Dear Gents,

I want to talk with you about the possibility of continous on-line monitoring actuation of valves by mean of ultrasound technology. I have thought in DS-386 (from UE System or similar) and monitoring sound with permanent probe mounting on down side of the valve, or perhaps monitoring the difference between B and C traditional checking method (one probe before, one probe after the valve). If you monitor the valve Db readings vs time, you do not need the four checking points A, B, C, D and sometimes E Method (from UE System too), no?

Have you got any experience in this kind of apliccations? What about the possibility to integrate suchs detection method in manufacturing process of valve or as a added value to the valve itself?

Please, tell me your suggestions about this issue or your own experience.

Regards,

Allen

You have a lot of possibilities with the UE386. However what I would need to know is what do you want the system to do? Detect a valve leaking when it is closed? What is your application? The 386 systems can be manually adjusted to a set alarm level, mounting 4 possibly 5 transducers this becomes expensive, but then again what is the application. The two points upstream of the valve would and should detect the static pressure of the energy before the valve, which in the closed position these sensors should read close to the same. While sensors 3 and 4 as well should read the same, remember 3 is placed at the downstream side of the valve any movement of the valve opening will create quite a delta P (differential) and trigger a false positive, but that is why the UE 386 should not register an alarm until it has heard the sound again or for several seconds, 4 will hear it as well but at less sound, while sensor 5 is merely to monitor downstream further to register any sound created from a leak or another anomaly further downstream and acts as a warning system for another problem. Again depending upon your particular need and application whether you can do this without three of the sensors. You may only need to monitor ONE position that is aft of the valve you can set the sensor only to respond after it has heard the elevated decibel range for several seconds after the valve has been fully stroked opened. The idea of valve leak detection (if this is your application) is to hear the valve leaking past its seats...as it is the 386 system is about TIME, real time as to when the sound is detected and if the sound is heard over a set period of time then an alarm is latched.
Good Luck.

This message has been edited. Last edited by: Jim,

Thanks dear Jim,

Not any application concrete, but only asking for general information. We have some leak failures and we intend to detect it earlier, in order to prevent dangerous consequences.

Could you explain again the method to diagnose valves with only one probe after the valve? It seems cheaper. Most of our valves ares on/off, but in case of modulating valves, is it valid that method you talk about?

Regards Jim,

Allen

Allen,

When a valve is closed and not leaking, then the ultrasound level is theoretically zero. In practice, the ultrasound level depends on the background sound in the piping system. An upstream measurement serves as a Reference for the piping system sound. If downstream measurement is equal or less than Reference upstream, then either the valve is not leaking or the sound from leaking is masked by the background. If the downstream sound is higher than the Reference upstream, then the valve is either leaking, not fully closed, or downstream system background sound has increased.

A single sensor mounted on the downstream pipe or valve body close to downstream side of the trim/seat can only be effective if the system background is equal or lower than the sound from a leak.

Measurement of ultrasound with valve open might be used to detect abnormal flow rate (clog or closed upstream valve) if calibrated against valve position. Pressure gages may serve same purpose.

I could not find DS-386 sensor info, but I did find the UCA 586 on UE Systems Web site. A critical specification is a maximum temperature of 60-C (140-F), so this temperature could limit potential valve applications.

I would be very careful about what to expect from ultrasound valve monitoring. I suggest you try one or more ultrasound meters and make many hand-held measurements in both controlled (test lab) and actual plant conditions before you try to specify/design a permanent monitor system. Be wary of sales person's claims.

Walt
w_f_strong [at] msn [dot] com

The UE386 is in fact zero'd out before beginning with the ambient or normal background. I have seen and sold this unit for this application and to my knowledge saw very little problem. With most of the UE specific units it is best to go the extra mile and expense if you have a lot of valves and have NOT the sales person but one of UE engineers to come on site for an evaluation. Also, as Walt states, try using the handheld unit for this application first and prove the application. Good luck....
UE386 info: http://uesystems.com/prod_custom.asp

Adjusting the sensor output to zero does not eliminate or cancel the background sound. My comments regarding the influence of the background remain the same. The DS-386 multichannel monitor would be connected to contact-type ultrasound sensors, so the temperature limit must be considered for the valve service conditions.

Walt

Whatttttt, come on go to bed.
I said that the UE 386 is zero'd out. You turn the unit on their is ambient or background noise at this time, hopefully this is the normal for the valve. You then take in consideration the background and set an alarm limit above the normal background, you then set the time that it looks at a rise in decibels before it latches an alarm. Yes, temperature is a factor that has to be considered before using this system. GOODNIGHT...
The horse is glue...

Speaking of temperature - you can simply install a RTD mounted in the valve cover plate or within a valve cover plate bolt. Interface the RTD to a temperature transmitter and send the 4-20 signal to an existing control system or low cost PLC. When a valve begins to fail, it usually begins to leak the process gas. This causes the process gas to be re-compressed, further heating the gas. Measured temperature differences of 4 to 20 degrees F when compared to average temp of all like valves can indicate a problem valve. Would permanantly installed ultrasonic sensors and an ultrasonic monitoring system be more cost effective than this?

We follow working in this project. So, I would like your opinion about this possible solution :

- Db measurement upstream and downstream of the valve with magnetic / threaded probe. Permanent probe.
- System of measurement : continous device with permanent connection to probes.
- Alarm levels : Defined in first measurements according to residual noise in system. Upstream alarm level detecting any change in upstream turbulence or background noise, and downstream alarm level detecting any change in downstream condition. If there were an alarm condition in both sides, both increasing upstream background noise and leaking condition of valve would be possible. In other cases, you always could distinguish between leaking and background noise.

It would be a valid system if you trend Db levels.

Regards,

Allen

Allen,

I suggest two leak detection alarms:

1) Downstream dB level increase
2) Delta dB increase (downstream - upstream)

If alarm-1 only, then increase in system noise, but probably no leak, although detection threshold increase may prevent detecting small leaks.

If alarm-1 and alarm-2, then leak is probable.

If only one alarm channel per valve is available, then use alarm-2.

Can you state whether the valves are used for gas or liquid service? Can you state the maximum expected temperature where sensors are to be located? Will the sensors be in a harsh environment such as dust, chemicals, radiation, etc.?

Walt

Excellent idea Walt. Sure it could be definitively good. But, why do you ask me if liquid / gas service. The Db level is proportional to turbulence ratio, and this last parameter is proportional to sizing and sharping hole, to differential pressure, and viscosity of the fluid. I suppose you refer to this last factor, but I suppose this could influence in alarm level or %change from the baseline established as our level of alarm.

Regards,

Allen

Allen,

I am just trying to understand your application. Compressible gas or incompressible liquid with possibility of Cavitation can conceivably affect background ultrasound level and leak detection threshold. Have you conducted leak detection tests, or are you installing the monitor and hoping that it works? I don't consider myself to be an expert on this application, but I have done enough tests to indicate that this is "not a piece of cake" especially on large valves with steam service.

Walt

We are hoping that monitor works, but no tests made yet. I agree with you in opinion that this is not a piece of cake. We are trying to find some optimal solution for monitoring valves. Perhaps, ultrasound is not the best one, neither technical not ecconomical point of view.

Allen