If you're making your very first voicebank, I would recommend starting with CV Japanese instead of jumping straight to VCV. CV Japanese will allow you to create a complete voicebank with a minimal amount of recording and become familiar with the process of setting up and configuring voicebanks. Here's a guide you can follow for configuring CV:
https://utaforum.net/resources/cv-otoing-guide.269/
The best way to reduce noise is to avoid sources of noise in your environment. Turn off air conditioning/electric fans/etc. Ask others near your room to be quiet. If there's additional noise from outdoors, wait for a time of day when there's less of it.
You might also encounter a noise floor coming from your recording equipment itself, which may sound like constant white noise or constant humming. This can be fixed with equipment upgrades, which will improve the quality of the recordings in general.
Only taking all of those measures should you consider editing files after they're already recorded. Audacity's noise reduction and other free noise reduction software are optimized for speaking vocals, which means that you will be able to understand words more clearly, but the tone of the voice will sound lower quality. Professional-grade software like Izotope RX can help preserve the quality of the vocal while carefully eliminating defects. You may be able to find people in the community who can offer this as a service.
You don't need to, but you can record multiple pitches or multiple expressions if you want to. Again, if this is your first UTAU voicebank, you should stick with the basics until you're comfortable with the process and are willing to repeat things on a larger scale.
You could if you wanted it to be permanent, but you can also just use g flags within UTAU. Go to Project > Project Property, and in Rendering Options you can type "g+(number)" to make it sound more masculine or "g-(number)" to make it sound more feminine.