There are two different things to ask here:
1. Why doesn't the railway install solar PV panels on it's own land or buildings to generate electricity and feed it into the grid?
2. Why aren't we seeing solar PV farms feeding directly into railway electrification?
The main answer to 1. is that we did see this (Blackfriars station roof was a pioneering project) when the financial climate was right (subsidies meant that there was a secure return on the capital invested). When those subsidies went down much faster than the price of the equipment, the economics didn't make sense, so the whole market ground to a halt. The subsidy regime for field-scale solar farms was different and rates changed at different dates, with some schemes able to qualify well before being completed, so it took longer for that sector to stop. Luckily the rest of the world didn't stop buying, so prices have continued to come down. That means that it is now economic to install PV, if a good proportion of the energy is used directly on site (and so doesn't attract charges for use of the energy network). This has lead to projects like GTR's Streatham Hill depot:
https://www.ianvisits.co.uk/articles/south-london-train-depot-to-be-covered-in-solar-panels-51633/.
The best sites for using energy on site are always likely to be buildings, so I would expect to see these sort of rooftop projects first. But there is a need for certainty about being their for the contractual payback, so it would a very hard sell right now with the GBR legislation not published! Best bet would be one of the long term train maintenance contracts like DLR Poplar Depot rebuild. The railway doesn't tend to have that much land which is flat, seggregated from trains (to make access for maintenance easy) and in sensible shaped parcels (not long thin strips!). Any such land generally got sold off at privatization, or is protected for future railway use, so they won't want to put 25 year lifespan assets on it. All these mean that ground mounted solar farm developers are likely to see the railway as a hard job, and go somewhere else for land first. Network Rail itself has capital borrowing limits, which makes non-core capital projects like this hard to fund.
Your link directly relates to the second question. If you are building a many-fields-scale solar farm, one of your key problems is getting connected to the electricity network without it costing an arm and a leg (if upgrades to the existing network are needed to connect your development, you are likely to be asked to pay!). Electrified railways are interesting because they have their own private electricity network alongside them, one of the very few that exists. In general, railway networks are much more tolerant of the sort of somewhat "dirty" power put out by small scale generators than the public supply - because the railway itself is an incredibly dirty user! This is one reason why getting supplies for railway electrification can be expensive, it has to be done in a way which doesn't make the neighbour's lights flicker! So there is a real opportunity to connect solar farms adjacent to (or on) electric railways directly, in order to get a cheaper connection - this is exactly what the project you linked to the report on set out to show. It no where near met the line's demand, but it didn't have to get a public grid connection in order to export the power it did produce. There are some technical hurdles about power going "the wrong way" even on the railway (despite the use of regenerative braking), which is what the project looked to overcome. I believe it was a success, but the DC rail network (or rather the 11/33kV 3-phase AC backbone behind it) can't absorb huge amounts of power, and largely exists in the crowded south east, where land for panels expensive.
The big win is being able to connect to 25/50kV single phase electrification system. That's rather harder (because it's non-standard for solar farm kit) but not any harder than a Static Frequency Converter - which are now in main line use in the UK. I believe that there was a next phase project to demonstrate this using a 25kV transformer and traction supply, with the panels connected in place of the traction converter side of the DC link. Effectively it would have been a stationary train re-generating into the OHL, but I've not heard anything about this since then. The drive for doing work will have been limited by the economics of solar power generation as described above for grid connected solar, although that should be getting better now.
