The effects of back-pressures on safety valves is a potentially serious problem, but one only now being recognised by the valve industry. Research has been performed by the University of Milan, as reported in the June 2002 issue of Valve World. This report is also printed in full below. In addition, please note that RWTH's Hans-Dieter Werker has presented a paper entitled "Influence of back pressure on function and flow rate of a direct loaded safety valve" at the Valve World 2002 Conference in Maastricht.
Back-Pressure effects on safety valves operating with compressible flow.
Laboratorio di Fluidodinamica delle Macchine, Dipartimento di Energetica, Politecnico di Milano
Superimposed or built-up back pressure strongly affects the operational characteristics and flow capacity of safety valves. It is common knowledge that this feature is connected to a reduction in the disc lift and/or to the establishment of a subsonic flow regime.
Laboratory tests on five different commercial safety valves, especially ordered for the purpose of operating under back-pressure conditions, show a difference between the performance guaranteed by the manufacturer and the actual valve performance. This difference may be so great that the protected equipment might operate over the maximum allowable pressure.
Laboratory tests under back-pressure conditions should be required for the certification of operational reliability and consequently plant safety; the characterisation of valve behaviour under back-pressure conditions is also required by the new prEN ISO 4126-1 standard which, at the time of writing in draft version, has been submitted to CEN members for enquiry.
Five years ago, a new test rig, for the evaluation of the performance of safety valves operated under back-pressure conditions, was designed and built at the Laboratorio di Fluidodinamica delle Macchine of the Politecnico di Milano University (LFM). In recent years many safety valves, manufactured in accordance with different European and International Standards, have been tested on this plant. The actual and the certified behaviour of the device when operating under atmospheric back-pressure conditions were close; on the other hand there was a considerable difference when back pressure was increased over the atmospheric value, particularly when the back-pressure ratio (ratio between the absolute back pressure and the relieving pressure) was increased over 25 %. The application of balancing bellows can help but not solve the problem.
The experience gained in working both as a "Research and Development" and "Certification" Laboratory led us to the decision of testing the behaviour of five different commercial valves under atmospheric and over-atmospheric back-pressure conditions; the laboratory results, in terms of operational characteristics and discharge coefficient, have been compared to the values declared by the manufacturer. The purpose of this paper is to highlight the requirement for operational and flow tests under back-pressure conditions, when certifying the performances of a single valve or of a valve size range.
The valves tested were ordered with the same technical specification, including the requirement to operate under back pressure. The orifice designation is "J", with 2" ASME 300 x 3" ASME 150 flanged inlet x outlet, and centre to face dimensions in accordance with API STD 526, June '95. Four out of the five valves (that we shall identify as A, B, C, E) are certified according to ASME VIII. We will call the fifth valve "D" and its manufacturer declares that the device can be applied at a maximum back-pressure ratio of 25%; however we have subjected the sample to the same tests.
Main technical specification for the five samples are reported in Table 1.
|Superimposed back pressure (min)
|Superimposed back pressure (max)
|Built-Up back pressure (max)
Table 1: Technical specification required for supplying valves A, B, C, D, E.
Test plant and test procedure
The test plant is equipped with an air storage of 5000 kg at 200 bar; the maximum flow rate is 8 kg/s, whilst the maximum operating pressure is 40 bar. The opening and re-seating transients, as well as the valve and flow characteristics at steady flow conditions, are described by means of high precision pressure transducers at a frequency rate of 5 kHz. The flow rate is measured by means of calibrated converging nozzles positioned at valve inlet, which allows a real-time reading of the flow rate through the valve. The application of the "Propagation Error Theory", has shown a flow rate uncertainty below 1.5% of the reading, with a confidence level of 95%. For the aim of this research, the valves were installed on a capacity of 1.2m3, the measuring nozzle had an orifice of 50 mm and the outlet of the valve was connected to a tank (1500 l) where the back-pressure is regulated as reported in the following.
Without removing the manufacturer seals and without any regulation of the spring or of any other part of the device, each valve has undergone the following tests:
||Set pressure measurement : the pressure in the supply tank is slowly increased: The set pressure value is detected through hearing.
||Discharge coefficient measurement: the discharge coefficient is measured at 10% overpressure of the set pressure stated at point 1.1
||Re-seating pressure measurement: the re-seating pressure is evaluated by slowly decreasing the supply tank pressure until the disc returns to the initial position.
||Built-up back pressure: the range of stable operation under built-up back-pressure condition is investigated. The valve is operated starting from the same condition as at point 1.2; subsequently, the back pressure is slowly increased by shuttering the exhaust of the back-pressure tank, whilst the relieving pressure is kept constant. The upper limit of the operational built-up back-pressure range is detected by the appearance of instability in the valve lift. The trend of valve lift and discharge coefficient as functions of the back-pressure ratio are recorded by the data acquisition system.
||Superimposed back pressure : the back-pressure tank is operated at constant pressure. Under this condition tests described at point 1.1, 1.2 and 1.3 are repeated at almost three different back-pressure values within the range specified in Table 1.
Results of tests at atmospheric back pressure
Table 2 reports the main results of the tests performed under atmospheric back-pressure condition for the four valves; K is the actual discharge coefficient, whilst Knom is the one declared by the manufacturer at atmospheric back-pressure regime. The most remarkable feature is the particularly high blowdown value of valve A, which is nearly 5 times that declared by the manufacturer. The rest of the data are reasonably in agreement with what was declared.
||Results of atmospheric back-pressure tests
|Blow down %
At 10% over pressure.
Table 2: Summary of atmospheric back-pressure test results
Results of built-up back-pressure tests
Figure 1 reports the results of the built-up back-pressure tests performed on the five valves; the plots show a reduction of both the manufacturer declared (dashed line) and actually measured (solid line) discharge coefficient versus the back-pressure ratio. The ratio K/Knom, represents the ratio between the actual measured discharge coefficient at a certain back-pressure ratio and the discharge coefficient as declared by the manufacturer for the same valve operating under atmospheric back-pressure condition.
Figure 1 : Reduction of the discharge coefficient referred to the one declared at atmospheric back pressure: experimental results and manufacturer declared values.
Results show that the expected de-rating of the flow rate of valves B and C and E is strongly under estimated. Moreover a very surprising result is found for valve E: in particular the valve shows a dramatic flow capacity decrease when back-pressure reaches the 18%. No further tests were performed on this device, because when the valve was operated during the first test at super-imposed back pressure, the bellows broke. In the author's opinion the bellows was operating correctly when the built-up back-pressure test was performed, but this hypothesis cannot be confirmed. However this accident underlines the importance of the reliability of the bellows, and supports the requirement highlighted in , paragraph 5.1.8: " In the case of failure of a balanced bellows, if any, the safety valve shall discharge its certified capacity at no more than 1.1. times the maximum allowable pressure of the equipment being protected".
Referring to valve A, the de-rating trend is clearly foreseen by the manufacturer and the flow capacity reduction is quite low. Note that at a back-pressure ratio of 68% the discharge coefficient has undergone a reduction of only 20%; on the other hand the valve shows great difficulty in the re-seating process, as already shown by the high value of blowdown measured in the atmospheric back-pressure tests. This will also be noticed in the superimposed back-pressure tests.
The manufacturer of valve B declares that the discharge coefficient is not sensitive to the back-pressure ratio until it reaches 50%, where a 3% de-rating of K is expected. On the contrary, experimental results show that the discharge coefficient experiences a progressive linear reduction starting from a back-pressure ratio of 20 % up to the 48%, where the discharge coefficient rate is about 2/3 of that of the atmospheric back pressure. Moreover at the same built-up back-pressure ratio, the valve shows clear instability and the disc tends to re seat.
As for valve C, the manufacturer declares that the built-up back pressure does not affect the discharge capacity; the plot clearly shows that this feature is not representative of the actual valve behaviour: the measurements show a dramatic reduction of the flow capacity, already at back-pressure ratios of less then 30% (K/Knom = 0.63); raising the back pressure the discharge coefficient is reduced to a minimum value of 43% of the nominal one.
Also the manufacturer of valve D states that no reduction in the discharge coefficient is expected for back-pressure ratio within 25% while measurements show a reduction of K/Knom to 0.95; the flow capacity reduces to 80 % of the nominal rate for a back-pressure ratio of 34%, beyond which the disc rapidly re-seats.
Results of superimposed back-pressure tests
Tests under superimposed back-pressure aim to verify the functional and flow capacity characteristics of safety valves within the limiting overpressure of 10%, evaluated on the basis of the set pressure measured in the atmospheric back-pressure tests.
Valve A has been tested at three different values of back pressure up to a back-pressure ratio of 65%. All tests show the same features: the valve pops before the 10% overpressure and the discharge coefficient reaches the same value measured in the built-up back-pressure tests for the same back-pressure ratio. As already mentioned, this clearly predicted behaviour is followed by quite a high blowdown; in particular the blowdown is reduced with respect to the atmospheric tests (37% in tab. 1) and keeps reducing with the increasing back pressure: moving (increasing) from a back-pressure ratio of 40% to 65% the blowdown reduces from 16.5% to 12%, in any case exceeding the limit of 7% declared by the manufacturer.
Valve B experiences the pop-up before the 10% overpressure is reached only for back-pressure ratio of less than 33%; for higher back-pressure ratios the pop up requires overpressures rising up to 13% at 37% back pressure, which corresponds to the limit indicated in the technical specification. Once the valve has opened, the discharge coefficient reaches the same values measured in the built-up back-pressure tests; clearly when the sample does not open, the flow capacity is very low. The blowdown is always within the declared limits.
Valve C opens within the 10% overpressure and closes within the declared limit; the discharge coefficient always reaches the same values measured in the built-up back-pressure tests, but it is necessary to remember that the flow rates are very small if compared to the declared values (see valve C in fig. 1)
Finally, valve D opens within 10% overpressure only for a back-pressure ratio of less than 25%, as specified by the manufacturer. Over this limit the opening of the valve requires an overpressure of 16% for a back-pressure ratio of 30%. The blowdown corresponds to the manufacturer's declaration and, once again, the discharge coefficient reaches the same value recorded in the built-up back-pressure tests.
The results of a wide campaign of operating and flow tests performed on five commercial safety valves, supplied according to the same technical specifications have been presented and discussed. Valve performance has been compared to that declared by the manufacturer highlighting the following main features:
- Atmospheric back-pressure tests show fairly good agreement to manufacturer's declaration for four of the five samples, whilst only one sample shows a blowdown of 37% instead of the declared 10%.
- Superimposed and built-up back-pressure tests show several differences with the manufacturer's declarations both in terms of operating and flow performance.
- One of the five balancing bellows failed at the beginning of the back-pressure tests.
Therefore it is quite possible that, during on-site applications, when the valve is required to open under back-pressure conditions, the protected equipment may be exposed to pressure higher than 1.1 times the maximum allowable pressure. This can be due to the valve's inability to pop-up within the overpressure limit and/or to the reduction of the flow capacity caused by a poor disk lift and/or to the establishment of a subsonic regime, also when the back-pressure ratio is below the critical level (0.528).
This data highlights the need for future standards which clearly require a well defined test procedure for the evaluation of the behaviour of safety valves operating under back pressure. Things can be further complicated when a full valve size range has to be characterised by testing only a few samples which comply with the test rig limits; in fact laboratory tests performed on safety valve models show that body size and outlet to nozzle area ratio strongly influence the flow capacity within the same size range. .
Only extensive laboratory tests will allow us to reasonably foresee the performance of a full size range when the valve is operated under back pressure and will improve the safety conditions of plants and pressure equipment.
[1 ] prEN ISO 4126-1: January 2000 :"Safety Valves for protection against excessive pressure- Part 1: Safety Valves (and ISO/DIS 4126-1.2: 2001-11-29, same title)
 V. Dossena et al, "On The Influence Of Back Pressure And Size On The Performance Of Safety Valves", to be presented at 2002 Pressure Vessels and Piping Conference (ASME), Vancouver, August 2002.
|About the author
Mr Vincenzo Dossena graduated in Mechanical Engineering from the 'Politecnico di Milano' University, [Milan, Italy] where he also won a PhD in Energetics in 1994. He started work as a Research Engineer at the same University in 1995, where he is also holding a course of Machinery and Energetic Plants. Since 1994 he has also been working as aerodynamic specialist, with particular emphasis on experimental measurements. He has published several papers on gas and steam turbine aerodynamics. Mr Dossena has been involved in the experimental analysis of safety valve behaviour, both for R&D and certification purposes, from 1996 onwards.
He can be reached on firstname.lastname@example.org