I did realize that it was non-linear.DH , friction loss is not linear , your calcs assume it is , this is why you have two different losses for the same gas flow rate (maybe).
Where does the 15 come from?23.4 metres of 22mm delivering 3.37 m3 of gas would have a pressure loss of...
23.4/15 = 1.56 mb.
Presumably you are using different tables.15 metres of 22mm tube will deliver 3.4 m3/hr of gas at a loss of 1mb at each end. It seems your chart uses 3.6 m3/hr for 15m of 22mm
I have found the same error on each of my tables I had summed 1-2, 2-3, and 4-5 to give the total at the boiler. It should be 1-2, 2-4, and 4-5.your calcs are also completely different to that of the chart you posted
However, I did not follow the instructions in the document as there seemed to be a flaw in the logic in that it assumes the maximum permitted flow is going down the pipe. To take just the section 1-2 as an example:
Assuming 28mm pipe, effective length of 23.4m and a flow of 3.37m³.
According to the instructions, the longest 28mm pipe which can carry 3.37m³ is 30m, so the drop will be 23.4/30 = 0.78mb.
But hang on a minute, a 30m pipe can carry 4.8m³ with a 1mb drop. But we are not putting 4.8m³ down the pipe, only 3.37m³, so the loss must also be lower due to the reduced flow. I assumed it was linear and multiplied 0.78 by 3.37/4.8, giving a drop of 0.54mb.