Selecting the k most relevant candidates from a larger set is known as top-k ranking. Traditional ranking methods prioritize candidates based on their relevance, which can lead to discrimination. Due to the AI Act, fair top-k ranking has recently gained attention. We introduce a new positional fairness metric that considers the ranking positions of groups in the top-k ranking. Secondly, we propose a novel algorithm, FairNormRank, that optimally fulfills the three fair top-k ranking criteria of proportional fairness, maximum relevance, and ordering consistency and accounts for positional fairness. Our method works for non-binary and intersectional groups, therefore enhancing its applicability in realistic scenarios. An evaluation on a real-world dataset shows that we outperform existing methods in terms of fulfilling the fairness criteria.