Industrial valves and fugitive emissions. A topic, that of the interaction between this component and its external environment, which has accompanied us with surprising regularity over the last 30 years, always creating a lot of interest and discussions, so much so that even today there is no Scientific Conference in which this topic cannot find space with significant follow-up. After an initial period characterised by sometimes not entirely clear guidelines, in which the manufacturers of valves and packings came into contact with the matter, the Low Emission regulatory framework was consolidated. Today the requirements of the products and the tests to be carried out are well defined and together with the LDAR – Leak Detection and Repair – protocol which monitors the correct emissive functioning of the Valves at the plants – they constitute a Reasonably Achievable Control of Technology (RACT) to reduce and contain the emissions of VOCs and HAPs from the Valves.
After so long and at the onset of this new season, in which the reduction of the environmental impact and the search for better sustainability are becoming increasingly important, it is perhaps worthwhile to take stock of the situation, to investigate what could be the improvement drivers for each of the players involved (O&G and Chemicals Producers, EPCCs, Valve and Sealing Manufacturers).
The regulatory framework for Low Emissions tests for valves and packing
Many publications have recently illustrated and compared the main Low Emission (hereinafter LE) test standards concerning valves and packings. In general terms, however, a quick summary will be useful to the reader.
The ISO 15848 test, both prototyping and production, is in its scope a test for Valves even if it actually approves the configuration of valve and packing as an assembly. In no way does it qualify the requirements of the packing which are instead defined by ASTM F2168 and F2191 and by EN 14772 section 6.7, sometimes amended in some part in the technical specifications of the End Users. The API std 622 test, on the other hand, is aimed at the approval of the packing and uses a fixture for the tests. It also defines the physical-chemical requirements of the packing by detailing the Packing Materials Test and referencing the MSS SP-120 for other requirements. The conferment of the LE attribute to the valves using API std 622 approved Packings requires the execution of the API std 624 (rising, rotating, rising and rotating stem valves) and API std 641 (quarter turn-off valves) tests. Finally, the TA LUFT VDI 2440 test, of substantially regional diffusion, which is aimed at approving both the packings, with a specific test conducted on a fixture, as well as the valves themselves. Finally, it is worth remembering that the three standards also differ in the combinations of temperature and pressure, as well as in other technical details that it is not in the scope of this article to examine in depth.
It is important to add that the IOGP association – International Oil & Gas Producers – published two specifications for industrial valves in 2019, the S-562 (Supplementary Requirements to API Specification 6D Ball Valves) and the S-511 (Supplementary Requirements to API 600 Steel Gate Valves and to API 603 CRA Gate Valves), in which ISO 15848, ASTM F2168 and 2191 together with EN 14772 with some amendments are adopted as standard to define the LE requirements of the Valves and those of graphite packing. It appears inevitable, since many of the IOGP associates are American and due to the importance of the American Petroleum Institute standards, that the aforementioned specifications are over time harmonised with the API std 622, 624 and 641 and the MSS SP-120. In fact, the now established trend, from the point of view of LE packings, is to acquire all three approvals in order to meet the demand for LE products in compliance with all standards.
The key aspects of Low Emission tests: oxidation, temperature and number of mechanical cycles.
What we can say after so many years is that « pure » graphite is unable to pass a Low Emission test without having undergone specific treatments. There are at least three things we have learned about graphite when facing an LE test: in relation to braided packings (wiper rings) according to the ASTM F2191 classification, Type I and II yarns (Continuous and Discontinuous Carbon Yarns) are unusable while the Type III (Flexible Graphite) is functional for the purpose, as correctly indicated by the IOGP specification S-511 and S-562 which prescribes this Type; the typical permeability of expanded graphite is too high, interfering in tests with Helium that last beyond a certain time; the friction coefficient of the expanded graphite is too high, affecting the endurance acceptance criteria of the LE tests.
To fill these gaps it is inevitable to resort to impregnation of graphite, in order to correct permeability and friction coefficient. But what is the price to pay by adopting this strategy? Typically, the graphite impregnation alters the chemical scenario of the packing with an increase in the risk of corrosion for the stem and making all the tests aimed at quantifying the detrimental materials questionable. But the real problem is that beyond a limit temperature, depending on the impregnating agents used, there is inevitably a weight loss of the packing due to their collapse, with an immediate reflection on the elastic thrust towards the stem and the stuffing chamber, verifiable by the torque reduction of the stuffing box nuts. The TGA diagram shown below perfectly illustrates the situation described above (in the valve stopping chamber the phenomenon occurs more slowly but the mechanism is the same). At the end of the first hour at 150°, to eliminate residual water, the weight loss is about 1%. As soon as the temperature rises to the test threshold (670°C) after a few minutes the impregnating agents collapse and subsequently the graphite, protected by oxidation retardants, oxidises by just 5 % in the following 5 hours. Based on the requirements of EN 14772 section 6.7 we could say that this TGA is not compliant in the first phase, because the oxidation is higher than 4% per hour, while it certainly is in the second phase and in its entirety we could still define the overall performance as excellent because the WL Weight Loss was between 10 and 12% in the 5 hours of testing.
Example of TGA test
Some things are immediately evident. The first is that the execution of a TGA test on a packing with LE target is perhaps conflicting with the main objective (to contain emissions) as it seems to be highlighted by the IOGP specification S-511 which states in section F.3. 13.12 Oxidation Test – F.3. 13.12.1 Purpose: « This test does not apply to packing materials containing polymeric lubricants (e.g. PTFE) or blockers ».The second thing that immediately catches the eye is that for the success of the LE test it is essential that the actual temperature of the tow chamber does not exceed a certain limit value. But let’s go into detail. For its purpose, API std 622 defines the applicability of the standard to graphite packing for use from -29°C to +538°C, prescribing, among the Packing Materials Test, the WL Weight Loss test, the low and high temperature Corrosion Test, verification of the content of PTFE and Wet Lubricants, and finally the measurement of the content of Leacheables (Chloride and Fluoride). We remind you that the Low Emission test is carried out at 260°C measured in the tow chamber.
Therefore, in this technical context, the packing design must use impregnation within extremely precise weight limits since service at 538°C must be guaranteed. The WL Weight Loss test intends to monitor this circumstance. Let us ask ourselves what we can ask of materials from the point of view of their resistance to temperature. Up to what temperature and for how long can impregnants perform their function? Asking the packings to fulfil the LE requirement and at the same time suitability for service at 538°C, generally steam, appears to be in some ways irreconcilable.
So far, the argument has been essentially about temperature and nothing has been said about the endurance acceptance criteria. In summary terms, we point out that the API std 622 test requires for 1510 cycles (310 for API std 624 and 610 for API std 641), the ISO 15848 test for the Isolating Valves requires 205 cycles for the C01 class and 1500 for the C02 class, the TA LUFT VDI 2440 is traditionally consolidated on 200 cycles, in the absence of substantial specific indications of the standard.
But why is the number of mechanical cycles so high in such a limited amount of time? This fatigue test that stresses the packing (and the valve) in a rather unnatural way, gives us indications of the quality of the packing (or of the valve together with the packing) or rather obliges the packing manufacturer to adopt every possible strategy to reduce the coefficient of friction of graphite, which in the absence of interventions is stably between 0.15 and 0.25. The problem arises above all with the ISO 15848 standard in which the test is conducted at the valve rating pressure while in the others (API std 622, 624, 641 and TA Luft VDI 2440) it is conducted at a maximum of 40 bar. This determines, with the same configuration of the sealing system, the application of a higher tightening torque of the packing with direct effects on the friction coefficient of the graphite which is not constant but increases according to the applied load.
In conclusion, from a conceptual point of view, can it be considered correct to design a packing to pass the Low Emission test or would it be better to design a packing capable of maximising its performance when the valve is in operation in the system? Does the first objective include the second or are they not coincident with each other?
But does anyone know how the Valves in operation behave from an emission point of view?
The information about the emission behaviour of the Valves in operation is available to the Plant LDAR manager, where the surveillance programme LDAR is implemented, but objectively very few aggregate data exist. A unique document of its kind is the 1997 API publication “Analysis of Refinery screening data”. The document illustrates the data collected at seven Californian refineries between the fourth quarter of 1991 and the second quarter of 1996 performed according to the EPA method 21 technique (the same adopted in the API std 622 test) at all the Equipment Leaks of the refineries participating in the project. The table below shows that in cumulative terms, i.e. on the total of screenings carried out, it is 1,450,000 for the Valves Gas Service and 1,340,000 for the Valves Light Liquid Service, compared to the Leak Definition of 500 ppmv the Leak Frequency of the Valves was about 1.00% while compared to the Leak Definition of 10,000 ppmv the Leak Frequency was about 0.25%. The following columns indicate the repetitiveness of the leakage on the same components or measure how many Valves identified as Leakers in the previous campaign were also so in the following campaign. The complete report, available in the API library, details the performance of the components of each Refinery Unit, highlighting what can be expected, namely that the Leak Frequency is correlated to the temperature, pressure and volatility of the fluid and therefore differentiated between different Business Units.
These data are very far from those that had been estimated by EPA in the publications of the 1980s and which attributed to the Valves in gas service a Leak Frequency of 11.40% and for those in Light Liquid service equal to 6.90% compared to Leak Definition 10,000 ppmv.
API Analysis of Refinery screening data – Valves highlight from table 3-1. Screening Results for Seven Refineries (5.5 years)
What does this data tell us and how could we interpret it? It is not known how much these Leak Frequencies are influenced by the rather widespread practice, even if it is not recommended yesterday as today, to place the many manual Valves when fully open with the stem in counter sealing. We therefore do not consider this interference in the analysis that is being conducted.
First of all we could think that they refer to a historical period in which the Equipment and the packings were not designed according to a Low Emission logic and that therefore it would be reasonable to expect that today a net improvement has been achieved, this following the introduction of Equipment and Low Emission Engineered packings.
On the other hand, the results of the API publication are still substantially confirmed today by the numbers collected in the field by the numerous companies that carry out LDAR monitoring, which rarely detect aggregate Leak Frequencies higher than 1.00% with Leak Definition 500 ppmv (the aggregate term intends to mean that there may be Production Units where the Leak Frequency is greater but considering the set of all Production Units the Leak Frequency converges to lower average values). On the basis of LDAR experience, we know that the Cluster of Leakers is made up of Valves characterised by one or more of the following attributes: they are valves with frequent actuation (on the other hand, those that remain in a state of rest for a long time tend to have a much better Leak Frequency ); they are Valves characterised by use at temperatures above 260°C measured in the tow chamber; are Valves characterised by a pressure use higher than the 600 psi class; are Rising, rising rotating, rotating stem valves (on the other hand quarter turn-off valves have negligible Leak Frequency). In conclusion, the area of possible improvement is identified in the resolution of the specificities of this Cluster in which the Valves and seals are called to perform more severe than this Cluster.
The Low Emission strategy linked to Industrial Valves is based on some cornerstones that are not in question. They are the qualification of products specifically engineered for the purpose (Valves and Packings) and on the control of Equipment in operation with the LDAR surveillance routine. It is desirable that thanks to the growing potential of information management it is possible to highlight in advance, right from the design and purchase phase of the Equipment, those Valves that belong to the Cluster with greater probability of high Leak Frequency in order to pursue practicable solutions. Finally, from a regulatory point of view it would be advisable for the test standards to clarify some points of conflict, this in the common interest of improving the quality of the products and their safety of use.