HAMMER Validation

Yusuf,

I have been modeling for nearly 30 years and doing site dataloggings of systems I have modeled.

I have used 7 softwares through the years, I am still using two with HAMMER being the most recent addition.

HAMMER is really becoming the dominant one we use, it is simply too versatile.

The MOC is the best modeling engine, that is my opinion, I tried the wave plan with one very well known product.

HAMMER coding makes good in using MOC, I met the original coders in 1995, I could not follow half of what they explained coding wise but they did love the site data I brought to them.

An example of why we trust the tool is this file we had in 2019 where an Engineer I had trained on a transient software 20 years ago  had just recently been hired by a city and called on us.

The city was about to commission a new valve chamber at the end of a gravitary profile.

When the Engineer in question asked the Consultant about his plans on valve closure for this over a half-century old 4.5 km long, 600mm pipeline, the answer why was:

"why worry there is only 20 psig of static pressure?".

That pipe is rated 100 psig and is the sole provider of water to this reservoir where the valve chamber is.

They used to take minutes to close the intake to the reservoir, now they have new actuated valves in place.

The city Engineer froze the commissioning and we got a call

We modeled the system with HAMMER at  the requested flow the city wanted to go in the reservoir with the valve's closure that was planned.

Looking at the modeling results we did not even want to test at thaf flow because of the risk involved.

When we did the testing on site, the automation guy made an error and did just that, same flow as our model, same valve closure as our model.

This is HAMMER prediction, screen picture of the laptop on the site:

This is the picture of the logger's screen that was right next to it:

The blue line is the equivalent head trace to the model, we were logging head at two locations hence the green line.

The pink one is the flow.

The Daughter and I were thrilled on site, people around us were either afraid or thought we were terribly nerdish...

We were lucky the pipe held, it had to be isolated from the reservoir and slowly flushed because turbidity went up following the test's transient event.

We have hundreds of files of all sorts where modeling predictions were compared with site loggings.

This file is really close model to site results, assembling model is hard work, when you see it's on the money on site, it's a rush each time.

We have a video of the valve's closure with logging traces and simultaneous view of the valve, sound and all (vena contracta at the end) on our private YouTube channel.

But it is in French, eventually we will do an English version and if it is of interest we will post the address.

As far as solving the problem shown here, modifications were made based on further modeling and today the system is safe, of course we retested it.

A note, that closure was started with the valve partially choked to limit incoming flow.  A further model predicted that with the way they were planning to close this valve from full opening to 100% closed, the resulting head at the valve would have been 194m./275 psig.  That would have made the nightly news broadcast...

HAMMER works up to a point that we are redoing major files done through the years on other softwares on it.

We are outputting to Epanet from other tools and importing in HAMMER with relative ease, you still have to go over all properties boxes but at least you don't have to redraw the grid.

And Jesse is right, the algorithms, the calcs matrix, all that stuff, is really complex, I gave up trying to code ages ago, I just started too old in this very specific field.

But aside from air valve models, that are tricky to manage, (all transitent softwares have problems with air valve modeling) if a model is well configured, you can trust HAMMER to give you a fair representation of a given system transient behavior.  We have air chamber models, acceleration-deceleration models, valve closure, surge relief, air valves; the deviation between predictions and reality is good.

I'm in my 9th year using HAMMER, and Caroline (the Daughter in question) and me like the tool, so many facets to it.

BENTLEY SYSTEMS technical support is simply the best I have seen for any software platform we have used and are using.

I hope this post is useful to you Yusuf.

Normand

Yusuf,

I have been modeling for nearly 30 years and doing site dataloggings of systems I have modeled.

I have used 7 softwares through the years, I am still using two with HAMMER being the most recent addition.

HAMMER is really becoming the dominant one we use, it is simply too versatile.

The MOC is the best modeling engine, that is my opinion, I tried the wave plan with one very well known product.

HAMMER coding makes good in using MOC, I met the original coders in 1995, I could not follow half of what they explained coding wise but they did love the site data I brought to them.

An example of why we trust the tool is this file we had in 2019 where an Engineer I had trained on a transient software 20 years ago  had just recently been hired by a city and called on us.

The city was about to commission a new valve chamber at the end of a gravitary profile.

When the Engineer in question asked the Consultant about his plans on valve closure for this over a half-century old 4.5 km long, 600mm pipeline, the answer why was:

"why worry there is only 20 psig of static pressure?".

That pipe is rated 100 psig and is the sole provider of water to this reservoir where the valve chamber is.

They used to take minutes to close the intake to the reservoir, now they have new actuated valves in place.

The city Engineer froze the commissioning and we got a call

We modeled the system with HAMMER at  the requested flow the city wanted to go in the reservoir with the valve's closure that was planned.

Looking at the modeling results we did not even want to test at thaf flow because of the risk involved.

When we did the testing on site, the automation guy made an error and did just that, same flow as our model, same valve closure as our model.

This is HAMMER prediction, screen picture of the laptop on the site:

This is the picture of the logger's screen that was right next to it:

The blue line is the equivalent head trace to the model, we were logging head at two locations hence the green line.

The pink one is the flow.

The Daughter and I were thrilled on site, people around us were either afraid or thought we were terribly nerdish...

We were lucky the pipe held, it had to be isolated from the reservoir and slowly flushed because turbidity went up following the test's transient event.

We have hundreds of files of all sorts where modeling predictions were compared with site loggings.

This file is really close model to site results, assembling model is hard work, when you see it's on the money on site, it's a rush each time.

We have a video of the valve's closure with logging traces and simultaneous view of the valve, sound and all (vena contracta at the end) on our private YouTube channel.

But it is in French, eventually we will do an English version and if it is of interest we will post the address.

As far as solving the problem shown here, modifications were made based on further modeling and today the system is safe, of course we retested it.

A note, that closure was started with the valve partially choked to limit incoming flow.  A further model predicted that with the way they were planning to close this valve from full opening to 100% closed, the resulting head at the valve would have been 194m./275 psig.  That would have made the nightly news broadcast...

HAMMER works up to a point that we are redoing major files done through the years on other softwares on it.

We are outputting to Epanet from other tools and importing in HAMMER with relative ease, you still have to go over all properties boxes but at least you don't have to redraw the grid.

And Jesse is right, the algorithms, the calcs matrix, all that stuff, is really complex, I gave up trying to code ages ago, I just started too old in this very specific field.

But aside from air valve models, that are tricky to manage, (all transitent softwares have problems with air valve modeling) if a model is well configured, you can trust HAMMER to give you a fair representation of a given system transient behavior.  We have air chamber models, acceleration-deceleration models, valve closure, surge relief, air valves; the deviation between predictions and reality is good.

I'm in my 9th year using HAMMER, and Caroline (the Daughter in question) and me like the tool, so many facets to it.

BENTLEY SYSTEMS technical support is simply the best I have seen for any software platform we have used and are using.

I hope this post is useful to you Yusuf.

Normand

Thank you for these! I would very much like to see more test vs. actual data if you have some. I have no doubt HAMMER is a very powerful tool; the validation is just something our site requires for any work. What version do you use? We have 10.03.01.08.

We have 10.03.02.75.  Really producing documents detailing model vs site results is a lot of work and requires that the client gives us authorisation that the site data can be released.  For the validation to be retracable you need to mention where the logging was done.  I had to do this a few times for other softwares and believe me it is demanding, time consuming and I even had to deal with questions from a Lawyer once!  But simply sending JPGs comparing traces without identifying the sites is doable but I wonder if that has its place on the community site?  If you tell me what kind of modeling vs site logging applications you're thinking of, I could search the database and send, eventually, some stuff.

Normand,

Thank you so much for sharing your positive experiences with HAMMER and our technical support team. We love getting user feedback on real-world performance!

Kristen Dietrich

Senior Product Manager, Water Infrastructure

Bentley Systems