Automated Alum Dosing

This aspect of the project involves the design of a component that helps automate the water treatment plant. A crucial constraint of previous designs was that they required an operator to manually vary the flow rate of alum entering the system. Heavy rains would sharply increase the turbidity of the source water and the operator would have to walk considerable distances to change the alum dosage to accommodate the turbidity. In addition to being inefficient, the delay allowed contaminated water to exit the plant because an insufficient amount of alum had been added. This meant that the manual system delivered ineffective dosages. The goal of the automated alum feed section of the project was to create a system that will sense the turbidity of the influent and automatically adjust the alum dose by changing the flow rate of the alum solution feed. Successful automatic alum dosing would allow for more effective disinfection of the water by chlorination before distribution.

Design Requirements

The design of this automated system needs to account for several considerations. The water treatment plant in Ojojona is designed to operate despite large variations in flow-rate. Even if the turbidity remains constant, the alum dosage requires adjustment based on the amount of water flowing through the system. The system should operate using very low power consumption, and this power source needed to be consistent, reliable, and sustainable. The entire system needs to be extremely cost-effective. In case of a system malfunction, the treatment plant needs to revert back to a manual setting. The automated alum feed should also be based on feedback from both influent and effluent sensors. And finally, the system needed to be simple and replicable in other rural communities.

The initial phase of the project involved determining the requirements of the automated system. As previous work involved intuitively determining the ideal alum dose based on the jar test, a more standardized method was necessary in order to create an algorithm. As influent through the system varies depending on several conditions in the environment, precipitation data for various periods of the year was obtained. The range of turbidity levels was also obtained in order to determine the ideal dosage. Besides these more critical numbers, information regarding temperature and humidity extremes were also obtained as any electrical components would need to function effectively within these ranges. We also investigated the environment around the proposed construction site at Ojojona, to determine the feasibility of providing power to our component. Based on this analysis, we determined that using solar panels along with a rechargeable battery would be a viable option. After considering these requirements we devised an initial system that would provide an automated alum feed. The limitations of this initial led us to make some modifications and to finally arrive at a prescribed system design.

Automated System Design

Selecting an appropriate turbidimeter was one of our crucial design steps. The turbidimeter needed to operate within a range of 0 - 600 NTU in order to provide useful readings for the environment in Ojojona. In addition to being that sensitive, the sensor needed to be hardy enough to function for a long period of time. Another important consideration was that the sensor needed to be easily integrated both with the other electrical components as well as the water treatment plant itself. We also found that the turbidimeters that operated within our specified temperature and turbidity ranges tended to consume a lot of power, so power reduction became another crucial consideration. And as is it is the most expensive component of the system we had to search for a relatively inexpensive one.

Company: HF Scientific

Model #: Micro TOL Online Turbidmeter with 24V option (part #20111)

This sensor fit all of our operating requirements and is available with a low power 24V (20-30 VA) option. Other turbidimeters we found operated at upto 80VA. Our project supervisor, Monroe Weber-Shirk has established a relationship with HFScientific as he has purchased several of their products for his lab at Cornell Univeristy. The HF Scientific turbidimeter has been tested within a lab setting and is known to work well for our situation. Furthermore, the influent fed into the turbidimeter is gravity driven and so there are no pumps that would use an excessive amount of power. Finally, the sensor is covered in a protective case which will ensure operation of the turbidimeter for a long period of time.

October 24, 2008

Accomplishments

• cleaned floc tank Sunday, Wednesday, Thursday and Halloween
• setup the tube flocculator
• manipulated the increment alum dosing in the operator selected state
• attempted to use peristaltic pump with tube flocculator
• replaced tapered baffles with uniform baffles to run increment alum tests
• worked on wiki page

Outcomes

• created turbidity profiles on Sun, Tues, Weds and Fri
• improved our method of creating turbidity profiles, by normalizing data and adding more data points
• replaced soft tubing in the tube floc with rigid tubing.
• Sunday's turbidity profile inaccurate because incrememnt alum dosing was not set
• also cleaning and sampling method varied each test so data cannot be compared
• two tests on uniform baffles will be run next week with increment alum dosing then the uniform baffles will be replaced by the tapered baffles

Issues

...