Looks like a pretty compelling use case, particularly for large-scale, commercial implementations where there are hundreds of solar panels. Not sure the cost incurred with a robotic-based system can hold up to smaller or even residential installations, though.
I would think this would be most cost effective on newly installed systems. I would be interested to know how this would work retrofitting existing installations. A 15% increase in output of panels would add a substantial amount of power to even residential systems. To reduce the number of panels needed by 15% may make, even residential systems, more viable. I would be concerned with the security of this system in adverse weather conditions, like how much wind could this system handle?
I would be interested in seeing a comparison of the costs of a robotic system vs. the cost of individual panels. Since regardless of the type of installation you implement, you still need space, I'm wondering if as the panels become a commodity it would be cheaper to jsut add additional panels rather than trying to relocate them over a robotic track?
In many parts of the US this monorail driven robotic positioner might be impractical. Think snow country with icing conditions thrown in for good measure. And even in locations not prone to icing and snow, what happens if debris is blown onto the track or the robot? Another issue is a single point of failure. If the robot fails you lose tracking for your entire array! I also assume the robot has a charging station unless the rail is energized to provide power on demand. So, now add robot battery maintenance to your operating cost.
I think the energy saving issue is overstated. The same total amount of locomotion energy would have to be dissipated to get every array refocused plus, now you have to waste energy on moving the robot from array to array! The only potential saving would be on peak current required as each array would not be simultaneously moving to track the sun. But you can do the same with trackers. If you want to reduce generation efficiency a bit you could have intellligent controllers stagger the moves and also not re-aim continuously but rather do so periodically with longer idle times between adjustments.
I'd like to see the numbers comparing the cost to engineer and install the monorail tracks and positioning adapter couplings versus individual motor drives. Keep in mind the potential savings of mass production of those AZ or AZ/EL drives versus the much lower production rate for the robots. And how much is ultimately saved if you need to utilize multiple robots and tracks to provide reasonable response time for a very large array? More important, I'd like to see MTBF comparisons. I'd think the monorail robot with all of its moving parts and complexity and its duty cycle would represent a higher probability of failure than any individual tracker.
Looking at the picture, since the monorail is below the panels, the ability to cover or enclose the robot is well beyond the ability to protect the panels. Enough snow and ice to immobilize a protected robot and track would probably render the array useless that day so the loss of use of the robot is immaterial. The robot looks pretty substantial so I'm guessing that there's an amount of power/torque/reliability that would be prohibitively expensive on individual panels.
As far as a broken robot goes, since the plans call for two robots, one for a backup, this should not be much of a problem. It doesn't take much engineering to add a clutch that is manually activated to disengage any drive mechanisms from the wheels. The device could be set up so that if another robot pushes from behind, everything disengages automatically. Or a handle thrown by hand that opens the drive mechanism to allow picking up the robot from the track.
Since I'm willing to get on an airplane, I should be willing to accept that they can build this robot to be reliable.
I think the point here is to not have motors on each pan/tilt axis of the panels. You drive the robot to the panel and make the adjustment, so there are no motors on the panels. This reduces the number of motors and drives in the system to no more than 3 - one to move the robot from panel to panel, and two to adjust the pan/tilt axes. A clever mechical design could probably get it done with a single motor and some clutches.
The robot moves the motors to the solar panels, makes the adjustment(s) on that panel, moves to the next panel, makes the adjustments, etc. That is why it takes 40 minutes to do the adjustments on a 300kw installation - it is a round-robin affair. It can only get to each panel so often.
Thanks, ttemple. I was unsure why this was better than existing tracking systems until I read your explanation. I'd really like to see a comparison, in terms of power consumed and cost, of this system versus conventional tracking.
Ttemple, instead of doing the rotation to individual panel, I would like to suggest another method. Fix all the solar panel in a single structure and by using a powerful motor they can rotate the entire structure in a single move. The power for motor rotation can be generated from solar panel itself.
Ttemple, I won't for rotating the entire farm by a single motor. Clubbing some panels in single frame and each frame can be rotating using a powerful motor instead of independent motor for each panel. This may be a cost effective solution when compare with many independent motors.
Until I read your comment, I totally didn't understand this system. The article should've made it more clear how this works as I had first thought the panels were being driven down the track. Then I double-checked on youtube, searching for qbot demos. Yup, the aerodynamic looking device travels along the track to the base of each panel. I imagine there's some patented mechanism involved to lock the panel in position when the robot moves on to the next unit. Perhaps a clutch system that releases automatically as the robot locks into position under a panel. I would expect the monorail itself to include bus bars to connect the output of all the panels. How much additional cost to add a full length tunnel to the track to keep snow from blocking the robot? Wouldn't it be cheaper to use a plastic pipe to run the robot through, with the interface of each panel mount sticking down inside of the tube? Bus bars can be inset along the inner wall of all the tube sections. This will help keep weeds from interfering with the robot, keep the weather off of the bus bars, maybe help reduce corossion along the bus bars so it's not a problem for the robot to always maintain a connection to keep it's battery pack charged. Data could be sent out the bus bars using off the shelf powerline networking to track the performance of each solar panel as well as track the condition of the robot itself.
I can't imagine anythong other than "proof of concept" being built next to a possible obstruction. Solar panel 101. However, driving down country roads does reveal a lot of badly placed panels on small government solar powered projects.
A suitable gear reduction on one motor could drive a whole series of "linked" panels in one axis since speed is not of the essence. The sun is far enough away to be considered a moving point source of illumination so all panels can point to the same place.
Using a traveling robot seems to be streching for a reson to use robotics. On the other hand it looks cute....
rspake: I believe the requirement to use a mobile robot is also based on the fact that the monorail structure can be placed in any location, routed between trees & buildings or even across rooftops. It's a very flexible design for being capable of placement in any possible situation, even one of those badly planned government jobbies you mentioned.
Year-round aiming requires more than just choosing what azimuth is needed every few minutes. The sun follows a path that changes in height every day. It won't matter if a Qbot on a monorail is used or a structure like what you described, the installation itself could be less than perfectly flat & level. So each panel needs a slightly different elevation adjustment as well as tracking the azimuth.
A one motor single axis array will need a very well trussed structure in order to move as a unit without flexing noticably, and will also need a very large circular space cleared of all objects, which means it wont work in a suburban environment or any kind of existing building sites such as in industrial parks or around a hospital for example. No matter how you gear down, you'll still be moving a lot more weight with that one single motor so it'll still need to be a lot more horsepower and consume a lot more power. Plus cost more to purchase.
And did you consider placing panels across the entire area your circular structure will be operating inside of? Maybe you could try floating the thing on pontoons in a circular ditch just to take the weight off of the central spans. And then a much simpler approach is to have an anchor pin to keep the structure from drifting into the sides of the ditch and use propellers mounted all around the outer perimeter to make it rotate.
Your points are good if random, odd, small locations are used.
If your plans were to invest considerable moneys then site location and site prep are paramount. Dips, valleys, trees and buildings can all be corrected. This brings down the complexity of control and ultimately cost of operation.
Two axis cotrol is assumed with two controlling motor inputs. I don't see a lot of success in future use of the super flexable robot setup unless PV panels get a whole lot more efficient than present. Return on investment is too slow.
I am sure PV efficiency will get there sooner than most expect but it may be 5 to 10 years.
I tracked satellites using very large antennas for 20 + years. I do understand what it takes to track and follow moving targets accurately. Those same antennas are still in operation nearly 50 years later but now with electric drive rather than the massive hydraulic drive system they started out with.
Your over the top suggestion (I hope) of propellers for guidance reminded me of a Jules Verne scheme. :)
The robot looks very interesting but time will tell if it is practical.
I can see them as a first step in an even better future design.
Scale describes the problem well. Go big enough and mountains look like check marks on a site plan.
I would be more concerned about how well the mechanical linkage on the Qbot method holds up to repeated interface cycles. It could fail to hold position after aiming and wind blows against the panel, putting torque on whatever it is that's supposed to keep it from moving while the Qbot adjusts all the other panels.
Don't worry, I wasn't really thinking about digging a ditch around my property...
I like the concept, it reminds me of tugboats on a river; a single tugboat serves to help steer many ships.
If this bot can steer 200 panels and bring the output up by 20% to 40% on any given day (30%) per year... and do so at a greatly reduced cost compared to 200 tracking motors and accessories... great idea.
That said, beware of the unintended consequences of leaving a robot to its own devices. I'm not sure what could go wrong in this case but I always assume that robots are dumb as a rock and will find the one thing we've missed and exploit that to create maximum damage.
In the 70s & 80s I worked at a place that was a NASA grant funded solar heated/cooled factory. We had about 1/2 an acre of water filled collectors (you don't hear much about those anymore do you?)
They originally had CC Cameras watching the system but someone decided that they were not needed... The panels could be rotated to face down to keep snow off during heavy storms. Well, someone turned on the rotation motors and walked away assuming the panels would stop at the limit switches... Yeah...
One switch failed and that array pulled on it's 4" diameter hose till it snapped... 10,000 gallons of water and very expensive environmentally safe anti-freeze pumped out. Believe it or not a second leak occurred a couple years later which also emptied the system before they finally unplugged it. I helped dismantle the system a while later.
Okay no robots were involved, but a true story about a very expensive solar panel mishap involving a single limit switch... Like I said, great idea... but expect and prepare for the unexpected.
Well, for some reason S. Jersey has a lot of solar farms going up. So far as I can tell they are all of the fixed panel type. I often imagine that they could improve output by going with tracking, but perhaps this is a case where simplicity has the advantage of being easiest to implement and maintain.
Not all solar panels have the same useable angle that they can be off by. The cheaper solution is to pick a wide angle panel and skip the tracking gear. But a panel with a narrower alignment could easily make more power for the same amount of light.
There's a company I saw years ago that made a tracking system that used a pair of reservoirs, one each mounted at opposite sides of a panel. Both cannisters were connected with a single tube. The system was filled with a fluid that turned to a vapor at very low temperature. Some sheet metal shielded each can so that when the panel was tipped to the east, sunlight would be able to warm that side but not the western cannister. The fluid would expand and travel out of that side to the other where it recondensed. At a certain point, the entire deal gets weighted far enough to tip over so it's facing west. Using a solar panel that makes the most of only half of the sky finished the design. In the morning when the sun starts to rise, the western side gets warmed up, the fluid travels back to the eastern side and the panel tips back to face east.
PROCAD, a leader in 3D and 2D piping design software, released its newest product PapriCAD 3D. Built on a scalable data-centric platform, PapriCAD® 3D offers a comprehensive solution that delivers 3D modeling, component design, material reports plus AutoCAD. Quickly transform ideas into data rich 3D piping installations with minimal training, accelerated productivity and lowered costs.http://bit.ly/NmdQGu
Elizabeth, I had seen a similar system in a solar farm. Al the panel structure is attached to a small motor, which will rotate the entire panel structure in synchronize with the position of sun. This will help to fall the sun rays directly over the panel and hence a better efficiency.
This picture may not represent the article well. Look at where the shadows are vs. the direction of the panels. I am all for Solar but there are too many magical mirrors and fuzzy government math claims to convince me it is even close to being economical. Why doesn't someone print the actual facts and figures?
Hi, I actually had a solar panel system installed 2 months ago. I don't have the RS485 data link on the inverter wired to my computer yet, but I do go out occasionally and select through the menu system to see what I'm getting for my money. This is what happened today 19/09/12 at 10:00am with panels at a fixed inclination of around 30 degrees and facing almost north. and a totally overcast day with that even grey cloud that blankets the sky such that you don't see the puffs of the cloud. The density was such that it was just shy of raining. I have about 4kW worth of panels and about 1/4 is shaded by a neighbour's tree that is soon to be shortened.
Our location is:
Latitude: 37°48′49″S Longitude: 144°57′47″E
Now this was 4 hours after sunrise and that would have been increasing later in the day as the sun moved higher and continued to about 2:00pm (4 hours before sunset) so that's at least 6kWhrs while still (barely) in Winter (southern hemisphere) and there would have been more before and after these times as I have seen about 1kW at 9:00am on a similar day about 1 month ago.
So seriously, don't belittle what solar can do, we have been quoted a daily average of around 14kWhr per day averaged out over the year for the system, and with the neighbour's tree lopped I see it truely happening. Once I have the data logger connected I will put some graphs up on the web.
We had solar panels installed this year. We have 44 panels, 20 south, 24 west facing. Our system was rated at about 9500 kWh per year. Last year our consumption was around 15,000 kWh, which puts the system at generating nominally about 63% of our usage. We have a grid tie system, which means the extra energy we generate gets pushed back onto the grid, and runs our "net" meter backwards.
In 8.25 months (the system went live 1/12/2012), we have generated 8734 kWh, and our net usage is 1350 kWh. That puts us at generating ~86% of our usage. Last year, in July alone, we used about 2100 kWh. We live in Colorado, which is one of the better places to live for generating solar, for sure.
We decided to lease the system for 20 years, which is what made it cost effective. All maintenance is covered under the lease, so if a panel goes bad, an inverter goes out, or other problems it's all covered. The lifetime of a system is about 25 years. However, we don't get any rebates or credits - these go to our installer, Sungevity. But, we pre-paid our lease, which gave us a significant discount.
We figured that it would take us 12 years to re-coup our investment, but that does not take into account energy prices (Xcel has already announced a 5% increase). If we are able to maintain our energy production at 75%, our re-coup drops to just over 9 years, which also does not account for price increases. Also, we have a 2 tiered system in CO, so from 6/1-9/30, any usage over 500 kWh in a month costs double (from 4.5 to 9 cents) - that makes things a little more complicated to figure out the numbers, but actually it speeds up our re-coup, because during the summer months, 2/3 of our spending was at the higher rate. I estimate it will probably take us about 7-8 years to re-coup our investment.
From March through August, our electric bill was a big fat 0 (except for the $6.75 facilities fees). That has been very satisfying. Last year our July bill was $340. This year it was $6.75.
Grid tie systems are a lot more economical - battery systems are extremely expensive, and almost double the cost of the system, plus the batteries rarely last more than 5 years, so maintenance costs quickly eat up your savings.
I hope this helps, and let me know if you need any more info.
Will the robot use the same alignment data for all panels or can some panels have individual settings? For example: if, over time, a tree on an adjacent lot blocks a panel at the edge of the array, can the robot select an angle that avoids pointing at the tree? Or if with wear & tear, the mechanisms get out of alignment, can the robot use specific data for that panel or get feedback from the panel as it's moved? Can the robot be programmed to move other objects that are mounted in a similiar fashion, such as reflectors used to increase the light falling on panels?
This seems to be one of those "why didn't I think if this" ideas. It's use of existing technology. It avoids a whole host of problems or issues with other systems. For example, individual motors on all panels driven by the output of the panel would fail as soon as it got dark enough to not power the motors. A mechanical linkage of all panels would allow a single panel's mechanical problems to affect the whole array.
Once the monorail system is built, it could have additional functionality. For example, a robot with a water resevoir and pump can wash down solar panels. It could even carry parts (or food) to someone working on the solar array.
Inflated overstatements, and half truths make things easier to shove down the throats of possible customers.
This is just another version of the old "better mousetrap" paradigm.
Here is when the law of "Dimishing returns" hits the road.
What king of "huge energy savigns" are we talking about, when the positioner moves the panels a few degrees every 40 minutes ? what is the return on the investment on a -more expensive- (and supposedly better) tracking system?
I think that this kind of article is just a bunch of glittery "hype" and of course, EDN has to make some money to keep feeding the payroll... so I think it's ok to publish this kind of "promotional" article. but please don't take us "engineers" as fools.
We will always question the purpose, usefulness and every kind of claim made about what is presented to us, with shine and glitter. We as engineers represent the most "critical thinking" part of society in all referent to our fields of expertise and then some.
Under present designs, there are already low function robots on every panel, they are considerably >underutilized< and >massively< redundant! It makes sense to me to consider replacing that capital (expensive, life-limited wear items) with a concentration of robustness and intelligence in a few mobile units.
Suppose you have a robot built to wash the panels and perform routine maintenance and repairs, including damaged panel replacement. Such a robot is practical and necessary. It would be relatively trivial for it to also perform the function of the QBotix monorail. It would not necessarily require rails; although rails increase efficient, long balanced arms and/or gantries on a wheeled maintenance vehicle might have similar returns. The panels could have passive breakovers that let the panels lay over in severe wind gusts and be reset by the robots.
I think there is no reason to suppose that the fault tolerance of the system must be realized in a single robot, nor even that multiple-redundant robots are needed to achieve some consensus as to which ones are "sane". Any supervisory functions that manage the solar farm would also monitor the correctness of each of the tactically redundant robots.
As to the question of the relative production volume of the robot, I suggest that mass-produced robotic fruit and vegetable pickers under development (e.g., Vision Robotics) could be downrated and repurposed to the relatively simple tasks of solar panel washing and tilting.
Once the "solbot" (Sol, as in Sun) breaks, the whole system is useless. They claim reliable, but put some bugs or animals in its way, on the track, and then what?
I like how they combated the old concept of having each panel move with its own motor. Each panel is then positioned optimally. This appears to be cheaper. I wonder how a typical day finds the Solbot. Is it able to power itself? Without the bot, how much energy is saved? And how would the same array fair with just moving all panels at the same time like puppets compare?
Great post. I agree, it seems to be a "better mousetrap". Then again, isn't that what we do, design and build better mouse traps. This is an intriguing system and obviously one that has gained some traction and much interest. Years ago, I was part of a team that developed a solar water heating system, significantly less complex than the one shown in the article. It did work and we were able to market over 250 during the four years we sold the unit. The biggest problem—it was not cheap. Quite expensive for a single family dwelling and not effective enough for commercial use. I do think devices of the type shown improve the overall technology and certainly improve marketability.
Fifty-six-year-old Pasquale Russo has been doing metalwork for more than 30 years in a tiny southern Italy village. Many craftsmen like him brought with them fabrication skills when they came from the Old World to America.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.