and is stored in 200 L tanks until it is used. Normally, the suspended solids load is low. However about ten times per year it rains and the suspended solids concentration increases to approximately 100 mg/L. The community has access to electricity through a rural solar energy program supported by the Mexican government. Transport of goods and materials occurs on a poor quality highway. A map is linked to the html version of this handout (www.4x4abc.com/baja/animas.html).

² Although there is a water distribution system at the refugee camp, we will assume that the water is contaminated and cannot be treated in a centralized treatment plant. The web site that is linked to the html version of this handout (www.lwfkenyasudan.org/kakuma_refugee_camp.htm) provides background in formation on the refugee camp.

³ This is a rural community with a population of about 5,000 people located along the North Coast of Haiti. The main water supply consists of a capped spring located approximately 5 km from the town. Water flows to the town by gravity through a pipe system that is prone to leaks. Testing has identified significant concentrations of rotavirus in the water supply.

Assignment #1: Chlorine Disinfection of Water

The first treatment method that we will explore is chlorination. In this treatment process, hypochlorous acid (i.e., HOCl) or sodium hypochlorite (NaOCl, which you may know as household bleach) is added to water. These chlorine compounds are poisonous to pathogens and inactivate the organisms by reacting with in their cell membranes and cellular constituents. Chlorine inactivates pathogens relatively quickly (i.e., typically in several minutes) with faster rates of inactivation at higher chlorine concentrations. Experimentally, it has been determined that the inactivation of the organisms is determined by the dose of chlorine administered as expressed by the product of the concentration and the contact time. Thus, adding 10 mg/L of chlorine compounds to a water sample and allowing it to react with the pathogens for 10 minutes will result in as much pathogen inactivation as exposing the water sample to 100 mg/L of the chlorine compounds for 1 minute (i.e., 10 mg/L*10 min = 100 mg/L* 1 min = 100 mg/L min). The product of the concentration and time is referred to as the C*t value. In full-scale treatment systems it is necessary to consider the decay of chlorine during the treatment process by integrating the area under the C*t curve, but for our purposes, we can assume that the chlorine concentration will remain constant during treatment, which allows us to multiply concentration by time to estimate C*t.

The inactivation of pathogens can be expressed in terms of C*t with the following equation:

[equation 1]

Where:

N = Concentration of active organisms remaining after treatment (#/L)

N₀ = Initial concentration of organisms (#/L)

L_Cl = Coefficient of specific lethality (L/mg min)

C = Concentration of chlorine, expressed in mg of chlorine/L

t = time (minutes)

The units used to express chlorine concentrations often confuse students. When you are given chlorine concentrations (e.g., on a bottle of bleach or in Table 1) the concentration is mg/L, which refers to mg/L of chlorine as Cl₂. To convert this value to moles per liter, we divide by the molecular weight of chlorine (71 g/mole). Here comes the confusing part: the disinfectant that we are add to water is NaOCl (bleach) and not Cl₂. To convert from mg/L of Cl₂ to the amount of NaOCl that you need to add to water, you will need to convert mg/L of Cl₂ to moles per liter of OCl^- by taking into account that each mole of Cl₂ makes one mole of OCl^- (i.e., Cl₂ + H₂O à 2H⁺ + OCl^- + Cl^-).

Thus, it is possible to predict the concentration of organisms remaining after treatment as a function of C*t provided that coefficient of specific lethality is known.

In our next class, we will evaluate a simple method of household water treatment by chlorination. In this process, the home water user will add sodium hypochlorite to water in a
5-gallon bucket. Your job is to come up with a simple, inexpensive way of accurately and precisely adding bleach to water. In class on Friday we will attempt to add 1 mg/L of chlorine (as Cl₂) to 20 L of water. Your assignment is to design a way to add the appropriate amount of bleach to water the bucket of water that would work for someone in a rural household. You must obtain the devices needed for measuring out the bleach and bring them to class. On Friday, we will decide upon the most practical method and test it out.

Question to be answered prior to second meeting:

Please hand in your answers to the following questions at the start of our second meeting.

1. For your water (Table 2), determine the volume of stock sodium hypochlorite (bleach) that must be added to 20 liters of untreated water to produce drinking water that will not cause illness in a person who drinks 2 liters per day of water. Make your calculation for a contact time of 1, 10, 100 and 1000 minutes assuming the concentrations of pathogens, infective dose and specific lethality coefficients listed in Tables 1and 2. Bleach is 5.25% sodium hypochlorite (i.e., 52.5 gm of NaOCl per kg of solution) and the density of the solution is 1.085 gm/mL. The molecular weights of the atoms in bleach are as follows: Na = 23.0 gm/mole; O = 16.0 gm/mole; Cl = 35.5 gm/mole.
2. How long will a 1-L container of bleach last for a family that is using chlorine to treat your water? Show any assumptions needed to make this estimate.

Second Meeting: Chlorination for Household Water Treatment

What do we have to do now?

As discussed previously, sodium hypochlorite can be used for household disinfection of drinking water. To employ this disinfectant in a household water treatment program, a centralized agency or non-governmental organization provides households with a bottle of sodium hypochlorite (i.e., bleach) and a container for water storage. The organization also has to provide simple instructions for adding the bleach to water that results in the delivery of an accurate dose of chlorine that would be sufficient to prevent disease.

During the first part of the class, the class will decide upon the best possible way to add chlorine to a 20-liter container of untreated water. For example, you may instruct the users of the system to add a certain number of drops of bleach with a medicine dropper and wait for 10 minutes before consuming the water. During the second part of the class, each person will use the agreed upon procedure to add enough chlorine to obtain a target dose of 1 mg/L of chlorine. We will then measure the actual concentrations added by each group and provide you with the data.

It is highly unlikely that everyone will end up with the exactly 1 mg/L of chlorine in the treated water. It is more likely that the chlorine concentrations will follow a distribution like the one depicted by the solid line in Figure 1:

Figure 1: Possible distributions of chlorine concentrations. For the solid line, the mean concentration is 1 mg/L and the standard deviation is 0.1 mg/L. The dotted line has a mean of 1.2 mg/L and a standard deviation of 0.1 mg/L.

As part of the design process, you would want to use information on the distribution of concentrations to determine the appropriate concentration of chlorine to add. For example, if the distribution that the class obtains looks like the solid line in Figure 1, you might increase the target concentration of chlorine to 1.2 mg/L (dotted line in figure 1) to assure that almost everyone always ends up with a dose greater than 1 mg/L. (For the dotted line, only 5% of the samples would be expected to contain less than 1.0 mg/L of chlorine.) However, if the standard deviation is too large (i.e., if the distribution is too broad), at some point you would need to start considering the possibility that some people might end up with more than 2 mg/L of chlorine (i.e., the value where the water starts to become unpalatable).

How do we analyze the data?

The data that you collect should follow a normal distribution similar to the shape of the curve in Figure 1. If so, you can estimate the mean and standard deviation of the data. Provided that a large enough number of samples has been collected, you can then assume that approximately 25% of the samples will contain chlorine at concentrations that are less than one standard deviation below the mean (i.e., -s) and 5% of the samples will contain chlorine concentrations less than two standard deviations below the mean (i.e., -2s).

Determine the mean and standard deviation of the data. You can determine standard deviation (s) with Excel’s data analysis function or by using the formula shown below:

[equation 2]

where:

s = standard deviation of the sample

X = any observation

= mean concentration

n = total number of measurements in the sample

Questions to be answered prior to our third meeting:

1. Calculate the mean and standard deviation for the chlorine concentration data.

2. What target concentration of chlorine would you choose to provide at least 1 mg/L of chlorine to 95% of the people?

3. In addition to providing bleach and measuring devices to the community, describe some of the other steps that you would need to take to assure the success of the project. In other words, after you buy the bleach and containers, what would you do next?

Household Water Treatment with Ultraviolet Light

Background

As an alternative to chlorine, pathogens can be inactivated by exposure to ultraviolet (UV) light. Ultraviolet light is classified as light with a wavelength between 10 and 400 nm. For the purposes of water treatment, we normally focus on UV light with germicidal activity, which has wavelengths between 200 and 300 nm. This treatment process inactivates pathogens because the genetic material (i.e., DNA and RNA) within the organisms is damaged by adsorption of the high-energy germicidal UV light. For water treatment, UV light often is produced with a low-pressure mercury lamp in which electricity is applied to mercury vapor to produce light in a narrow spectrum with maximum intensity at 254 nm.

UV light is used to inactivate pathogens in water, air and on surfaces. For many years, UV treatment systems were too expensive to use at large drinking water treatment plants but recent advances in lamp design have made these systems practical for use in treatment systems for cities. UV treatment also can be effective in household treatment systems. Several organizations have developed simple, inexpensive UV treatment systems for villages or households in less developed countries. Additional information on these systems can be found on the UV tube project website. These systems either rely upon the availability of electricity or include solar cells for producing their own electricity.

How do we determine how much light is needed?

As was the case with chlorine, the inactivation of pathogens by UV light depends on the overall dose that the organism receives. The dose of UV light depends upon the intensity of the light produced by the UV lamp and the amount of time that the water is exposed to the light. Therefore, the inactivation of pathogens by UV light can be expressed in a manner similar to the C*t expression used for chlorination.

[equation 3]

Where:

N = Concentration of active organisms remaining after treatment (number/L)

N₀ = Initial concentration of organisms (number/L)

L_UV = Coefficient of specific lethality for UV radiation (cm²/mW sec)

I = Light intensity in the water, mW/cm²

t = time (seconds)

Using this formula, it is possible to estimate the dose of UV light (i.e., I*t) needed to achieve a target concentration of pathogens in the water passing through the reactor.

The dose of UV light can be increased by increasing the intensity of the UV light or by increasing the amount of time that the water is exposed to the light. In most cases, the UV light intensity is fixed by the number of lamps in the system and we can only adjust the amount of time that the water is exposed to the light by varying the flow rate.

The average light intensity in the water will decrease if the water contains substances that absorb or scatter significant amounts of UV light. The most common substance in untreated water that absorbs UV light is dissolved natural organic matter (e.g., tea is an example of UV light-absorbing natural organic matter). Light scattering is usually attributable to suspended clay particles in the water. The effect of these substances on light intensity can be expressed with the following equation:

[equation 4]

Where:

I_d = Light intensity at depth d, mW/cm²

I₀ = Incident light intensity (i.e., light intensity at the top of the water column), mW/cm²

a = Diffuse attenuation coefficient, cm^-1

d = Depth of water, cm

The value of a will increase as the concentration of suspended particles or dissolved
UV-absorbing compounds increases. Typical values for a at 254 nm are 0.1 for particle-free groundwater and 0.5 for particle-containing river water.

Third class meeting: Planning the experiment

In this part of the module, you are going to investigate the use of a UV lamp system for household water treatment. The system that we will be using consists of a 15-watt UV lamp suspended over a stainless steel pipe (see Figure 2). The water flowing under the lamp is exposed to UV light. By controlling the flow rate of the untreated water into the pipe, you can set the dose of light received by the water.

Figure 2: Household UV treatment system.

The reactor system that you will be using is referred to as a plug-flow reactor (PFR). In this type of reactor, the water should move through the pipe as discrete parcels that do not mix with the water already in the pipe (see Figure 3). Given information on the flow rate of the water through the pipe (Q, cm³/sec) and the volume of water in the pipe (V, cm³) you should be able to predict

the amount of time that the water is exposed to UV light in the plug-flow reactor, which is also known as the hydraulic residence time (V/Q).

Figure 3: Top view illustrating the movement of a parcel of water (shaded region) through a plug flow reactor with a 10-second mean hydraulic residence time.

Questions to be answered prior to our fourth meeting:

1. Assuming that the average light intensity in the reactor is 30 mW/cm², estimate the hydraulic residence time needed to make your water safe for consumption in the home water treatment system. Ignore the possible effects of suspended particles or UV-absorbing substances on UV light intensity in the reactor.

2. How much will the presence of suspended solids increase the necessary mean hydraulic residence time for total suspended solids concentrations of 1, 10 and 100 mg/L? Estimate the mean hydraulic residence time needed to treat water at a depth of 5 cm using the following values for e: 0.05, 0.3 and 0.7 cm^-1 to correspond to suspended solids concentrations of 1, 10 and 100 mg/L.

This calculation is more complicated than it seems! Water at each depth will receive a different amount of UV light because the value of d in equation 4 is a function of the position of the water in the reactor. You may want to break up the water into slices (e.g., 1 cm thick as a function of depth) calculate the concentration of pathogens leaving each slice, calculate and concentration after the water mixes and solves for the residence time needed to achieve the desired level of treatment.

Fourth class meeting: tracer tests

Two additional factors will complicate the estimation of inactivation of pathogens in the UV treatment system. First, the assumption of ideal plug flow behavior may not be valid, especially when very high or very low flow rates are used. For example, at very high flow rates some of the water might find preferential flow paths through the reactor resulting in a distribution of residence times. This phenomenon is best illustrated by considering the results of a tracer test, in which an inert substance is added to a parcel of water in a brief pulse (i.e., 1 second) as shown for the shaded region in figure 3. Under ideal conditions, the pulse of the contaminant should emerge from the reactor over a 1-second period starting 10 seconds after it was introduced into the reactor (the solid line in Figure 4). However, if non-ideal flow occurs, some of the contaminants may emerge from the contaminant before and after the expected retention time (the dotted line in Figure 4).

Figure 4: Concentrations of a tracer measured at the outlet of the reactor depicted in Figure 3. The solid lines illustrate the expected shape of the profile for ideal plug flow conditions and an initial pulse of 1 second centered at time = 0. The dotted lines depict non-ideal conditions.

Deviations from ideal plug flow conditions are important because they can decrease the overall efficiency of the reactor. For example, under the non-ideal conditions depicted by the dotted line in Figure 4, some of the water spends much less than 10 seconds below the UV lamp. If the contaminants are traveling through the reactor at a rate faster than the average hydraulic residence time, it is possible that the inactivation of pathogens will be less than predicted from equation 3. The errors introduced by non-ideal flow conditions become particularly important when you design reactors to remove pathogens.

For example, if we were to design a system to remove 99.995% of the pathogens from water under the ideal plug-flow conditions depicted by the solid line in Figure 4, the non-ideal conditions depicted by the dotted line would lead to 99.990% removal overall. While this seems like a small difference, it would mean that your treated water would contain approximately twice as high a pathogen concentration as your design target.

As part of our laboratory experiment, we will perform tracer experiments to evaluate the mean hydraulic residence time as a function of flow rates and inlet configuration. By examining the shape of the curves of the tracer experiment, we also will assess the importance of non-ideal flow in the reactor.

During the tracer experiment you will add a pulse of water containing sodium chloride (i.e., table salt) to the UV reactor after setting the flow rate. By collecting data on the conductivity of the water leaving the reactor as a function of time, we can determine the mean hydraulic residence time. (Note: we have added a colored dye to the salt solution to help you visualize the tracer test.) From the width of the peak (i.e., the standard deviation) we can assess non-ideality.

So what do we do in class?

Measure the mean hydraulic residence time as a function of flow rate over flow rates ranging from 10 to 30 gallons per hour. We also will examine the effect of changing the inlet configuration on the residence time. Don’t forget to measure the total volume of the reactor. Your instructors will help you to set up the experiment and will make the data available to you as a spreadsheet file after class.

Questions to be answered prior to our fifth meeting:

1. Use the data from the tracer test to estimate the mean hydraulic residence time for each flow rate and inlet configuration. Is this relationship consistent with predictions based on the flow rate and reactor volume?

2. Evaluate the shape of the curves to assess the occurrence of non-ideal conditions. This can be done qualitatively by examining the general shape of the curves. To evaluate the data quantitatively you will need to quantify the width of the peak (e.g., how long does it take from the first appearance of the tracer to the HRT?). Are non-ideal conditions more important at high or low flow rates? What are the implications for use of the household UV treatment system?

3. How would you change the design of the existing UV reactor to improve the treatment efficiency?