I'm glad you like the idea! We'll make this work for you.
While it would be great for you to get 100 people for your project, that's probably not feasible and it's not totally necessary. You'll certainly want to get as many people as you can. You could try asking everyone in your class to participate, or you could ask family and friends. (Keep in mind that age, gender, and fitness level can all influence blood pressure.)
Let's say you chose 3 drinks to compare. If you had every subject in your study perform experiments with all three drinks, then your sample size would be fairly large because it would be the same for all three drinks. You would have to do a special statistical test, called a repeated measures Analysis of Variance (rmANOVA), but we can worry about the statistics once you have done your experiment.
If you have each person only experiment with one drink, then your statistics will be a bit easier but your sample size will be much smaller. So it's a trade-off.
Ideally, I would aim for about 25 people, and study 3 drinks (2 with electrolytes, and water as a control). You will want to test all of your drinks (at least 3 different bottles of each) for conductance, and calculate an average conductance for each drink - inlcuding the water.
For the experiments, you will want to take resting blood pressure prior to the drink, then ask the person to drink a set amount of liquid (which should be calculated as a fraction of their body weight, such as 10mL of liquid per kilogram of body weight). Then take their blood pressure at set intervals (say, every 30 minutes) for 2 hours after drinking the liquid. You should also have a control group that drinks no liquid, but still has blood pressure taken every 30 minutes for 2 hours.
Thus, you will have four groups: (1) no liquid, (2) water, (3) electrolyte drink #1, (4) electrolyte drink #2. In each group, you will take resting blood pressure, then have the person drink (or not), and repeat blood pressure measurements every 30 minutes for 2 hours thereafter. If you have 25 people in each group, that would give you 100 people. If you have 25 people do all four treatments (on different days), then you will still have a sample size of 100.
If you can't get that many people, try seeing how many you can get, and whether they are willing to have you measure them multiple times. You should either randomly assign people to the different groups OR do all 4 measurements on everyone. You would not want to repeat measurements on some people but not others.
Let me know if this makes sense. Do you think this experimental design will work for you? How many subjects do you think you can get to participate?