Safe Safety Valves (1)

Sentinel 7109's new safety valve system is taking shape as shown below.

The Flange and Supporting pipework

I introduced the new pair of Bailey-Birkett type 716SSL safety valves a little while back. I've now obtained the supporting steel pipe fittings and constructed the new support manifold.

Whilst on the surface this construction seems fairly straight forward, it is a safety system and therefore requires rather more formal engineering processes to ensure it is fit for purpose. (It's also my safety I'm concerned about!).

Before I retired, I spent many years on railway signalling and control centre research and development projects. Most were computer based and hence involved formal safety related system/software development project 'V' life cycles. ('V' implies: Define what it has to do; design it; implement it; test the implementation; test it meets the design; test/assess whether it does what it was defined to do originally. And don't forget the traceability!).

Safety valve pipework is somewhat more tangible than software so it would be over the top to do all that but some allusion to the principle is worthwhile.

Requirements:

Firstly, what are the requirements (i.e. WHAT does it have to do)?

The pipework shall:
1. Withstand 275psi.
2. Withstand 230DegC.
3. Support the safety valves.
4. Support the exhaust outlet pipework.
5. Allow the full output of the boiler to pass to the safety valves (4600 lbs/hour).
6. Connect to the original Sentinel boiler flange mounting.
7. Prevent condensation accumulating in the exhaust (i.e. to prevent showering nearby onlookers when the safety valves blow off!).
8. Enable easy removal of the Safety Valves for hydraulic boiler testing.

Design:

Requirements 1-8 are ultimately tested or observed; however, they all need to be taken into account during the design process.

Requirements 1 and 2 (275psi & 230DegC) are satisfied by choosing to use mild steel pipe and fittings and avoiding the weaker malleable iron which is not up to the job. It also requires that a flange to connect to the existing Sentinel flange has to conform to BS 10 (1962) Table F or better.

Requirement 3 is satisfied, firstly, by ensuring that there is a vertical 3/4" male BSP thread to connect to each safety valve and, secondly, by using sufficiently heavy gauge material.

Requirement 4 is satisfied again by using the heavy gauge material. Additional support from the cab roof may also be included later to support the exhaust pipework.

Requirement 5 is satisfied by ensuring that there are no parts of the pipework that are of a smaller diameter than the 3/4" inlet to the safety valves.

Requirement 6 is satisfied by using a compatible flange. Easy you would think but when Sentinel made their safety valve mounting, there were no standard flange sizes and ratings! So the nearest standard type needed to be chosen that involved minimal adaptation.

This resulted in a Carbon Steel flange as follows:
A screwed flange of nominal bore 1.5" (to allow a male 1.5" BSP thread to be attached).
A BS10 Table 'F' type capable of 300psi at 232.2DegC.
5.5" Diameter.
Four holes 11/16" diameter on a pcd of 4.125" to fit 5/8" studs.
0.5" thick.

The four mounting holes had to be elongated slightly to fit the Sentinel flange.

The flange dictated that 1.5" pipework had to be used which conveniently led to the considerable strength of the final structure and avoidance of any diameter less than the 3/4" inlet to the safety valves (requirement 5).

Requirement 7 is satisfied by including narrow bore draining pipes in the exhaust pipework.

Requirement 8 is satisfied by incorporating unions into the exhaust pipe elbows so that the pipe can be detached easily. The safety valves can then be unscrewed without having to take the exhaust pipework apart. A cap is screwed onto the 3/4" male thread to seal for hydraulic boiler testing.

Implementation:

Below is an early mock-up I did at South West Engineering Supplies to get the hang of the idea. It was not the first attempt; I started off using 3/4" pipe bends from the 'T' piece to the base of the valves but did not believe it was a strong enough structure to carry the weight.

Early Mock-up

The implementation evolved to the final version by using shorter 'running' nipples having no plain centre section to keep the links as short as possible and by using single piece reducers to convert from 1.5" to 3/4" diameter fittings. Thus I believe I've minimised the number of thread interfaces and maximised the strength of the structure.

The Final Version

The black 'stuff' on the threads is a jointing compound from Rocol called 'Steamseal'. It is more technically called 'Foliac Graphite and Manganese' and is specified to be able to withstand steam pressures up to 2800psi at temperatures up to 600 DegC. Good stuff! Easily exceeding requirements 1 & 2.

At this stage, requirements 1-6 have been embodied into the design and implementation.

Requirement 7's condensate draining pipe is shown below.

Condensate Draining Pipe (bottom left)

To prevent a condensate pool accumulating, the small brass fitting had to be machined so that it would not protrude into the large pipe.

Flush Drainage hole

I hope I've demonstrated how setting out the requirements before design and implementation drives towards a compliant solution. Often, it's easy to assume that what has been done before will do but that approach tends to only just get what you want if you are lucky and doesn't cater for doing something that hasn't been done before.

Of course, if you haven't got ALL the requirements identified at the start, you may still not get what you want!

Next to do is the upward exhaust pipework.