Targeting the minimum flow rate for drinking-water distribution

Hello,

It’s difficult to give you the value you’re looking for, because it will vary according to the type of network you want to monitor. Is it an urban network, a rural one…? Are there any large industrial customers who may perform certain operations at night?

You can start by setting a Linear Water Losses Index of 150 L/h/km of mains. But once again, this value may not be suited to your network and will require adjustments.

Hello cams,

I agree with LR06 that it’s hard to give you a definitive value without knowing the network type or size.

In theory, a rural municipality (watch out for any agricultural operations) should be below 1 m3/h (I’m not putting 0 because it’s difficult to ascertain depending on the pulse value of the meter contact head).

You can use exported volumes from the customer file as a baseline.

Once you know your annual volume, divide it by 365 to get the average daily volume. Then multiply this number by the desired output and that will give you your daily volume objective.

After observing the volumes recorded on the district meter, you’ll be able to see whether consumption is linear or fluctuating (seasonal or activity-dependent).

You need to look at the theoretical daily volume calculated above in order to determine the volume losses per day and convert it to m3/h to find out what to aim for. Your objective of m3/h at night is therefore easy to calculate.

To give you a more tangible illustration: If your network does not drop below 5 m3/h at night and, in calculating your performance, you are looking for 3 m3/h, your objective is 2 m3/h at night.

To monitor your network’s linear consumption, the m3/day numbers are sufficient, but for a fluctuating network, you absolutely need your nightly m3/h.

Where it gets complicated is that a network of 15 m3/h at night can be a good performer while a network of 2 m3/h of nights might be lousy.

You need to spend some time observing the distribution curves to ensure your performance goals are in line with the network.

Hope this helps a bit.

I assume that your district meters for both wholesale imported and wholesale exported water are equipped with remote management. Your settings have to incorporate data collection intervals of 15 minutes max.

That way you can calculate your minimum flow daily on your sector. I am talking about the minimum flow—don’t use the night flow from 2 to 4 a.m., for example, because that won’t give you an accurate reading, particularly if you have certain types of customers like industrial plants. This min flow rate must be very close to 1 or a few m3/h according to your sector customer base. The best way to obtain your ideal level is to use the lowest value every time you identify it.

Seems to me at least that it’s the fastest and easiest way.

Hello,

Beginning with minimum flow rates, I calculate a Linear Water Losses Index (LWLI) for each district that I then compare to the data available from my local water authority—see below.

Where D represents the number of subscribers per network km and where the Linear Water Losses Index is m3/day/km.

Hello,

Already quite a few leads provided.

1 obvious point: be familiar with the network set-up and properly identify any meters managed via remote.

Next: retrieve the collected data. Fluksaqua offers a solution that is easy to set up and gives really interesting results

In order to set your thresholds, we have to go back to etlemondeserameilleur’s table above. In my case, I actually start with the Linear Consumption Index (LCI) to determine whether we are RURAL, INTERMEDIATE or URBAN (often it is pretty obvious). For the Linear Water Losses Index, I take the minimum flow rate, since I consider that it accounts for all losses (NB: this means losses both before and after the meter).

I calculate the difference between the daily volume and this minimum flow rate x 24 h. If the Linear Water Losses Index is 30, then URBAN—so INTERMEDIATE is anything in between the two.

Using the same logic, I calculate the Linear Water Losses Index and classify my metering districts as GOOD, ACCEPTABLE, POOR & UNACCEPTABLE. We focus our leak detection efforts on the UNACCEPTABLE districts, and I set alert thresholds on the GOOD and ACCEPTABLE districts so that I know when we move to the upper level. (You have to give yourself some wiggle room initially and establish thresholds that correspond to the POOR level or else all the districts will be in the red zone and prevent you from prioritizing. After dealing with the UNACCEPTABLE, then we address the POOR and ACCEPTABLE.

Then to ensure your operation meets the highest standards, you’ll want to review your efficiency action plan and examine other key elements such as the fleet of meters in the network, maintenance of air release valves, purging of water hammer arrestors, district meter metrology, pipe refurbishment schedule, etc.

NB: take into account any contractual or commercial “priorities”.

A+

Hello,

Thank you so much for your answers.

You read my mind about Fluksaqua—it’s precisely so that I can use their benchmark feature that I wondered about the choice of a daily reference volume and minimum flow target.

I have already identified the remote meters. For my distribution meters, I chose as daily reference volumes the consumed volumes estimated from 2015 annual data to which I applied the performance required by the contract.

For the minimum flow target, I had used the same method as etlemondeserameilleur. I am now going to crunch the numbers using fredblock’s method to determine whether my districts fall into the same categories according to the two different methods. I basically have rural and intermediate districts.

Thank you again.

There’s also a more empirical method proposed by Lambert at http://www.leakssuite.com/concepts/uarl-and-ili/.

With this method, you set Unavoidable Annual Real Losses (UARL) as the minimum target. The UARL is a theoretical reference value that predicts what real losses would be for a given system if its infrastructure were in good condition and the best technology could be successfully applied. As a result, it is also a key variable in the calculation of the Infrastructure Leakage Index (ILI), which effectively attests to how well the distribution network is being managed and maintained.  

The detailed info you need is in the link, but the basic formula is as follows:

UARL (liters/day) = (18 x Lm + 0.8 x Nc + 25 x Lc) x P

OR

UARL (gallons/day) = (5.41 x Lm + 0.15 x Nc + 7.5 x Lc) x P

Where

Lm = length of mains (km or mi.)

Nc = number of customer service connections

Lp = average length of customer service connection piping, property line to meter (m or ft.)

Lc = total length of customer service connection piping (km or mi.) Lc = Nc x Lp (km or mi.)  

P = average pressure (m or psi)

Hello cams,

Agree with Ouananiche. AWWA uses the same formula as those referenced by Lambert.

Good answers and good discussion by all. Here are a few other things to consider when refining the District Metered Area (DMA), and establishing target levels.

Once the DMA is initially implemented, an acoustic leak detection survey should be conducted to identify existing leaks and have them repaired. The goal is to remove the backlog of existing leakage, and bring the flow into the DMA down to a baseline low level that reflects the supply only needed to meet the customer demand. An initial target level can be set using some of the approaches already discussed, but remember that each network and each DMA is unique, especially regarding cost. The cost of the water being lost to leakage (usually the variable production cost to supply water) should be calculated. Then the cost to deploy leak detection crews should also be known. The actual target level should be the leakage rate at which the cost of the water being lost to leakage will equals the cost to conduct the leak detection work. This is different for each system.

As previously noted the Unavoidable Annual Real Loss (UARL) can be calculated, but setting the target level as the UARL may not be cost-effective unless the production cost of the water is very high, or water resources for the system are very limited. The UARL is the lowest level of leakage that could be achieved in a DMA or network, but it can be very costly to attempt to bring leakage down to this low level, and the water savings achieved often won’t be sufficient to justify going to this low level of leakage.

I hope that this gives some additional perspective to the use of DMA’s and flow monitoring – this is a very useful technique which can help most water systems better control leakage.