This is a technical guide to sizing rainwater pre-filters for your local rainfall patterns and roof size.

We have had a lot of questions about how to properly size PURAIN filters (or any type of self cleaning filters for that matter).  The questions stem from the difference in goals for handling stormwater runoff during extreme rainfall events versus optimal rainwater capture sizing.  Sizing rainwater pre-filters for extreme rainfall events may lead to excess capital cost and suboptimal performance.  Sizing pre-filters too small may cause low capture efficiency and possible water handling issues.

The considerations in sizing a PURAIN filter are:

1.  We want to make sure we catch the maximum amount of rain.

2.  We want the hydraulic jump to occur often enough to maximize performance and minimize maintenance.

3.  We want to avoid oversizing which can lead to extra capital cost.

4.  We want to avoid problems with undersized filters which include filter overflow other cases where and stormwater goes where is not supposed to go.

In addition, we have heard a lot of feedback about the use of standard roof sizes as guidelines for sizing pre-filters.  Perhaps this works well if you filter is being installed in Germany or someplace with similar rainfall patterns, like say the Great Lakes region of the US and Canada.  But, this may be an issue in the western US, Texas, or in Miami with radically different rainfall patterns and intensities.

This guide is intended to give you a hopefully easy to use method for sizing your PURAIN filter in any location.  It is based on local rainfall in your area so it will not use a blanket assumption for roof size which is only accurate for certain geographies.  You will have different size recommendations depending on the rainfall intensity in your area.

Background – Ideal Operation:

It should be possible with proper sizing to achieve optimal performance of you PURAIN filter.  To us, ideal operation is when you achieve water 100% capture for most rainfall rates with occasional hydraulic jumps to flush any accumulated debris out of the filter.   The amount of debris accumulation should not be enough to affect filter efficiency.  We think that the frequency of hydraulic jump (and filter flushing) should be a few times a year depending on debris load and rainfall pattern.


The reason that the PURAIN catches all rainfall up to a certain rainfall GPM is due to its design with a ramp up to the reject port.  Up to the hydraulic jump flow rate, the filter acts like a simple basket filter will all captured wateer falliing through the screen to the rainwater tank.  When the hydraulic jump occurs, water rolls back on itself like in a river rapids to clean the screen.  The amount of water flow required to achieve hydraulic jump is shown on this graph.

There will be some water flushed out of the filter when the jump occurs.  Luckily the amount of rainfall lost is small if the PURAIN filter is sized right.  In almost all geographies the distribution of rainfall looks something shown on this graph where even though there are some rain events with heavy rainfall, most of it falls at a lower flow rate:

The trick is figuring out what size filter achieves this kind of results for your geography and for roof size.  That is what we show next.


In sizing pipes for runoff from roofs, engineers need to look at how the design will handle infrequent, high intensity rainfall events.  In designing a rainwater collection system, we want to do a great job in handling lower intensity events such that water capture is maximized and water quality is the best possible using rainwater collection best practices (see ASPE 63).

If a rainwater prefilter is sized to handle water for all rainwater including for example 100 year rainfall intensity, then the filter will be very large and most of the time it will not perform well.  All types of self cleaning filters for the PURAIN jump filter to other cascade filters to the spinnning vortex style will not perform at their best if oversized.  Excess capital may be spent on installing bigger and more filters than is necessary and water capture and self cleaning will not be optimum.

One the other hand, if the system is sized for optimal rainwater harvesting, then heavy intensity rainfalls may leads to flooding and possibly building and equipment damage.

We must design systems that handle water well in extreme rainfall events and maximize rainwater harvesting performance.  Below is a description how this can be achieved.

Design Criteria:

In order to have an approach to prefilter sizing that works for everyone, we want to accomplish the following:

1.  We want a simple method that uses local rainfall data that is easy to find.

2.  We want to choose the design rainfall event such that the resulting filter size meets the criteria of catching almost all rain while achieving hydraulic jump frequent enough to keep the screen clean.

3.  Avoid any water handling issues that may occur during infrequent, high intensity rainfall events.


Note:  If you just want to skip to the easy to use cheat sheet, click here for a color coded PURAIN sizing chart.  Easy-Peasy…

1.  The first step is to decide know the maximum flow a prefilter can handle hydraulically.   This is the amount of water that can flow into the filter without it overflowing.  In the case of PURAIN filters, we use the method laid out by DIN 1986 to characterize maximum flow.  Using this method builds in some conservatism since DIN 1986 assumes the flow at which the flow to the filter fills the incoming pipe to 70% with a 1.5% fall.  In reality, the PURAIN filter could handle somewhat more flow than this without overflowing.  Here are the flow ratings for each size PURAIN filter:

2.  The next step is to decide upon a rainfall intensity and frequency (return) and match the flows above to that intensity.  We know that we do not want to use the same 10 year, 25, year, or even 100 year intensity values with longer intervals that engineers use to size pipe for stormwater management.  It must be a higher frequency and thus lower flow event.   We have looked and tested a wide range of intervals from as low as 2 months (tough data to find!) and as long as 5 years.  We feel that using the 5 minute intensity on a 1 year interval is a pretty good choice for sizing and is readily available through NOAA.

This data is readily available on this NOAA website —  One simply selects the location of interest and reads the 5 minute, 1 year intensity value.  For most staes of the US, this is easy to get right off the website:

For some states, it is necessary to use a PDF report from the same website.  For example, you need to use the PDF to get data for Massachusetts.  In this case, it is necessary to use the data in this report for 5 minute, 2 year interval and multiply by a conversion factor of 0.84 to approximate a 1 year return.  For example, the 5 minute, 2 year value for Boston is 0.34″ so the 5 minute 1 year value is 0.28″.  Here is the link to that report:

3.  Now take the 5 minute rainfall amount and calculate the gallons per minute flow for your roof size.  Here is an example:

0.4″ 5 minute rainfall in Atlanta.  On a 5,000 sq. ft. roof this equates to 0.4 inches x 12 x 5,000 Square Feet x 0.623 Gallons per square foot per hour / 60 minute per hour = 249 gpm

Since 249 gpm is less than the maximum for the 6″ filter, we recommmend using the PURAIN DN150.  Would you be OK using an 8″ filter?  You probably would be OK with the larger filter since the flow on the 1 year event is near the 6″ maximum flow of 268 GPM.  Choosing the larger filter would lead to more rain capture but hydraulic jump would occur less frequently.  We estimate a jump every 2 months in Atlanta with a 6″ filter and about every 4 months with an 8″ filter, both of which are in range for reliable operation.  This means that there is some leeway.  Given the unpredictability of the weather, this is totally OK.

You can find the filter recommendations for 48 cities across the US and for roof sizes from 1500 square feet to 110,000 square feet by clicking here.

The table shows the distinct differences in filter sizing recommendations for various parts of the US.  In some places (like San Francisco and LA) a 4″ filter will work well on a 4,000 square foot roof where in others (like Tampa and Miami) an 8″ filter is recommended. It appears that areas like Detroit or Boston may closely be able to follow the standard roof size guidelines we see with self cleaning pre-filters.  It may be because these areas have similar rainfall intensities to the areas of Germany from where the roof size guideline came.

Gravity Bypass Lines:

Many of you are now asking what we should do to take care of situations in which the maximum flow of the filter is exceeded, which will happen sooner or later.  On average, this will happen every few years so this often must be accounted for.  In some cases, it is OK to let water overflow pipes and filters since water will flow to other stormwater management devices and BMP’s.  In other cases, we must be much more careful.  In those cases, which happen fairly often, we have a solution.  Our approach is to design the rainwater system for the “normal” rainfalls and then also design bypass lines so that when maximum pre-filter flow is exceeded, water is directed to a place where it is directed to a place where problems are minimized.  There are many ways to acheive this and your particular project site will require you to be creative in finding the best solution.  here we give you a few concepts that may spur those creative solution for your particular siuation:

1.  If the main concern is to prevent the pipes upstream of the rainwater system from overflowing, then overflow pipes can be placed upstream of the pre-filter to handle excess flows.  You can use your trusty pipe models to estimate maximum flows in pipes and then size the conveyance pipe how you normally would without any rainwater system.  For a direct burial PURAIN HD filter, a way to deign for ideal PURAIN flow and required stormawater management piping might look like this:

You may say have a 25,000 sqaure foot roof that requires a 20″ diameter drainage pipe for stormwater management, but our receommendations for PURAIN filter sizing is for a 12″ or 16″ DN 300 or 400.  Under the most frequent rainfall events, all the collected rainwater goes to the filter.  When the capacity of the filter is exceeded, the the 20″ bypass pipe takes the excess flow and directs water to the next stormwater BMP as shown on the diagram.

2.  We have heard about a lot of cases where self cleaning pre-filters were located indoors in a utility room or basement and has problems with water splashing and in many cases with more concerning flooding and equipment damage.  Thankfully, none of these filters were PURAIN filters.  What we do to assure these problems never happen is:

  • Rely upon the PURAIN design which naturally handles water better and avoids water splashing out, especially compared to those filters that create a swirling action.
  • When splashing is a worry, we fit a completely water tight cover on the filters.
  • We work with the engineering team to design a bypass line that handles excess flow.  It usually looks something like this:

This solution can be designed with a wet pipe situation or direct gravity feed.  The overflow pipe is situated such that rainwater flows to the filter under typical rainfall conditions.  When an unusually heavy downpour occurs, water rises to the overflow pipe level which we recommend to place such that the invert is just above the top of the filter feed pipe.  At this position, there will be a slight pressure on the filter and it would overflow slightly if left uncovered.  But, with the water tight cover, this pressure will be handled and there will be no leakage.  In this way there should be no leakage or mess and stormwater will be handled well in extreme rain events.

Final Comments:

This sizing guide is intended to show how generally how the best filter size can be selected.  It is possible to choose a larger or smaller filter depending on specific needs of a specific project, but there will be trade offs in capture efficiency and self cleaning action through the PURAIN hydraulic jump action.  As always we encourage your feedback on this design proposal and hope that you have found this information useful.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>