You could use an ac electric motor of known RPM and gear it down to the RPM of interest (some people use a drill press). You measure the voltage across one pair of leads of the star connection (open circuit voltage) at the speeds of interest. You should also measure resistance between the lead pair. The maximum power that the alternator could then deliver can be estimated by terminating each lead pair in that resistance and calculating the power into each load resistor and multiplying by three. The problem is that you probably won't use the alternator in that way. If you rectify the voltage and feed the current to a battery you have both the battery voltage and battery resistance being important as well as the voltage drop across the diodes. Each lead pair will provide the battery current one third of the time. While the lead pair current will vary somewhat from the DC average, you could assume the DC average to find the power out. You have a circuit loop with a voltage source equal to the lead pair open circuit voltage in series with the lead pair resistance, two forward biased diodes, the battery resistance, and the battery voltage. Note that the battery voltage will oppose the lead pair voltage as will the diode voltages. From this you can calculate the current through the loop and the power delivered to the battery or the power lost in the alternator.