What’s Smart about the Smart Grid?
I recently had the opportunity to be a part of a panel discussion on smart grid technology at APEC 2010 in Palm Springs, CA. The panel was made up of some outstandingly bright folks (which begged the question on why they chose me). The central theme of the discussion was on smart grid communication. The main take away was “YES, the folks that sell IC’s for a living are ready, willing and able to take money for their communication chips and everybody claims to make the best chips in the world”.
But I’m not sure that solves anything. If we look at the load base, the loads are everything but smart. The multi MVA wind turbines break on average every two years. As engineers designing these machines, we need to go back to the drawing board and redesign the motors and alternators for the applications. Until we do this, we will all be penalized by the utilities.
The statistics vary, but all seem to be in agreement that nearly 30 to 40% of the present load base on the grid is caused by induction motors. They appear in furnaces, condensers, less expensive clothes washing machines and clothes dryers, pumps, fans, blowers, draft inducers, etc. That seems odd to me. Most off the shelf induction motors don’t have an efficiency beyond 66%. We then band-aid this in applications where variable transfer rates are required with a voltage sourced inverter to vary the frequency of the applied waveform. When the induction machine is operated outside of it’s shamelessly dogmatic designed frequency range, it’s efficiency is less than the 66% peak. Not to mention the power factor of the machine is terrible, typically around 0.7 lagging.
Let’s not congratulate ourselves for this and blitz wildly toward a grandiose communications solution to make our grid smart when our loads are plumb stupid. We need to go back to the drawing board and take advantage of the newer motor technologies (sorry Tesla!). We have the technology and the fundamental knowledge to build machines that truly do have high efficiencies and broad speed ranges, it just seems that the industry is stuck in a rut. We keep designing more and more induction machines for all the wrong applications. All of the induction motor designers benchmark other induction motors and make small improvements in the stack, the shorting bars, the slots in the stator, etc….all the while ignoring the obvious questions: What is the application? Why not design the right machine for a given application? Why not take advantage of far superior motor technologies including PM brushless machines and high pole count machines for low speed high torque applications?
We know the answers to some of this: it’s not the path of least resistance. The machine designers are faced with mature tooling and production capacities, designed to build induction motors. There’s likely thousands of tons of laminations for the stators, tons of shorting bars, end bells, etc in stock. Also, in these economic times, I can’t imagine the notion of retooling the production lines, developing new machines, and re-validating these machines would gain a lot of traction with upper management faced with endless ROI justification, reductions in force, and decreasing time to market expectations from the financial community.
Now here’s where the path of least resistance approach gets problematic. Let’s take a look at the city of Bakersfield California (as one gentlemen that asked the question of the panel reminded us). Pacific Gas and Electric (PG and E) decided to add some value and update the residents of this fine city with “smart” watt-hour meters on their homes. Much to the resident’s surprise, their bills went up 10x on average with the advent of the “smart” meters. Smart for whom? The answer seems clear. Why then would this happen? Well, first, the smart meter knows what time of day it is. It knows the demand on the grid. During higher demand periods, the smart meter applies a scalar multiplier. When power is perceived as scarce which corresponds to a heavy load on the grid, the rates go up. At minimal demand periods, the scalar multiplier is rolled back to unity (wouldn’t it be nice if it were less than unity?). The city of Bakersfield is now in litigation with PG and E for what they claim is an unfair adjustment. If the meters are so smart how could that be unfair?
If 30 to 40% of the loads are induction motors, we are spending a huge amount NOW on induction motors and soon to be an ASTRONOMICAL amount running these squirrel caged beasties via a “smart meter”. Roughly every 2HP output from an induction motor makes 1HP of heat that we also pay for. There’s nothing smart about that.
How do we fix this?
In retrospect, the white goods industry has done a fantastic job designing high end appliances with machines far better suited to the applications. Permanent magnet synchronous machines are being used to drive the load directly without gears, transmissions that leak oil, etc.--a commendable job indeed.
The HVAC industry is still struggling, however the path of least resistance is soon to change. When the SEER requirements on your AC system/Condenser go up to 14 and above, an induction motor simply can’t be had with its inefficiencies. This WILL drive PM solutions in truly variable transfer applications as opposed to slamming off and on. The best solution at present is the Turbocor compressor from Danfoss. They went back to the drawing board and built the right machine for the application. They opted to eliminate the need for oil in the refrigerant to lubricate reciprocating pistons or screw compressors. Instead, they built a turbopump with a bldc motor than runs efficiently from a few hundred RPM to tens of thousands of RPM. To eliminate bearing failures, they used mag lev bearings and microprocessor intelligence. For medium scale commercial applications, this IS the best solution available. The turbocor from Danfoss IS SMART! Unfortunately, the overwhelming majority of the loadbase from refrigeration and heating is still induction motors. There’s nothing smart about that, but change is coming.
The wind power industry is perhaps the most guilty of all. By analogy, a 3MW wind turbine is a high power machine in a small space that takes energy from a turbine moving at low speeds (at or around 20RPM) and converts this to power to be distributed or sold back on the grid. All gears and power electronics are located several hundred feet in the air on a tower. The locomotive industry did very similar (sans the tower) only the power flowed the opposite direction. In a locomotive, there is an engine-gen set capable of producing 5000HP or so…roughly 3.7MW. When diesel-electric locomotives were first being prototyped, it was determined that gear boxes simply couldn’t handle those power levels and speed reductions in a small space, hence traction motors were built into the trucks (yes, there is one gear set per traction motor, but it’s certainly not a planetary gearbox). Locomotives last a LONG time. History has taught us valueable lessons that has allowed us to build them very well. The folks that make the large scale wind turbines are presently transferring the 3MVA output capacity through a planetary gearbox to the alternator in spite of hundred year old, proven historical teachings.
This is sub optimal and history has taught us this lesson clear back to the locomotive industry and LeTourneau’s work on prime movers. As validation of this reliability problem, the actuarial scientists that insure these wind turbine investments rate these gearboxes to fail every two years. Please recall that this is several tons of gears, located several hundred feet in the air on a stick. And this solution is smart? locomotives of the same power level last 40 years! The technicians that maintain the large wind turbines change out gearboxes every 2-3 years per wind turbine, consistent with the actuarial models. The tool and die makers and hobbsman that make the gears curse the solution as being a huge reliability problem in spite of being mandated to use the strongest steel available, single tool hobbs, slow heat treating, magnaflux testing, xray testing, hardness testing, etc.
Most anyone actually doing the work will tell you that the multi-MVA wind turbines are a prime candidate for the right machine design -- and that’s putting it kindly. Norm Rittenhouse, world class innovator and founder of EVT has clearly demonstrated that a multi-pole, transverse wound, PM machine can not only make the same amount of power, it is MUCH more efficient in that it ELIMINATES ALL GEARS and runs the alternator directly from the turbine blade. Imagine that? The right machine for the job? With Norms machine, we eliminate the frictional loss of hundreds of KW in the gearbox, also we eliminate the DOMINANT failure mode of gearbox failure and the associated multi-million dollar service calls with a Grove 300,000LB crane every two years to fix broken gears. The EVT MACHINE is SMART!! This machine will bring wind turbines into the same reliability timeframes as locomotives. Again: the right machine for the application!
Before we get all wrapped up in which communication protocol is best and whose chip is the cheapest, perhaps we should take a look at the fundamental problem. If energy is getting scarce enough to meter nearly every VA delivered with a smart meter, and adjust rates dynamically based on the load on the grid, we clearly have a load problem. I suppose a shiningly brilliant communication chip can shut those loads off for us, which will have to happen if we don’t change our present machine designs…..of perhaps we could change our present machine designs, design the right machine for the job and take advantage of PM machines using increasingly cheaper neodymium magnets, whose efficiencies approach the ideal.
When judge Harold Greene swung the gavel on telecom deregulation back in the 1982 I don’t remember the phone bill getting magically cheaper, nor the service any better. The line entry points that I admired as a kid with drip loops, orthogonal runs, and perfectly plumb drops are now reduced to some skinned twisted pair jammed into a ragged hole by somebody with a magnetic decal on the side of their truck that should have worn a hardhat more often. The main claim that sealed Ma Bells fate of “bundling services” is now presented on my “media bill” as “BUNDLED SERVICES”—only now it’s more expensive and there’s more taxes on it.
Sometime after the verdict was rendered, Judge Greene stated that he did what was right under the constitution, but this was a horrible mistake that would take decades to iron out. We’ve seen the same kinds of things with power grid deregulation, only this time we are headed toward rate hikes will be VERY high compared to what we saw when Ma Bell broke up. As engineers we can impact this problem at the source by going back to the fundamentals and designing the right motors for the applications.
We simply can’t band-aid Tesla’s squirrel cage machine with drives that make it less and less efficient. Just like Edison’s light bulb has by and large come to pass, the same needs to be said of the induction machine before we can call our grid “smart”. Further, we need to take advantage of the multipole, tranverse wound technologies offered by EVT and make our wind turbines into LONG TERM, HIGH RELIABILITY, generation units rather than a bunch of gears on a stick that breaks every two years.