These papers should help point you in the right direction Vahid

(PDF) Online data processing for proactive UK water distribution network operation (researchgate.net)

https://www.researchgate.net/publication/269124410_Optimal_Locations_of_Pressure_Meters_for_Burst_Detection

John

Using your hydraulic model to identify low and high pressure points in your zone of interest is a good first step.

Hello Vahid Nouri ,

I was under the impression that valves in a water distribution system were primarily for isolating parts of the system for preventative maintenance, construction, repair, isolating in the event of a pipe failure, etc. Can you provide any real world examples of municipal water distribution systems that utilize active valves? Here, I am distinguishing passive valves as those directly under control of people, as opposed to active valves, which are under computer control.

Regards,

Tom Cuff

Hi Tom,

I want to schedule the valves to provide suitable pressure for all parts of a zone. But I think if I tune my schedule based on some critical points which endanger with pressure drop, I can manage the rest of zone easily. Do you believe this idea is a practical idea?

Or is it better to follow different optimisation methods to do this schedule?

Thanks

Dear John,

Thank you very much for your helpful articles especially for " Online data processing for proactive UK water distribution network operation ". This article is very helpful. Do you have any updated article about the UK water distribution network operation ?

Again thanks for your consideration.

Hello Vahid Nouri ,

The only thing that worries me about your idea is what happens in the case of a failure? For example, how will your system react to an electrical, mechanical, or hacking failure? In other words, will your sysem react to any and all failure modalities in a graceful manner?

In most modern gravity-fed water distribution systems, pressures are usually controlled by spring operated regulators. With active valves, as you propose, if your sysetm loses electrical power does the state of the valves allow the system to work as a passive system, i.e., return to its default mode? Active valves add one more thing that may malfunction mechanically. How do you assure that a mechanical valve failure leaves the valve in a safe state? What about deliberate hacking? An active system, such as yours, would seem to be a prime target for such disruptions.

Regards,

Tom Cuff

Dear Tom,

Thank you for your viewpoint. Yes, it is very important. In this idea a Data Validation and Reconciliation(DVR) system, would validate the input data for predicting the behaviour of system. So, if a failure in electrical or mechanical occur, the DVR system can detect if the data is valid or invalid. This can help us to control the valves when something abnormal happen in the system.

About hacking the system, we need to apply some security protocols on the system to protect it. But for mechanical failure at the moment I do not have any idea and again thanks you for consider this issue to me.

Please let me know if you have any further idea about this system.

Thank you very much for your support and time.

Best wishes,

Vahid

Hi Vahid Nouri ,

Let me give you a single instance where a mechanical issue with a valve contributed to a major disaster: March 28, 1979, Three Mile Island, Unit 2, Nuclear Power Plant. The Electromatic Relief Valve (a.k.a., Power Relief Valve or Electromatic Operated Valve) located at the top of the Pressurizer in the Primary Cooling Loop, experienced a mechanical failure, i.e., it did not close when instructed to do so, thus allowing coolant to leave the primary cooling loop. This valve was electrically operated by a solenoid. It could be controlled either automatically (by a control system) or manually (by an operator in the control room). Unfortunately, the annunciator for this valve did not actually read the state of the valve, but rather the annunciator indicated whether or not current had been applied to the valve's solenoid, from which the state of the valve was inferred [1]. Note, although there was no direct sensor to sense the position of the valve disc directly, there was a water level gauge connected to the pressurizer [2]. Unfortunately, in the situation where the residual heat of the scrammed reactor was not being removed because of loss of coolant, this gauge could and did give an erroneous water level (too high) in the pressurizer causing the operators to turn off the emergency cooling system lest a condition known a "solid pressurizer" occur, i.e. no steam bubble at the top of the pressurizer. The erroneous reading was caused by super heated steam mixing with the coolant.

[1] Ellis Rubinstein & John F. Mason; An analysis of Three Mile Island; IEEE Spectrum; Vol. 16; No. 11; November 1979; pp. 32-42 + detailed foldout. [Detailed drawing of the Electromatic valve + piping of the system.]

[2] Mike Gray & Ira Rosen; The Warning, Accident at Three Mile Island; Contemporary Books, Inc.; 1982; 287 pp. [See the diagram of the water level gauge connected to the pressurizer on p. 38.]

[3] Victor Bignell & Joyce Fortune; Understanding Systems Failures; Manchester University Press; 1984; 216 pp., see in particular pp. 9-39, Chapter 2. [Note, see the detailed pictures of the massive control room.]

[4] James R. Chiles; Inviting Disaster, Lessons from the Edge of Technology; HarperBusiness; 2001; 338 pp., see in particular pp. 37-63, Chapter 2.

Regards,

Tom Cuff

Dear Tom,

Thank you for your interesting example and the challenge that you have mentioned. We have about 20 data loggers in the zone, so we can understand if one of the valves does not work correctly because the pressure behind the valve would be increased while the pressure in other data loggers would not changed!

So system would understand a failure about the valve. In this case we can use a couple of parallel valves in the same place to increase the reliability of system. The second valve would be actioned if the first of valve does not work correctly.

I think it could handle the issue. Do you agree?

Regards,

Vahid

Hi Vahid Nouri ,

Consider the following failure mode. A valve, which is supposed to be 75% closed (25% opened) fails in the fully open position (100% opened). In this case, a redundant valve in parallel will be useless because the redundant valve cannot reduce the flow, only a redundant valve in series with the failed valve would be able to reduce the flow.

Each valve would have to have both a parallel and series redundant valve to be able to deal with all valve failure modalities.

Regards,

Tom

I have several questions related to my question. In all of British papers about WDN, DG2 refers as risk point for low pressure point in the WDN.

1) What is the key word in non-British papers?

Based on different papers burst or leakage detection are the common ways for finding DG2 point.

2) Do you know any other ways to find DG2 points in a WDN?

3) Is there any standard or simple method to find DG2 points in a WDN?

Thanks in advanced

Vahid