Solar Design Lab Text

Welcome to Solar Design Lab.

For class videos, click here.

This class will explore the areas where computer-assisted design improves the accuracy of performance estimating, array layout, project documentation, all the way through permitting. Follow along to complete solar designs on your own, with input from qualified personnel as needed.

Let’s start by going through the steps of what’s needed to prepare a residential solar project for permitting. There’s a variety of software available to speed up this process and one of my favorites for small residential projects is SolarDesignTool.  

A good solar project should not be installed in an area that has significant shading. Partial shading is resolved through module level panel electronics like micro-inverters or DC optimizers. So many solar designs do not need an accurate shade analysis because the rooftops are in fact unshaded from early morning to late evening. So without much ado, let’s assume to already know where  a solar array will be installed and attempt to complete the permit documentation. Within Solardesigntool, click new project and click go to create the site. It is already keyed into the the same database as Pvwatts to develop its its unshaded energy estimate.

Already the software is asking for advanced project details which is why I love this software so much. Although limited to small residential projects, its unique focus on creating final project documentation ready to be dropped off at the local permit office makes it a worthwhile endeavor to sign up for its free trial and experiment with a complete design yourself, to discover useful site survey information. 

For example, on the interconnection application, it will commonly want an electric meter number and electric service information. This kind of project information can only be determined

through on-site inspection but a site owner can typically take photos of this information to be forwarded to a trained eye. 

Solardesigntool has many different configuration options and if it does not have the exact configuration you need, it can create a CAD that is close enough to a complete document that it can be easily modified to appease a local permit office in an external CAD software

such as AutoCAD or DraftSight 

If you don’t know some of these terms, such as the difference between a top fed service panel, a center fed service panel, or a main lug only service panel, as opposed to having a main service panel breaker, you might not be competent enough to perform this level of solar design detail (at which point you should consult with an electrician to complete these steps).

Now other project information is inputted, some which is technical. This can be omitted but there is no reason why this level of detail cannot be established at the front end of the project.  Then pulling all the information together, such as your local authority having jurisdiction, site conditions which were considered in the design, wind speed, surrounding terrain, and other information will then be called out in the final permit documents themselves.

If you do not know certain information, you might find yourself checking back with manufacturer provided sizing tools, such as racking or inverter sizing software, but this sort of project information should be known ahead of the project. 

For the next step, it’s asking us to draw out the rooftop.  There’s a few options to do this ranging from simply inputting the dimensions of the installed area, to tracing over an aerial image overhead map, to actually buying a mapping service to do the heavy lifting for you, so depending on what you’re already purchasing you may already have a 3d model. 

The most common starting point for solar design software is to develop the array layout by tracing over a satellite image.

Later we will explore some other options where images found on Google Maps may be insufficiently clear to design based on a remote analysis.

But at any rate, here’s the job site and it’s immediately apparent that there is a large southern oriented unshaded roof surface that is an ideal roof surface for a solar project although it might be worthwhile to consider solar on the west facing roof surface and even on the north facing roof surface depending on the roof tilt.

Other kinds of solar design software are better at determining the roof tilt than SolarDesignTool so toggling the Street View and the tilt view can provide enough visual detail to attempt to guess at the roof tilt, as well as identify other rooftop objects.

 When complete the orientation of the roof is specified.

We are then prompted to enter the roof pitch. The process is then repeated for all roof surfaces.

How accurate do these roof areas need to be? The answer is it depends. Many jurisdictions require a three-foot offset from the roof ridges and rakes, which are the sides and the tops of the roof. Both for structural, accessibility, and easy of design reasons, allowing a little more space, such as a four foot offset, would be even better, to give the installer a greater  margin of error by reducing the need for careful site layouts.

Roofs with a substantial amount of roof penetrations such as plumbing vents may require more accurate planning within the design software. Regardless of the software used, the overall design procedure is the same. First define the usable roof areas and then add on any obstructions second.

So here we’ve mapped out different obstructions on the rooftop with these pre-existing holes in the roof:  chimneys, plumbing vents, and flashed-together valleys are pre-existing holes in the roof which become common leaking hazards as the solar array is being built and after, as  workers are stepping around these locations, which have become walkways due to the newly limited space on the roof, a risk that is increased with particularly with a compact array design.

At any rate the areas around a chimney, plumbing vent, or roof cap might need to be repaired at some point in the future so it’s best to allow for some clearance around these obstacles even when the clearance is not required. There are exceptions. An ugly solar array does not add as much value to a home as a pretty solar array. So if the array’s look is significantly improved by putting a solar panel a little bit closer to that chimney, it may be acceptable so long as the client understands the design decision. In any event, the clearance space around the chimney is more important than the clearance space around a shorter, smaller plumbing vent but  staying clear of all of these obstacles kind of removes the need for shading nuance.

With that in mind, it’s time to locate the equipment. SolarDesignTool is asking the designer to locate the main service panel utility meter, as well as any sub panel or disconnect switches. Google Street View is your friend here as it tells you you know where the meter location is without having a step foot on the site. Depending on the client relationship, this information can be picked up at a site evaluation or more simply by chatting with the client over the phone. 

Here SolarDesignTool is asking for the building’s least horizontal dimension, useful for engineers in determining wind load. Wind load is typically resolved by over-engineering the racking or offsetting the array from the roof edges. Google Earth with its built-in ruler tool can be useful in quickly measuring this dimension. In this case the building measures 50 feet wide.

It asks some roofing and rafter questions, again which is information to be supplied during a site visit. 

I find the rafter space is important for finalizing the racking roof attachment locations, and we will have more on that later.

Anyway the roof is sketched and we are presented with three defined roof areas to lay modules on to.

This menu is similar to using a manufacturer stream sizing tool combined with an array layout generator. 

It’s not as easy to use as other solar design software so sometimes the layout needs to be pre-determined, like when a detailed shade analysis is needed, before  remodeling the design within SolarDesignTool.

After selecting the module and the inverter, the designer is presented with various allowable circuit configurations. This is a micro-inverter design and the micro-inverters per their specification sheets are limited to 17 modules per circuit for this particular module on this particular inverter. It gives us different branch circuit configurations. The north roof gets 17 module circuits In very long even circuits. The south roof circuits are divided up to make the installers job easy.

So in this example a 320 watt black-on-black 60 cell module is combined with a DC optimizer inverter system. Taking a quick look at my array layout see that 31 modules fit on the south side of the rooftop in portrait.

I’m a big fan of ordering solar panels by the pallet to optimize supply chain logistics and similarly, portrait mounting makes quick and easy work on a shingle roof because of how the rafters and purlins are laid out. On a metal roof with east-west battens or purlins, it can be the opposite way. Australia has a lot of metal roofs and they commonly use a landscape orientation which can fit the roof a bit better. In the  United States we have a lot of shingle roofs with north-south rafters and so we do a lot of portrait orientation. In some cases landscape can be the better choice, for aesthetic or structural reasons.

For this particular solar panel looking at the specification sheet, one pallet of panels comes with 25 solar panels so if there is only budget for one pallet of solar panels, then it can easily fit on this roof regardless of whether it’s in a portrait or landscape orientation.

Here’s an example that would be a very easy to install and not too ugly of an install where 25 solar panels fit on the south side of the rooftop. If I’m going for low budget, fast, and easy with a one pallet installation, this is probably the array layout I would end up with.

In landscape, 33 solar modules fit on the south roof. 

32 modules fit in a different layout style,  and so if I wanted a larger system, I could fit a two pallet solar design between the south and west side of the roof in a compact, aesthetic manner although it would be a more difficult installation and have some shade from the chimney.

An installer in an active market may offer this design to show a best fit, space efficient system but in a cheaper market, another installer might go for an easier solution to lower cost, such as one that covers all three large surfaces of the rooftop in an easy installation layout. The empty space allows for minor adjustments as well as  aesthetic considerations and maintains the portrait orientation for ease of installation. It’s not just about how many solar modules can fit on the rooftop.

In the same screen we get the array layout, we also have the string sizing built in. This is just like a manufacturer or a string sizing tool that you would find on a manufacturer specific website. For example, in Fronius’s string sizing tool, picking out the inverter and solar module will give a variety of circuit options. We could have two strings of eleven or two strings of eight. The designer would have to choose which circuit configuration would make the most amount of sense.

Well in SolarDesignTool, that string sizing capability is built right into the software. Here we’ve selected a solar panel and a particular  microinverter and this particular micro inverter has a maximum number of micro-inverters per circuit at 17. Right here in the design software the micro inverters capabilities are limited to that number. 

SolarDesignTool has a plethora of detailed design documentation, more than any other software right out of the box, which is pretty impressive given that it’s on the low end of price range, making it a great value for installers who primarily install small residential projects. For larger projects, I’ll model smaller iterations and then kind of stitch them together in CAD afterwards. 

Once the design is put together, you can modify things like wire size, supply verses  load side connection. It’s nice to see on the design detail whether or not there’s a production meter interconnection meter. 

Here we are modeling a sub panel with the micro inverters all landing on different branch circuits and then tying into a solar production meter going through a main disconnect breaker, before heading out to either the utility main service panel or the utility meter, because it’s been  denoted as a supply-side connection.

So these one-line diagrams do get automatically generated at some point in the chain. There are some imperfections, such as when the tool does not model exactly where the rooftop transition is made. 

Here’s a popular rooftop transition box that’s for transitioning the cable to go down into the attic. Being able to interrupt these home runs and say no, right here is where I want the junction box to transition into the interior of the attic, sometimes that level of detail needs to be kind of redrawn in CAD after the model is complete.

 The exact location of the subpanel where the circuit is going to be run, how long those home runs are, can be adjusted by hand such that SolarDesignTool will connect the dots together to reveal how  many feet of cable and conduit are required, what the exact voltage drop is of each circuit, and other details. 

Here’s an example of a conduit schedule where the fine print tells the combination of three-quarter inch and half inch conduit. A designer might go back into this software and say it’s all gonna be in three quarter inch, and so the half inch will be upsized, which will then change the conduit fill  calculations. 

A quick glance at the balance of system material can reveal some ability to simplify the design and installation. In this case, it is revealed that a change from #12 AWG to #10 AWG will both simplify the balance of system material and reduce the voltage drop.

This is a one long diagram created as a result of the survey, processed completely by computers with no hand drawing. It’s not like the designer fills out a survey, uploads it to an overnight design team, who squirrels away to hand draw the detail in CAD. This is all computer generated and exportable. 

SolarDesignTool is intended to allow you to print out and staple the permit package to show the permit office the design has been duly considered. I think it’s main limitation is that it’s only good for small residential, not having the capabilities to generate line diagrams and reports for large projects.

2. Shade Analysis with Helioscope and Aurora

So the other drawback of SolarDesignTool is that it does not do shade analysis. Trees are tricky and not being able to model shade accurately because of nearby trees will result in problems. We used to get up on the rooftop with a survey tool called a Solar Pathfinder and it would tell us where the shadows would fall on the roof. 

The shadows would correlate to a chart which could be converted into a percentage of performance loss on the year.  Now, a detailed shade analysis can be done within a computer, for every solar panel on the jobsite, at high resolution, and be as accurate as going to site to perform a manual shade analysis. I actually think these computer models are more accurate, because  trees grow over time and so modeling the the tree growth is better done in a 3d environment. 

Helioscope is the most popular solar design software on the market. Actually I don’t know those numbers for sure, but that is the feel I get on various solar discussion boards. They’ve been

on the market for as long as SolarDesignTool but have been fighting tooth and nail to be the industry standard shade model ever since. This has resulted in a very rapid

Modeling interface that is primarily done in a 2D environment, but is actually a 3D shading map which in most cases includes nearby tree topography called LIDAR. 

Aurora Solar is their primary competitor, which is more feature rich and more expensive. Both of these softwares show the advantage of 3D shading modeling. You can draw some interesting design conclusions with a good 3D model. Let’s imagine this customer did not want to chop down this tree. This image shows that with the tree, it doesn’t really matter if the modules were facing north or facing south. The panels closest to that tree are the ones that are less productive. Whereas these southern facing solar panels will  produce less energy than

these northern facing solar panels because they’re shaded by that tree for a substantial portion of the day. So if the homeowner refuses to chop down that tree then the best solar array would be the one farthest away from the tree in order to get the most amount of production from that roof. 

Helioscope calculates the wire sizing and voltage drop calculations. The generated documents are not quite as detailed as other software more focused on permit documentation. Helioscope has focused on preliminary design and kind of assumes detailed design will come elsewhere later. That means you do shade assessment in Helioscope and then once you know your final array layout, you then replicate it in the permitting software. Helioscope will directly export its designs to EnergyToolBase.

Here is a tool Aurora makes, which attempts to guess the home load profile, which is how the  the customer uses their electricity throughout the day. This is only a guess based on monthly monthly consumption data, which is divided into a 24 hour time frame and then adjusted based on a survey of what electrical devices are in the home.

Start at 1:20 and then go back and comment how Aurora Solar does hourly consumption data inputs…

Aurora assumes that if the air conditioner is used, such that it will increase as the day gets hot. Here is electric heat modeled as part of winter load, which is region dependent. Swimming pools or all-electric home can cause the load profile to jump around wildly. This load profile data can be useful in economic analysis and we will later explore a software exclusively focused on economic modeling. 

But since we are discussing Aurora, let’s see how they develop a shade model on a residential site. Aurora has nice templates for modeling complicated rooftops quickly. Clicking the edges can auto-detect the remaining roof shape. If Aurora does not have LIDAR available in your region, then the model is simply built to match street views and aerial views. But if you do have LIDAR, items like measured the roof slope are less important as that information can now be picked up remotely with the model fitting the LIDAR data. 

Difficult 3D model at 45:00 through to building plans to 54:00

Aurora has  different shapes of trees and makes it easy to model obstructions in 3D. Now now they’re doing a shade analysis and the south side works well with the north side being the least sunny. We can note just how large the area around the chimney is to know to stay out of the area if using that East side of the roof.  So if I want to design the array to avoid broken up subarray sections, I might stick to the south and west side of the roof.

Now Aurora is selecting the module and inverter to do a string sizing calculation.  Because it is building a true 3d model, it can calculate performance and voltage drop for every single panel on the roof. It is a very robust design  engine which can handle difficult and complicated roof modeling to an exact degree. 

So even on inverted folds on the roof, modeling a seamlessly transitioning from one layer to the next is ok. So here they are breaking the roof into sections. They have a smart roof system which allows the designer to select roof features to model, so here is the feature for this inverted fold roof style and it applies it to the roof shape. So assuming the designer is practiced, even a very complicated rooftop, adjusted ever so slightly to make sure the model is preciselyvmatched up, with a little awning built on the back porch, can be readily designed.

The porch is a different shape than the previous roof, and so the model is quickly readjusted. Now we have a very detailed 3d model of the roof. From that point within Aurora you start by defining your field arrays, defining obstructions, performing string sizing to eventuall end up with an array layout, performance estimate, and one-line diagram.

Making solar designs based on remote roof imagery is not  always as easy as it sounds. Bad image quality leads to guesswork. Helioscope has Nearmaps built into it to get better resolution than then when using Google Maps images. I’ll use Microsoft’s Bing maps instead if I need alternate imaging, which is often better than Google’s images. But Nearmaps is a higher resolution paid service with the best images. Even when nearmaps is a software feature, it is not always included in the software subscription price. 

5 Minute Residential Design in Helioscope

They’re defining the array layout and considering landscape verses portrait racking including staggered layouts. It it looks pretty good although I am a fan of making the array as rectangular as possible, particularly when the roof surface is asymmetric to begin with.

In Aurora you build the 2d model quickly and tweak it within the 3d view using LIDAR data. Helioscope also has lidar data but draws more heavily on an overhead drawing view even though it is actually a 3D interface. 

Here is a shade analysis and we find out that the southeast side doesn’t actually get much sunshine due to the trees, whereas on the left-hand side the shade is less important.

They’re locating the inverters and finalizing the design.  Even though they are doing very detailed performance estimating by calculating the lengths of home run conductors for voltage drops consideration, both Helioscope and Aurora stop before becoming full-fledged construction document generators.

That can get a little annoying as a contractor on a commercial project where the only design work has been in Helioscope, as the true nuance of running home run conductors, say in conduit verses cable tray, are not picked up in Helioscope by itself.

Again, Helioscope is more focused on being an accurate shade analysis tool, whereas Aurora is trying to build something more comprehensive. But Aurora’s finesse takes more time than Helioscope. That can result in better images to impress the customer, but Helioscope’s response is to say that rapid modeling is a better approach, especially on commercial design where tree shading is less important.

5 Minute Commercial Design in Helioscope

Now they’re modeling clearances around electrical equipment. Stat six feet off the edges of a flat roof and at least four feet around electrical equipment. Slight adjustments are made to give the array a bit of a haircut, removing four modules that aren’t really needed and would make the install a little bit easier. 

Both Aurora and Helioscope generate single line diagrams, which have detail but are still basic enough to be intended to export it into your own CAD file with your own gingerbread. While they do voltage drop and wire sizing, they do not want to be the company driving the final electrical design on the project, leaving that task to the designer.

Good through 2:10 – 4:00 Previewing Sketchup w/Skelion Plugin

For shade analysis, what if you’re already using the popular architecture software Sketchup? So many firms out there have a Sketchup Pro. If you’re already paying for Sketchup Pro, then shade analysis can be performed within it. This plug-in called Skelion calculates the impact of that shadow of the chimney on the surrounding solar modules. These are the  unshaded modules showing 100% and then one closest to that chimney is showing 96%,

Skelion has a nice process for laying solar panels onto a rooftop and then and then evaluating the shade. That’s probably a good enough energy estimate for module-level panel electronic circuits. So if already doing projects in Sketchup, the ability to quickly model solar arrays on jobsites not only looks interesting but to some extent provides the value that 3D imaging software such as Aurora or Helioscope charge substantially more to provide. In this image, you can see how quickly energy performance of rural fields can be modeled into solar farms. 

Just like we’ve seen with Helioscope and Aurora, it starts out with an overhead image, this time using Sketchup tools to draw out the house. Perhaps Sketchup knows a thing or two about rapidly developing 3d models. 

Here’s some roof data used to add solar panels to the 3D models. Custom components can be added in. Setting the roof pitch and row alignment is simple enough. Some of the lower end  design software can be problematic in determining how granular the calculations are but here they really are analyzing the shade on each panel on each subarray. 

Start at 1:27 – 4:46 Building a 3D model

Here they are coloring sections of the array for the model, which are separated out when the production is calculated so you can actually differentiate between like the west surface, east surface, and north surface to get production values for all those different surfaces, which is something you can’t do in pvwatts by itself without manually doing the model for each different array, so that’s kind of neat. 

What they’re showing is the pvwatts interface having the standard system loss number of 14% and shading is only 3% of that number, so if Skelion calculates the total shade loss percentage on the subarray array, it then sends that average shade loss to PVWatts to replace it’s preexisting 3% number. So PVWatts will be adjusted accordingly.

0:30-6:45 Plugging into PVWatts

Here’s another thing from Skelion that I was very impressed with.  Here they are importing the local terrain into Sketchup which is cool. The slope of that mountain is now factored into the  solar production estimate. So even faraway mountains can be considered in solar performance estimation.

Great through 2:30 to understand impacts of mountains!

Intro to PVComplete

Let’s watch another one. This software that I only recently evaluated and it was very impressive. PVComplete is comprised of two component software. One is like Helioscope for overhead sketching, whereas the other is a CAD environment for detailed design. The downside of PVComplete is it does not have LIDAR data unlike Helioscope or Aurora. In other words the only way to give things height is to actually program it in.

For generating project permit documentation in house, it’s hard to beat PVComplete. It’s a plugin for AutoCAD but if you don’t have an AutoCAD license, it provides the AutoCAD work environment at not additional cost. A true CAD environment means the various project details are captured in layers which have other spreadsheet-friendly functions. They’ll have a layer for their wiring diagram and then another layer that is actually the wire path between the panels for determining material lists.

Why don’t we fast forward a little bit. What is showing here is different versions of the jobs, perhaps with different inverter or racking manufacturers. Here they’re stringing the array and go for a “snakes in the basket” approach. Their auto-stringing option is a bit chaotic. I’m a fan of manually stringing arrays anyway, and they have an easy interface for it that will show where the circuits both start and stop. Then PVComplete works with racking manufacturers to count racking balance of system material such as attachment footings and conduit fittings, which are not user inputted but rather generated from the line diagram itself and then exported in a way that is kind of procurement friendly.

Now we see their single line diagram tool, which is very easy to make changes within. For example if a solar production meter was originally not planned but needs to be added in later, or alternately removed, it is easy to click and change. Here they are alternating between different kinds of internal vs. external disconnects and overcurrent protection. 

The actual equipment models is not manufacturer specific, but instead features numerous options that are similar amongst various manufacturers. This flexibility actually makes it easier to design based on the real features of the exact component you are specifying. 

If the software lacks a template of a product that your company commonly installs, it is easy enough to save it into their AutoCAD environment. 

Here they show voltage drop calculations and other National Electric Code detail. This is similar to SolarDesignTool but what is particularly neat about PVCompletes is that the actual calculation can be selected and easily added into the one line diagram. So if you have a calculation or feature that you know the inspector will be looking for, you can put it right into the single line diagram  or project documentation. 

The benefit of going this in a CAD environment is that the final project documentation includes a circuit diagrams that tell where the circuit starts and where the circuit stops. In other words electrical details that show what is actually being built and installed on site rather than leaving it for the installer to guess. If instead, the solar company opts to generate line diagrams through a third part service, if they are not asking this level of detail, they will not be able to deliver this level of detail.

From beginning – 4:20 + 7:30 -for versioning, 10:00 for stringing, 14:00 for excel file 17:45 for one line diagram adjustments, up until about 21:00

There’s nothing better when doing commercial solar sales than getting the CFO to smile and presenting a CFO with a clear and accurate depiction of the true project economics is the first part of the conversation. The second part is delivering projects within acceptable parameters of a good economic deal. 

In college I would take engineering economics classes to be trained to produce cash flow diagrams like this. It’s evident that EnergyToolBase knows what they’re doing. It’s a very accurate economic modeler particularly useful for commercial. The same level of detail that Aurora brings to its 3d modeling, EnergyToolBase brings to economic modeling and visualization of site energy use. 

The best way to use  EnergyToolBase is to have 15 minute usage data of facility available and most commercial businesses should be

able to get this data from their utility. Some residential sites have it as well. 

What we’re seeing here is a day where here dark blue the building’s energy load is without changes, and then we have a solar array that turns on in the middle of the day and so

the new building load is thus. Instead of a big tall midday peak, the facility now has a little morning peak and a little afternoon peak. But the solar array has reduced the building’s

peak demand throughout the day and so that’s good demand management at times when electricity is expensive.

But most solar installers who have looked at commercial solar know that the true demand

savings for this day don’t matter if the very next day a partly cloudy day keeps the building demand high when the solar array is turned off because of an inconvenient cloud in the sky shading the array.  If that’s the case, it could eliminate the demand savings for the entire month because demand savings are calculated not on a day-to-day basis but what is the maximum.

The maximum demand charge can be up to 80% of a commercial electric bill in extreme cases, so EnergyToolBase takes the interval data from the building and adds in solar production from PV Watts as well as battery capabilities to accurately model cost savings. So this is a partly cloudy day when the building load would normally be high but the solar array is turned off and a battery is running to keep the building consumption flat.

Because of the battery, commercial solar becomes economically viable and can be the best payback in a given region, whereas in the past we have not had batteries and so commercial solar cost savings was a random mess with a very little economic return. Now solar

sitting on top of a commercial building can provide enough demand electricity to provide cost savings during the day and run a battery with just a small amount of storage that is perhaps 3-4% of the facility’s daily energy use, but we’re not trying to eliminate the entire electric bill of the building, but rather to just shave off the peak in a cost-effective manner.

It’s interesting that even without a large budget, just focusing in on the very peak can be all the more cost-effective. So the same solar array with a battery for a luxury residential whole house backup with a long payback can be placed on top of a commercial facility and because of the demand rate structure, the payback year could be in single digits. The most

cost-effective mainstream payback in solar right now is in the commercial sector.

EnergyToolBase even has my little electric cooperative in Mississippi modeled within their rate database. What they have found is that despite online tools out there to comb through various utility rate structures, the most accurate way to model rates is to do it the hard way by opening up every rate structure in the USA and hard-coding it into a database.  It’s not a perfect process but a company whose job it is to keep track of that stuff is going do it a little bit better than

what you can do on your own. 

I recently moved apartments although my electric cooperative does not have a very good solar policy they do have a pretty interesting time of day rate structure. In the summer time between 3 p.m. and 6 p.m. Monday through Friday, I have a peak rate and an off-peak for all

other times. In the winter, peak is only a  two hour period in the morning, again on weekdays only. Without time of day metering I get a flat rate of nine cents per  kilowatt hour, and with time-of-day during the peak period the electric rate quadruples to 36 cents, but during off-peak periods the electric rate is cut in half down to 4.5 cents per kilowatt hour. 

So three hours a day, five days per week is nine percent of the year. It’s called peak for a reason, so let’s fudge the numbers a bit and assume that 10 percent of my energy use comes during this peak time. Increasing that rate by 400% and then reducing the rate by 50% for the remaining for 90 percent of my usage will decrease my rate by 5%. 

This shift could come about by having a solar array on the roof, as well as a battery, or simply by using home automation to regulate the thermostat and other continuous loads such as dehumidyers, refrigerators, water heaters, ceiling fans and more.

So even though my cooperative does not have great solar policy, I can get better value out of a small renewable system by moving to time-of-use and designing a system around it. Or I could save money without doing a solar project at all by load shifting to reduce electricity use during peak times. If I go further into my cost savings, by turning off my thermostat completely as an example, my electric savings could be 30% or more.Sometimes these time of day rate structures can be particularly advantageous, meaning the system doesn’t need to generate or store the entire facility’s electricity to be optimized for the best rate structure, and to know the return on investment for various system configurations, EnergyToolBase is the fastest way to model the savings. 

When modeling energy systems, it’s useful to see what is going on to develop a narrative to explain system behaviour, which can help with fine tuning the design for maximum cost-effectiveness. On this day, the battery is not being fully drained down to zero, but on the next day it is draining further. The difference between the first day and the second is a larger midday peak use. A larger battery would not just shave this peak, but start generating baseload, which could indicate an oversized battery for demand rate structures. A downsized 

Battery would cost less and perhaps be the optimal system for cost-effectiveness.  

EnergyToolBase has some optimization algorithms built in to steer the designer towards the best system. Here a time-of-use rate is modeled for during the week and then a separate rate structure for weekends, reflecting my specific rate structure. 

Now I can see how the economics would work, with or without solar, with or without batteries. I plug in my monthly consumption data because I do not have interval data, sometimes called green button data. 

All I have is monthly consumption figures from my power company, but that’s it.

Because I lack 15 minute interval data, I have to tell  EnergyToolBase how much is peak verses off-peak, so I modeled 10 percent of my electricity coming during peak times. 

Now I can select between the flat rate and time-of-use rate to be more accurate with my projected savings.

Even before the final report,  EnergyToolBase provides some economic figures to guide the design along.  I found is that by doing relatively little I could save about three hundred dollars a year off my electric bill by switching to the time-of-use metering. If I kicked up my digital controls, I could add another two hundred dollars of savings to my bill, even without solar which would push the savings further. A right sized battery would only need to be store my peak energy use, anything larger than that would have more backup power capability but be less economic on the day-to-day.

So that’s that’s kind of interesting. The customer doesn’t really care where the savings come from, so perhaps the solar design could be more cost-effective on a different rate structure. The customer may be happy to see the lower electric bill, and not particularly care where the savings are actually coming from. 

Helioscope is integrated right into EnergyToolBase, which would otherwise take its production figures from an uploaded spreadsheet or directly from PVWatts without shade analysis. Helioscope is good at making those kinds of strategic alliances.  

At any rate, we’re putting in the array size cost, as well as the tilt angle and orientation. Here’s our pvwatts info after the solar array is specified. 

There is a separate battery component which is really where EnergyToolBase becomes a design tool rather than a simple economic modeler.

Other software isn’t even close to nailing the optimal battery size for a facility, for a variety of different battery case uses, and the transparency that EnergyToolBase brings to its modeler can help you narrow in on the best option for your case use real fast. 

We don’t have time to discuss battery nuances, other than to assume you will want to use some form of lithium ion (used to top-shelf) for your project. This tells us the larger the system gets the lower the savings, as a ratio to system size. Starting out, a smaller project will be more cost-effective. For my time of use rate, I need a battery which cycles almost every day, with just enough power to cover my peak use.

A large battery being less cost effective makes sense in this case, as a larger battery which stores more than peak load would not be used as efficiently. Electrical storage is expensive. Storing only peak energy is more economic than also storing off-peak energy.  Starting the project off with a small, cost-effective battery might be better for everyone than selling a larger, less cost-effective battery.

I was interested in exploring how the model was actually handing my time-of-use savings. Here is the solar array, with the battery only being charged by the solar array. So the battery is

being charged only when the sun is up. 

That’s a mode of operation for the battery.  If you charge the battery only with solar, then it qualifies it for the 26% solar tax credit. The solar array is oversized, resulting in plenty of outflow onto the electric grid. In a situation where the outflow onto the grid is bought back at a very low rate, then there is little need for a solar array to be so large. An array even half this size would still produce outflow. An array one-third this size would have no outflow, produce enough power to charge a time-of-use battery, and reduce the home load during the day. One way to make the project more cost-effective when the utility buyback rate is bad is to reduce the size of the project, exploring other options such as batteries or digital load control.

Here we’re programming the battery model to target energy arbitrage and rather than demand savings. Energy is measured in kilowatt hours, like time of use rates. Demand rates are measured instantly, in kilowatts. These settings are a little tricky to play with, and so visually examining the model to see how the system is behaving builds confidence that the settings are correct.  I’ll take a look at the visualizations just to make sure the system is charging and discharging that battery when I want it.

Here we have a couple of different configurations such as with or without solar, and with or without batteries. It’s also good to test different combinations, such as large or small solar, and large or small batteries. A $10,000 ten kilowatt hour battery is going through 129 cycles a year per year and each cycle saving at $38 per kilowatt hour. Here we have a larger battery, at reduced unit pricing, that’s going through fewer cycles which are not saving as much money per unit. I keep my eye on this effective generation rate number when trying to determine the optimal battery configuration. 

For my electric rate, the battery needs to run for three hours in the middle of summer. EnergyToolBase helps us identify the one day in summer which is the peak day, on the day of maximum usage, where the battery discharges down to almost a 20% state of charge over that period. I don’t need a larger battery than that and even a 20% margin might be too large. A 0% margin with a little bit of peak use might be more cost-effective than an oversized battery. That will vary by rate structure, a commercial building on demand shavings should have some reserve capacity built in.

At any rate, a report indicates how much of the building is solar, grid and battery information, as well as what the system payback is. 

There’s some economic graphs and cash flows. 

This is in the reports section. 

There’s a chart of the utilization rate of the battery. My experience is that the least cost-effective batteries have shallower utilization rates. An efficient battery has a deeper, more regular utilization rate. This is different than off-grid design, where you would want reserve capacity. But for grid-tied economics, if you know that on the day of maximum discharge under the time-of-use rate that the battery is only being used 80%, it’s an oversized battery. 

Knowing the right size is important because lithium-ion batteries are expensive and there is a top shelf that is even more expensive. Buying just enough for the three hour time frame will have the best financial metrics. 

Let’s take a quick peek inside EnergyToolBase. Here we have the proposal tool. The economics for solar and batteries aren’t particularly good in Mississippi, which is why we know a little more about smart home digital controls (which can be more cost-effective than solar or batteries for time-of-use rates). So your economics will probably look better than these economics. 

But in Mississippi we have to be keenly away of payback because solar policy makes it so difficult to achieve. EnergyToolbase helps me hone in on a $9000 system with a  14 year payback. Larger systems might have even greater paybacks. But this very small solar array with a very small battery can take advantage of the time of use structure to reduce the electric bill, without the need to outflow power onto the grid.

Here is the modeled energy use for the kilowatt hour charges. I like visualizing this demand

load profile. Let’s see how it works. I like turning off all the noise and taking a look at the building demand load profile. 

During this time-of-day period the building load is effectively being powered by the battery instead of the grid. Then the power surges as the battery charges back up again afterwards. I find it particularly useful to look at the battery state of charge during this time, and looking over the entire month, it is easy to find the maximum time of use period which in January is on the 19th. Taking a closer look at that time frame we have a solar

array coming on, and the solar array size can be evaluated. The calculations being built right into the software tell us that the payback period. 

If I model a large solar array at a lower installed cost of $2.50/W instead of $3/W and

click save, it says my payback is 6 years because I put in $3 as my installation price instead of $3 per watt. Well the system is not going to cost $2.50 for a 15 kilowatt array.  $2.50/W at 15kW is going to be a $37,500 system. 

Now we see that the paybacks of 24.5 years for the larger array instead of 13.5 years for the smaller array. Even though the array was installed at a lower unit price, because most of the solar array is worthless outflow onto the grid, most of that extra power isn’t doing the customer any good. 

So we have tiny solar arrays in Mississippi in order for them to have any sort of payback, unless the whole house is taken completely offgrid which is another class for another time. 

Now we are on the energy storage component. Here is a $9000 battery and that blended savings number that lets us know if a larger or smaller system is worth examining. A 5kW, 5kwh battery for $9,000 is on the high end of lithium ion. A lower product like the Tesla Powerwall would be cheaper unit pricing such that capacity could be purchased for slightly more cost. In other words a battery twice as large might only cost $5000 more, except that the battery must be discharged over a two hour period instead of a one hour period. More expensive lithium ion batteries can have more rapid discharge rates. 

Well with the larger battery, the savings figure has gone down and the payback is not very good, so the smaller, more expensive battery such as lithium NMC would be the better choice. Putting things back, the  blended savings is now a bit higher. But the solar array project component is still dragging out the payback, even with the reduced size. With the solar array the payback is 24 years and without the solar array the payback on the large battery is 22 years. So a small battery is all that is needed. If the customer wanted a more rapid payback, then they would be left with digital controls.

Before we end, I’d like to show you a solar design software that is more of a proposal only tool, so that you understand its limits. This is solargraf, which has good reviews from installers. Solargraf claims to do shade analysis so I wanted to look further into that. 

There are solar lead generation companies out there, such as EnergySage or SolarReviews. These companies sell leads to solar installers and SolarGraf puts that into a user interface right within the software. 

A marketing company gets a nibble, sells it to SolarGraf, who then resells it to the installer. The installer knows their typical components and can generate a quote quickly, and it even incorporates financing companies for loan approval as well as other sales values. So far as te design elements, the question is the shade analysis good enough to run a solar company off of by itself? The answer is probably not. There is some  awareness of utility rate structures and they have Nearmaps imagery built in, which makes a 2D proposal look much better. Maybe this photo might not be of questionable definition, but there are some horrid satellite photos out there as well.  Solargraf’s layout tool is rinky-dink but should be okay so long as the designer is not pushing the design to the edge of the roof. It does offsets and is somewhat like solar design tool in its questions, that were at the beginning of this program. It’s a bit clunky of an array layout tool.

Here is where they’re integrating with solar finance companies right into the system. Greensky and loan pal are very popular and  widespread solar financing companies 

Solargraf makes it easy to buy permit set documents within the software using 

questions are very similar to the questions from SolarDesignTool at the beginning of this program. They might not be as useful as  PV complete but it may be good enough. There re more services such as engineering stamping within the software as well. so that sounds pretty cool.

Here’s their shade analysis tool where you sketch out the roof. It gets your shade analysis and with this thermal image of the Sun on the roof. This is imagery from Google Project Sunroof and I can tell that these are sunny and this image is useful enough to say where I might put the solar array. This west facing array over here by the tree line needs to be avoided because that’s getting shaded.

So it is good for a trained professional to look at and know absolutely where not to put an array. But it is not accurate enough as a detailed performance analysis on its own, so I’ve heard anecdotally from installers that use it. 

Its better for just telling you things the solar designer should already know and acts as a bit of a safeguard to say hey don’t get too close to that chimen or dormer. But it is not true 3d model that’s giving you module level shading around chimneys or trees or things like that.

Here is a trick I use in Google Earth to determine tree height. 

Google Earth has a ruler tool where you click on the map and on another portion of the map to get the distance between two objects. You also get the bearing of the shadow you are measuring with the rule tool. So here is the bearing of 160  degrees. Google Earth also gives the date that the photo and so by knowing the bearing of the shadow and the date of the photo, 

I can use a Sun angle azimuth chart – the US Navy office makes a really good one online – to put in the location where the photo was taken. 

Here we are putting in the date the photo was taken and the bearing of the photo is 160 degrees on November 24th. So November 24th in my location when the Sun is at a 160 degrees azimuth, that photo was taken at 10:30 when the Sun is 27 degrees up in the sky.

If the Sun is 27 degrees up in the sky and the length of the shadow is known, trigonometry can determine the height of the tree. So if you’re doing  spit ball shade analysis, giving yourself plenty of clearance for tree growth, then Google Earth can give some some estimates of trees height, particularly if you will have a flat area. 

Just don’t forget to also compare the height of the tree against the height of the

house because rooftops are good 12 feet up off the ground. 

So remote analysis is possible with a few photos sent from the jobsite. Free manufacturer sizing tools ask similar questions and in some cases provide better data than commercial software. 

But in many areas commercial design software exceeds the capabilities of human design or can otherwise quickly prepare you for project permitting.

I’ve yet to find a good, free 3d shade modeling tool pr one-line diagram creation tool. I do my own economic analysis on my own spreadsheets but then pay for EnergyToolBase when needed. Commercial design software improves accuracy when approaching the final phase of preliminary design.Lastly, this is a company called solar roof check. They’re an engineering stamp as a service and I just decided to put in kind of their survey form so that you could put this data into your site evaluations. The kind of attic is important to an engineer and the spans between rafters and kind of roof decking material is useful in the design phase of the project. 

This concludes the program. Please join our discord community for discussion.

Not Permitted for Redistribution. Licensing available on request.

Posted January 7, 2020 by John Cromer in category "Classes

Leave a Reply

Your email address will not be published. Required fields are marked *