Koestler introduced his theory by analysing humour. I think that was very wise of him, because in spite of much thinking about it, I have never found any aspect of humour which the theory of bisociation would totally fail to explain!
All the jokes in Koestler's books, and all the ones we looked at in class, are purely verbal. Every one involved some ambiguous word or phrase, and it was the clash between two possible meanings which constituted the joke. It is hard to imagine any verbal joke which could work in a different way, so the place to look for a problem with the theory would be to examine other things apart from jokes which make us laugh.
A while ago one of our cats was relaxing on my wife's desk as she often does, when she decided to roll over onto her other side. Unfortunately she was too near the edge of the desk for a movement like that, and just fell bump on her back on the floor. My wife laughed a lot at the incident (after making sure the cat hadn't hurt herself, of course!), and I laughed too when I heard of it. Where is the humorous bisociation in that? Actually it's not so hard to find. One always imagines cats as having such grace and elegance in their movements, jumping accurately from one small space to another, always landing on their feet, and so on. So when a cat falls flat on its back, there is a clash between the sudden incident and the image of cats in general. That surely forms the bisociation.
Have a think about humour, though - maybe you can find something that cannot be explained by bisociation, though I would be very surprised if you did. Koestler's theory seems unassailable on this point.
It seems to me that we are much more likely to find weaknesses in the theory in this area. Think again about the story of Archimedes. It is easy to imagine the mathematician/scientist having a "matrix" of ideas about how to measure volume - a network of ideas linked in all kinds of ways. But at the moment of discovery, that matrix was supposed to become connected to another one, a matrix of ideas about bathing. But surely thoughts about bathing are just a loose set of associated ideas, hardly comparable to a set of mathematical concepts. Archimedes was obsessed by the problem he had to solve, so measuring volumes was something always close to consciousness in his mind. When he climbed into his bath and noticed the water rise, he would surely wonder about the amount, instead of just idly noticing the phenomenon as usual.
In other cases it is very hard to imagine what the second matrix could involve. The German organic chemist, August von Kekulé discovered the structure of the benzene molecule in the mid-19th century. The chemical properties of benzene had been very difficult to explain by any theory of its molecular structure, but von Kekulé guessed that it might have a ring structure, and this turned out to fit the facts. The story goes that von Kekulé made his discovery while dozing in front of a fire. He had been struggling with the benzene problem, and while half asleep he began to imagine snakes writhing around. Eventually one of the snakes twisted right round and seized its own tail in its mouth - von Kekulé immediately made the association with a ring structure for the molecule. But what on earth could the "matrix" have been that prompted a vision of snakes? In any case, it seems that the story may be only apocryphal, as you can check for yourselves in this Wikipedia entry for von Kekulé:
August von Kekulé
Of course the Archimedes story itself is apocryphal: there are no historical records of what actually happened, and the bath legend seems to have started a long time after the event.
In fact it may be impossible to find a clear example of bisociation in this kind of discovery, although you may try looking up the details of some classical breakthroughs on Wikipedia.
Think about your own discoveries as well. We all make them from time to time, even in the trivial tasks of everyday life. For instance, while I was trying to think of a good example of a scientific discovery for this bit of writing, I dozed in my chair, and when I woke up I immediately thought of von Kekulé and benzene. A little bit similar to von Kekulé himself, but without any interesting dream! In fact I usually find that my better ideas come about in a similar mysterious way. Doing mathematics homework in my schooldays, I would often be unable to solve some problem before bedtime, but the next morning I would be able to finish it immediately. What happens in our unconscious minds during such events? It doesn't seem to be connected to bisociation. What about in your case?
We tried an example of a poem, but it seems to me that appreciating poetry will inevitably involve bisociation, in a similar way to appreciating jokes. The "melancholy, long, withdrawing roar" of the ebbing tide may be associated with the disappearance of something we have valued. On the other hand, why should the sound of the sea be "melancholy"? Is there something else here as well as bisociation? And what about other art forms - other kinds of literature, drama, or the visual arts? I'm not a very artistic person myself, so I'll leave those to you!