It sounds like the idea is to use a temperature transducer to detect the temperature and develop a 4-20 mA current signal proportional to temperature. That signal is the used as a setpoint that controls the frequency output of the vsd. This is fairly easy to implement. All the drives I've worked with in the past have allowed an analog speed reference signal input; many accept 4-20 mA signals.
Check out manufacturers like Allen-Bradley, Cutler-Hammer, GE Fanuc, Siemens...all will gladly provide selection and application help if you're thinking about purchasing their products. Don't let them give a hard sales pitch, though. It's best to shop around before selecting.

A good general-purpose drives reference is:
http://www.ch.cutler-hammer.com/unsecure/cms1/TD_08H17TE.PDF
Chapter 6 gives assistance in selecting a drive.

I agree with xnuke. If you are not adept at VFD applications, then seeking help with a mfg is a good idea. Please speak with a couple of companies, you will find the advice and pricing often different. The basics are usually the same however.

One caution; really take a look at your power distribution system. Understand all of the specs on the drives are far as voltage tolerance, use of line reactors, and environment. Power issues are the most common cause of drive failure.....which is caused by mis-app and improper installation/set-up.

Carl

Thanks gentlemen, I will read the suggested document and perhaps post further questions. When you say line reactors may I assume you mean inductors? Are these used to control the dV/dt and hence EMI?

This link will help explain the use of line reactors to decrease the sensitivity of the drive to line effects:

http://www.mtecorp.com/reactors.html

Most drive manufacturers will provide guidance as to when they recommend the use of line reactors, either from a tech rep or in their product literature.