It depends on which message you want to explain and what the main focus is when combining data.
Classically, the spatial distribution of material properties, water quality and the flow field (measured / modelled) are shown first. Depending on the statement, Piper, Durov, etc. are added.
To my knowledge there is no extension of Piper and Durov. That is also difficult, because these diagrams already show certain complex correlations and additional information would make these representations difficult to understand. The best thing is perhaps to try out which "additional" combined representations best convey the relationships.
There is e.g. the possibility of Swarmplots (Scatterplot with categorical variables) but even Heatmaps one could consider if there is enough information for a statistical evaluation.
I think you might find a good overview here: https://seaborn.pydata.org/examples/index.html
A look at https://pandas.pydata.org/ is certainly also helpful.
Piper's diagram is commonly used to characterize the groundwater irrespective of the host aquifer. I have also used Piper's diagram to characterize the groundwater type of my study area consisting of Alluvium aquifer (Article Fluoride abundance and their release mechanisms in groundwat...
).
Piper's diagram illustrates the dominance of the major ions whereas Durov's diagram can also provide an idea about the hydrogeochemical processes.
Some researchers have used a combination of the similar diagrams such as Piper's, Durov's and Chadha´s diagrams to illustrate the same. However, Chadha´s diagram is the latest one compared to others.
Piper's is traditional but worth using it. Durov's & Chadha's must also be used to show the trend using yours, in combination with other published data to understand the qualitative aspects.