It's not "well-known" at all. As best as can be made out, the scale of the problem has been exaggerated. One need only look at recent, wholly unsubstantiated, claims that "millions" of votes in the 2016 US election were fraudulent to see that it suits some people's interests to take a problem that exists at a small scale and then massively inflate it.
In terms of electoral crime, voter impersonation simply does not register. Even when it does occur the effect is simply swallowed up because the odd one or two fraudulent votes will have no impact on the result of elections with margins of hundreds. Far more of a concern would be campaign finance violations, for example; although I'm not going to argue that some elections are "bought", it still stands to reason that parties campaigning should be able to do so in a level playing field, so that they have in principle an equal change of getting their message heard; financial violations break that principle even before you ask questions about where the money came from.
Instead, concerns should be directed towards why turnout remains generally very low, and why even with full turnout, something in the region of 10% of UK adults aren't even registered to vote in the first place. Since that leads to millions of "lost votes" -- albeit by choice -- as compared to the rare cases of fraudulent votes, it should be clear that the way to improve the trust and the reliability of elections is to encourage further engagement, rather than set increased restrictions, however well-meaning.