Recently, I have been looking into more effective methods of harnessing nuclear power and making an attempt to move away from conventional steam plants and the use of turbines, thus reducing heat losses, maintenance and construction costs, etc. That being said, my current concept uses a reactor to transfer heat to a coolant, which is moved through piping via a pump. The piping is coated with a thermally conductive material that is not electrically conductive. Around this layer are multiple thermoelectric generators, all connected in series to increase voltage output. On the outside of the TEG's, a refrigeration unit would keep the outside of the units around -10 to -20F. In areas with cold climates, outside air may be used as a heat sink for the condensing unit of the refrigeration system instead of a body of water. The temperature difference (around 350K) will be enough to generate small voltage in each unit, then the series connection adds all of the voltages. This output is then run through a step-up transformer to get it high enough to be used for electrical distribution. Waste heat due to I^2R losses from the transformer will be removed via another coolant system, which dissipates that heat into a regenerative set of TEG units, cooled by the same refrigeration unit as the main units. To compensate for the drop in current while stepping up the voltage, multiple transformers would be used in parallel to achieve an appropriate amount of current. A single reactor would have between 8 and 12 coolant loops, each loop having multiple TEG piping units. All units would connect to 2 different outputs to supply electricity to the power grid. Please share any opinions, suggestions or questions you may have. Any input from engineers would be greatly appreciated.