welcome to the design analysis and operation of photovoltaic power systems with a tab for the next 45 minutes or so we'll be talking about photovoltaic technology and how do we essentially utilize a tab to enter information regarding photovoltaics and essentially how do we migrate that model and move it into an online system and then once we move it into an online system as well how do we then go back to our simulation and kind of compare what we design versus what we are now seeing and in the actual system so before we actually go through the entire process just a quick recap on the solar technologies there's many types out there such as solar thermal photovoltaic which is semiconductor based concentrated photovoltaic that essentially utilizes light concentrating technology to pass the solar radiation onto a semiconductor and there's other various technologies that are up and coming and under research such as infrared solar cells Hughley solar cells that you can put on your windows 3d solar cells that give you more watts per square area different types of materials besides semiconductors exotic materials they can increase efficiency to almost 100 percent and then P V thermal hybrid where you can actually utilize the the heat generated by the solar panels and use it for other applications so for the next 45 minutes or so we'll be focusing on photovoltaic only and again just as a recap regarding the photovoltaic electrical characteristics typically when we look at the semiconductor depending upon the manufacturer there's obviously an internal equivalent circuit model that you can look at that has the diode resistances in series and parallel but at the output of the photovoltaic panel the manufacturers typically publish a PV curve and an IV curve so the PV curve essentially is power versus voltage and the IV curve is current versus voltage so here we can see an example of an IV curve and how the curve varies as the cell temperature exhaust so if your ambient temperature is increasing that will actually have an impact on the overall cell temperature and the performance essentially will drop on the PV panel so it's important to cool them especially if the temperature ambient temperature is very high there's also dependence on radiate irradiance which is basically the sunlight that's typically in watt per meter square so as we can see from 200 to about a thousand which is a theoretical maximum in the atmosphere the current generated by the PV panels increases as you have more sunlight so that's intuitive when we look at the power curve and the IV curve and overlay them on top of each other we can see a correlation between the two at the peak off the power curve it essentially corresponds to the knee point of the IV curve so essentially on this point right here that's where we essentially going to get maximum power so all the inverter technologies try to make sure that they have a concept of maximum maximum peak power tracking or MPP control where you can try to adjust the the inverter such that the output is at the midpoint or essentially at the MPP point under normal operating conditions during peak sunlight hours so again that's just a basic background and this is a typical utility interconnection so most of the PV panels are typically connected in an array of multiple PV modules the block sizes vary depending upon the size of the entire solar farm that we we won MBA blocks to five MBA blocks each with their own collection system with its own inverter with the step-up transformer going into the main power collection station and it stepped up further into a grid environment the the connections are very similar whether it's solar farm versus commercial facilities however there are slight differences between the techniques and the reliability techniques used to connect them together but the basic principle is the same you have a PV array going through an inverter there may or may not be a step-up transformer and you're connecting into the grid so that's the basic connection as far as the modeling and analysis is concerned we obviously want to design the PV system size various equipment like cables transformers inverters and perform various power system analysis like low flow for current our clash calculations any disturbances that we want to introduce such as a fault on the grid side and look at the right through characteristics of the inverter and the return on investment analysis such as estimating the losses and the overall kilowatt hour or megawatt hour injection into the system so there's the different types of calculation modules utilized for each of these types of modeling and analysis and within the e type solution we'll be covering some of these technologies that have been introduced to help design and operate PV farms and PV systems in general so in E type you can have series and parallel connection of arrays there's a comprehensive manufacturer model library that has IV and PV curve characteristics we do adjust the performance of the PV and IV curve as as I mentioned earlier based on various coefficients that are published by the vendors there's even a sword solar irradiance calculator that will help us calculate the theoretical power for a brand new site various inverter modeling including MPP again which is maximum peak power tracking mode you can actually have grid sign generator modeling which means that the program doesn't really want you to have too much detail on the individual bits and pieces on the string cables and so on so forth on the DC side but you're essentially using ETA up to model a equivalent PV form connected to the grid and the focus is grid analysis so ETF does support that kind of connection or you can go into a full detail discrete DC model where you can build every single combiner box junction box every string cable and do a complete detailed DC analysis from solar farm or a solar system designer perspective you can also do AC and DC analysis and we'll look into that as well the PV system is integrated with all of the e type analysis modules so you can also use it for more starting transient stability and other types of calculations including protection and it's also integrated with e type real time micro grid energy management system so we'll take a quick glimpse at that as well so the idea was to take one system in this presentation and walk it through design into commissioning into operation and basically the analysis at the end so we decided to essentially create a system at our corporate facility I'm just using that as an example in this PowerPoint so we took undertook a endeavor to essentially learn more about PV systems a few years back so we essentially planned and did a lot of designing we finally built this system and as we were building yet to go back and redesign our inverter strategies we went back into the construction phase commissioned and integrated the system and now we essentially managing this solar system as part of the corporate micro grid so besides obviously understanding the the way PV systems work we also want to understand the issues faced by customers and typical issues faced during designing and constructing these types of systems so I'll also talk a little bit about some of the lessons that we learned through the process and obviously to conserve energy and lower the overall energy cost for our corporate building so the objective for the overall system was to add panels and offset the daily peak especially the summer peak our corporate office is based in Southern California we are blessed to have clear blue skies warm weather so we should be able to take advantage of the summer Peaks and offset the daily peak demand with the power from the PV panels and overall reduce the overall energy consumption and also we wanted to introduce electric vehicle chargers for evey vehicles to essentially get some of that solar energy in addition to energy we purchased from the utility and and do a composite so we also added a weather station to measure the real-time data so we can take a look at that as well and then finally we wanted to add a micro grid controller so we could see the performance of the system and adjust the strategies according to our energy-saving objective so this was the overall objective of the project so I'll go through the planning and design phase of the project the system was designed for about 180 kilowatts we used the NREL multiplier to account for losses any performance degradation of about 0.75 so we were expecting a peak kilowatt power of 133 kilowatts so later on I'll show you what what we're actually getting versus what we estimated but this was on paper in order to do a return on investment calculations the expected savings due to the PV generation of about to 120 mega watt hour for the whole year and the annual savings in electricity used to be approximately 50% now not going through all the numbers in details but as you can see in any commercial facility whether it's a data center or a commercial office building whatever it may be the the biggest factor is or the biggest energy consumers cooling and any cooling fans or HVAC units and then the lighting computer servers which are receptacles and any other process related items they're up there in in kilowatt usage but the biggest consumer of power obviously is the HVAC units so that was the main goal to cut down the demand especially during peak afternoon because that's where in the ACE HVAC units are running at their maximum power and that's really the peak that we wanted to shave using the PV panels so we started with that and we went into the economic justification and this is just a quick summary of the output of the economic justification and as you can see the calculation was done for a whole year and we were expecting savings of about 27,000 dollars on our electricity bill for each year so if you multiply that times 10 it would essentially give you about a quarter million or three hundred thousand dollars of savings over ten years which was essentially the cost of the solar panels alone that wasn't really the total cost of the project but essentially was offsetting the direct cost of the solar panels and there are about five hundred sixty four solar panels which I'll show you later on so we obviously had to go into the architectural phase and that's typical of any PV system project as well and design the from an architectural point of view how the panels will fit what are the easements and the structure layout I'm not showing every single architectural drawing but this is the overall plan view showing that we essentially have the e tab corporate office building we have our parking lot and essentially there's a solar car shade covering two rows of parking and then one all the way in the back so the three of them put together is about five hundred and sixty for individual PV modules that were essentially going to be placed alright so after the initial architectural drawing another typical thing that is done is an artist's rendering so this is what the artist's rendering looks like for those of you may not have seen the e type of office have not had the opportunity to come visit us you can see that's really our corporate office and this is the artist's rendering of the solar carport and the outcome is very similar to what you see here and I have some pictures later on for you to see so again the panel was supposed to do daily peak shaving in here the peak shaving is indicated by the load which is in red the resource is essentially in yellow so what we expect to see is this blue line which is between about 10:00 to 4:00 p.m.
We lower our peak demand and then obviously PV chargers I've already mentioned we added two of them into the system we did not use a sophisticated way like this to determine our weather this is what would be used in Hawaii if the coconut is wet it's raining we did something a little bit more sophisticated we added a weather station that has communication back to our SCADA server that measures different variables such as wind speed humidity irradiance and so on and you'll get a look at that in a little bit we added electronic control onto the thermostat so we can regulate our demand as well and then we went to the actual electrical schematic and this is just an overview of the electrical schematic and I'll just highlight some of the main points here these are essentially the PV strings this doesn't represent individual PV module but a string that's going into a junction box and into a combiner box and each of these combiners you have about nine to ten strings each but they're variable you can see two tens 911 s and sevens and this was essentially done because of the layout of the physical space not for any other intentional purpose I really would like this to be balanced as far as possible but we can see that we have a little bit more room to grow here in a combiner box number seven if you want to expand further into the remaining car spaces the combiner is essentially land up going into the inverter which is shown here with this rectangle so we opted for a 250 KVA inverter grid tied inverter and that goes essentially back into the mains switchboard that's fed by the local utility here which is Southern California Edison and that's a step-down transformer which is the main service entrance t1 and that's a 750 kV a service 11 kV down to 480 volt service so that's basically the overall setup of the system and we'll see the same thing on the on the ETF side so we went to the construction phase I'll just skip through these slides very quickly so you can get a flavor of the kind of effort that went into constructing the system here's a finished product almost finished product of the PV panels and these are essentially 305 watt poly crystalline modules that were plugged in the communications were essentially set up as well for each of the incomers from the utility from the inverter itself and then we also installed brand circuit monitoring for some of the loads that we wanted to measure in a little bit more detail especially the HVAC units we also put metering onto the UPS which serves the data center that we run within the facility as well as the critical computers especially for the developers that are working very hard on the e type software we went through the commissioning phase as you can see here which is going through a test right now and we actually had our first injection into the meter so we can actually see negative power going into the Southern California Edison meter so we were actually selling the moment we energized the system now when we talk about energization and working on equipment one of the things we always talk about is arc flash we were on our time scale to get the work done so some of the work was done energized but typically you would do this with hot gloves and appropriate PPE in this case you can see these two gentlemen actually adding a new circuit breaker on to a live bus but using hot gloves and wearing appropriate protection and then we finally went through and connected everything together so the solar inverter the combiners the weather station the UPS the thermostats and the data center are all communicating into one single location that's being monitored by each app real-time and it's being broadcasted into a web dashboard that we'll take a look at in a little bit and that's a first version of that dashboard that you can see on the web or a tablet so let me first talk about the lessons learned before we jump into the eat up site these are lessons learned during the construction phase so what we did learn was when we are purchasing combiner boxes there's typically an option to either have a fuse disconnect or just a disconnect and we land up with an option to have a simple disconnect and that disconnect is from the PV panels to the combiner box and you can see that disconnect is shown here in the rectangle and the fuse is right here what we were not told from the vendor that the fuse disconnect was optional so obviously we went ahead and purchased what was standard and we later on found out from during the commissioning phase that we should have had a fused disconnect so it had to be retrofitted into the system as you can see here the fuses were retrofit in but that it's essentially caused additional work so it should pay attention to your local codes to make sure that from the DC side into the combiner whether a fused disconnect is required or not and this fuse is there to protect the cable going to the inverter itself the inverters themselves have to be fused on the DC input side which is also an optional extra and again you have to make sure that when you're ordering the inverters check with the local codes to make sure whether this views on the DC input sign is required or not and that fuse is shown here in this orange color so these two fuses are there to make sure if the fault on this cable is cleared and the cable is protected from a fault on either side of the equipment or an internal fault the other thing that we learned very early on was to factor spares into the design in the previous picture as you may or may not have noticed we actually had to do some trenching to lay out cables the practice to always make sure that there are sufficient spares or conduits essentially provided to accommodate for illumination any plug-in electric vehicles security cameras plus the communication and control power because the combiners themselves may not be self powered they will require control power 12 volt or 24 volts DC input and that has to be provided from the facility to the combiner box itself it cannot be provided from the pv panel because at night you will lose communication altogether so it has to be sustained power one of the things that does create a problem and it has to be done in parallel is illumination again you have to do a lumen test and make sure that whenever there's any PV system going in what are the illumination requirements around that facility to make sure that again the appropriate illumination type is used in in our case we used LED lining to reduce the amount of power and then appropriate cables are run to make sure there's sufficient lumens available throughout the facility and that has to be determined by the local inspection that's done at the facility so here's a quick overview of the yearly production this is one years worth of data that we did from 2015 from last year February 2015 to February 2016 so it's very current and as you can see that the output from the system appears to be cyclical so you can see in February we were somewhere about 65 kilowatts of production and we cycled down to about 45 or so and then we are actually coming back up and this is February right here and we're coming back to our 65 kilowatt levels if I trend the line between the two I should not see any degradation in performance because of aging because it this is only one year's worth of data and the system has been in place for the last two years but over time you will see that from month-to-month it may not return back to the exact same levels and that's just degradation of the material itself the other factor in case it does not return back to the same level could be just simple dirt on the panels and typical practice is to clean them every three to six months six months preferably complete cleaning to make sure they're operating at optimal efficiency now if you remember in the initial estimates we size the system at 180 kilowatts we applied an NREL factor of 0.75 as I had mentioned and we expected the output to be about 133 kilowatts peak summer so we did not really achieve on 133 kilowatts but we did it to you about 125 or so so we were not that far away from the NREL factor that kind of estimated it and and this is actual production data showing us the same data is actually folded together so this is again almost 1 years worth of data folded together from 6:00 a.m.
All the way to midnight and as you can see most of the production starts around 9:00 a.m. and it essentially ends anywhere between 6:00 to 8:00 8:30 p.m. and and this large shift is also because of daylight savings and the fact that we are and another and latitude so we do have later Sun in summer we do see a shift plus we have an hour time change that causes the production time to kind of shift around so that has to be factored in when we are doing the yearly analysis so instead of looking at this data in this fashion I took a single day and this is what we're gonna analyze inside eat up right now so I took consumption versus PV generation data which is for February to 15th so it's about two two days ago and this blue line that we are seeing is the load the red line is the PV generation and this is essentially my 24-hour clock in a circular fashion so again the same type of behavior can be seen the PV generation is happening mostly it starts around 9:00 to 10:00 a.m.
And it's maxing out around 1:00 to 2:00 p.m. and it starts dropping off till till it gets to about 5:00 or 6:00 p.m. and we are essentially in the winter to early spring months so we do see the the Sun come up a little bit early and then set around 5:30 p.m. so that's why you will see the output dropping that fast at least in this location so if I take a slice and that slices let's say 1:00 p.m. so 1:00 p.m. we can see that the PVE generation 65 kilowatts the load is almost about a hundred kilowatts or so so I'm going to switch over to eat app and show you the model here we have two different types of models and we're gonna run that analysis in this model and see what we get so first of all I have the entire system zoomed out so you can see the overall structure the lines that you see in blue are DC the lines in black are AC and there's two different systems here I have one system up here which is a detailed model of the PV system having every single string cable and the combiner box to the inverter into the main switchboard and there's another system down here which is taking the entire DC system making it equivalent into a single source and I have the entire AC network so if I was doing grid interconnection studies I would be using a single source generation single pv source generation and then modeling the details of the grid on the right-hand side and my main focus would be to see what happens in the grid when I have pv in my system I'm more interested in the reliability and security of the grid I don't need to see the details of the PV system so I can make all of these panels equivalent have an inverter and inject the power into my AC bus and that component which is a PV array equivalent AC PV array Copland component is available in the e type toolbar right here so if I open it up we can see that I have my PV panel selected I actually grabbed it from the library so this is Renee Sola 305 watt panel polycrystalline and this is the PV and IV curve as I was describing earlier so this is actually coming from the library these are the performance co-efficient were provided by the vendor so I'm essentially utilizing those and then I have about 12 panels in series so that basically makes one single string of the PV system that we have and that one string actually combines together in these panels to create essentially 440 volt DC so it's like putting 12 triple-a batteries together and making a larger voltage battery so that very essentially putting 12 of these in series and 47 of them in parallel to give us a total kilowatt DC of about 172 and a total DC voltage of about 440 as I mentioned earlier about 564 panels and about 400 amps DC that we expect from the all these panels going into the inverter and then I can describe the inverter characteristics itself as I mentioned earlier it's a 250 KVA inverter at 480 volt AC side and we plugged in the efficiency at different percent loading now for inverters the efficiency curve is not very dramatic it is quite different for induction machines or any other type of load but because PV array is mostly electronics based the output for larger inverters or the efficiency of a large conversion is not that dramatically different as the load increases your decrease in the system but nonetheless it has to be entered in appropriately if you want accurate simulation of losses from from the system so if we look at the same PV array which is an equivalent I have 564 panels here with an inverter injecting power into the network if I step back to my diagram here this is exactly the same system if I open up one of these panels you will see it's the same panel here twelve of them but only one in parallel because I'm representing PVA one as one string and that's my choice in this particular system there's no inverter option here because this is a DC element and this element you can find from the PV array toolbar here the DC toolbar and you can drag and drop that into your model but from the PV array I have my string cable which has its own individual fuse going to the combiner cable and this bus essentially represents the the combiner point inside the inverter so actually technically this DC bus is part of the inverter and this bus right here is the combiner and these are the individual fuses that we retrofit later on inside the combiner box to create a fused disconnect so you'll see about 47 of those here so the entire system you see on the top is discreetly modeled DC with the combiner string cables with the inverter er to the AC side and then the network you see on the bottom is all of the blue stuff combined into one equivalent piece back into the grid so when we actually go ahead and run calculations within a tab typically when we run simulations in this case I'm actually producing about 150 kilowatts my load is about 150 7.9 I'm actually buying about 7.5 this is one of the scenarios that we had created early on in the design phase but coming back to the simulation that we were looking at on the slide this one p.m.
One where the load was ninety nine point nineteen and the PV generation was 65 let's see what happens here so if I go ahead and change the load to 99.99 just manually on the rating site you can also change it using the loading categories here I'm just just simplifying things and we will go ahead and put the PV generation as well over here so the the PV generation that was measured was about 642 and when I say PV generation I mean the irradiance on the PV panel not the actual kilowatt output itself theoretical was 875 so let me explain what these two values mean so back to eat AB now based on the location 33.6 north and 117 west if I look at February 15th at 1:00 p.m.
And go ahead and calculate you will see that the theoretical irradiance should have been about 875 so this is the irradiance that would have been falling on the panels provided I had clear skies no smog no fog no atmospheric issues whatsoever the sunlight comes in perfect perfectly diffused into the atmosphere and falls on the panel so this is what we call the theoretical value and if I update I press on that Update button right here what you will see is that 875 gets plugged into one of my generation categories there's another factor here that you have to plug in which is the ambient temperature because again if you remember from those characteristic curves the ambient temperature will affect the IV characteristics and will affect the overall output of the panel the difference is not a whole lot but it's sufficient that it should be factored in if the ambient was 30 degrees you can see that the output drops above or kilowatts or so but in a much larger system that for kilowatts can add up into significant difference in the output so let's assume the ambient was 30 because this is about 1 p.m.
And last this week actually has been a warm week record-breaking temperatures in Southern California so I'm going to go ahead and run that calculation real quick and we see that the output based on ideal conditions should have been 140 kilowatt but obviously we did not have ideal conditions based on my weather station that I mentioned earlier instead of getting 875 watt per meter square I actually got 642 on that day and that was from the archived data so in reality if I was to simulate this correctly this would be 642 watt per meter square and that simulation would have shown that I should have been producing hundred kilowatts or so and selling about 8 kilowatts so my net meter would have shown negative 8 kilowatt going into the utility however when we look at that and compared with the PV generation values so simulated theoretical simulated value based on theoretical irradiance we saw was 1:40 simulated value based on actual irradiance which is 642 was about a hundred which we just saw but measured was 65 if I go back to my chart you can see at 1:00 p.m.
Measured was actually 65 so we can see that there's an approximately 65 to 75 percent difference between what we are simulating versus what's actually happening in the system now why is there a difference now it kind of matches up with that NREL factor that I was mentioning earlier 0.75 but why does that happen so some of the factors in in this particular net network or in this particular system are as follows we are utilizing averages so I'm looking at and average value to do this comparison I'm not using exact instantaneous value so if I want to be more accurate and get a better feel for this deviation I should look at an actual instantaneous value instead of looking at averages for a whole hour because I could I could have multiple measurements and come up with an average that doesn't really correlate very well between the irradiance and the actual generation then we actually have shading differences so depending upon the location of the structure with respect to surrounding buildings shading is a factor now in commercial spaces that may or may not be controllable but obviously if you're dealing with the solar farm that is easily managed by just adjusting the spacing between the panels so that the shadow of one does not affect the other there's also another factor which is surface dirt and other variables such as losses and things of that nature but overall we have our different time slices for that day we have anywhere from 65 to 75 percent deviation from real time measured versus simulated and we kind of expected that again using that NREL factor so what we intend to do is come up with factors for each of these items the average is the shading the surface dirt and we're going to essentially adjust our model so that we can get a little bit more accurate prediction a little bit accurate what if without really using forecasted irradiance if I want to know what's going to happen tomorrow and adjust my generation accordingly I can use this model and really not rely on weather data necessarily and then we're going to do further analysis and hopefully in another webinar to follow and compare the forecasted irradiance from a weather service versus this forecast model that we're gonna come up with so again this was just a single day analysis to give you an idea what lends up happening between theoretical versus actually measured values they may or may not match but there are various reasons associated with it now some of the other lessons that we learned especially with shading is to avoid strings with panels at different tilt angles now you can take a look at this structure and you can see the the problem right away the the irradiance is not evenly distributed along the panel's though from this picture the curve may not look that much but there is sufficient tilt in in these panels that are different from each other so each string is tilted in a in a different manner if you recall when I put in an actual irradiance of 642 watt per meter square again we're assuming that all the panel's are facing the Sun appropriately or they're being tracked in such a manner that the tilt angles are optimal but as you can see here the tilt angles are not optimal and that really accounts for that difference in the model versus the the real-time measured value so the string should be organized again such that panels of similar orientation are strung together so you can actually have strings going this way so that they are all at the same tilt angle instead of having strings going around or across the so the solar panels now in our case the the strings are actually going across and that was done purely for construction purpose the tilts are more for aesthetic purpose there was no other scientific reason for doing it in this manner it was mostly for purposes so this is a significant factor if the panels are not at the same tilt angle so let me go back to eat up here real quick and go back into my DC side and show you the same calculation here slightly different numbers but if I go ahead and run a DC power flow you will see that we obviously calculate the power injection from each of these panels you can see that each of them is producing 6 3.5 kilowatt which will really not happen because they're not at the same tilt angles but because this number of strings are different you'll see 21 17 21 and so on so you can see the voltages at each combiner you can see the flow through with each of the combiners the input into the inverter and then we kind of stop right here because this is the extent of the DC power flow you can save this operating value into the inverter and then follow up with the AC calculation so you can now see the AC injection into the grid this is one way of doing it but what we've done is we actually introduced another type of calculation inside inside a tab which allows us to do a complete AC and DC simulation in one shot so you can see my DC power and my AC power injection into the the grid in one single calculation without actually having to go back and forth and iteratively trying to solve the AC side and the DC side obviously if you're not interested in the details of what's happening on the DC Network you don't really need to model in this fidelity you can just look at an AC only equivalent circuit okay so some of the other lessons learned to summarize them real quick so the other thing that we learned was grounding now this is a little bit more with respect to National Electric Code and that's a little bit more us-based but I'm sure there's similar electric codes in other countries such as the Canadian UK they all have their local codes so foregrounding definitely you have to follow the local code of the region you're in in u.s.
North America California we're following NEC so NEC requires that all PV systems or 50 volts have at least one current carrying conductor connected to ground so that's from 690 point 41 and all exposed metal parts conduits enclosures everything associated with the PV system regardless of voltage should also be grounded at 690 point 43 and that's just to make sure that if in case there is a ground fault because of an accidental short in ground one or more in one or more current-carrying conductors you can actually have a DC arc depending upon the voltage level and the amount of current being carried and if the fault is not properly cleared the arc can cause a fire hazard so there's another calculation that we've added within a tap and unfortunate we don't have time to go through all the details which is DC arc flash so this will essentially allow you to place a fault at certain points in the system so you can either right click and fault or go into the steady case and fault select the the short circuit model for various sources such as batteries rectifiers motors inverters and so on select from different arc flash methodologies and etabs actually the only program right now that supports three different methods based on the complexity of the system for doing arc flash calculations such as maximum power Stokes in Apple Lander and you can actually use the DC arc flash with the protective devices as well just like the ACR clash module to calculate what is the arcing current the fault clearing time instant energy and so on and so forth now in my system I've not set up all my protective devices appropriately so we will essentially not get an output but we do see what the fault current levels are over here so there is a DC arc flash calculation program or module available inside ETA to help you make sure that the grounding is done appropriately and it is safe especially when if you're planning to use a plan to do work on the system energized it will help you understand what is the appropriate PPE to be used obviously in PV systems the best thing to do is operate on the system at night so there's hardly any irradiance there's hardly any generation from the panel's and most of the work can be done safely at night but in cases if it has to be done during the day because of some kind of Fault in the system is not possible to shut down the entire system you may want to run this calculation and determine the appropriate PPE and label the equipment accordingly on the array grounding the equipment grounding conductor which is EGC and the grounding electrode conductor gec should also be installed appropriately 690 point 47c tells us that and then what we ended up doing in our system was we put a solid bare copper number 8 size wire and that's a WG american wire gauge size to essentially have a common ground common EGC and grounding electrode conductor to have a appropriate system ground and the equipment ground going finally into the the main ground of the building so the PV structure was also grounded to the main ground or the u4 ground as we call it named after and the inventor of the this type of grounding is essentially a ground conductor in encased in concrete so there is a you for ground available for most commercial structures here in North America Peevy structure was actually grounded to it according to article two fifty point 52 so again this is just a quick overview of the grounding scheme that we used in this building this is the PV string from our solar carport and we took the bond from these aluminum frame of the PV panel to a grounding bus bar within the combiner box that itself was connected to the structure ground and grounded and the equipment ground was carried through and brought into the ground of the inverter which also is grounded back to the actual switchboard voltages incorrect is actually 480 volt and everything is bonded together through a common you for ground so that's essentially things that we learned on the grounding side this was through again through the commissioning phase we had initially just grounded the system directly to the u4 ground and we had not carried the ground wire over to the main system ground so we said she had to go back and add this EGC wire back in to make sure that the these two systems were tied together and going to a common ground so let's really talk a little bit about the savings before we go into the the real-time side of things so we installed the system back in 2013 we've been operating it since then I just picked a sample month which is August because it happens to be a peak month as far as power consumption is concerned in Southern California at least for our building so as you can see in 2012 which is in blue this is before the PV system was plugged in the kilowatt hours users about 41,000 number of days is basically the billing cycle from the local utility but the other key component here is the maximum demand it used to be about her and 44 kilowatts and this is the actual bill about seven to eight hundred or eight thousand dollars per for for the month of August and and this is essentially that tou rate or time of use rate and that varies from utility to utility this is the schedule that's provided by Southern California Edison which is gs2 general schedule number two used for commercial facilities and that's the schedule that we were on and then August 2013 we plugged the system in which is shown in green so that's essentially when we went switched over to green energy and you can see that the kilowatt hour usage dropped significantly down to twenty-seven thousand kilowatt hours and we are essentially on net metering hence you will see that this offset or this drop is because we were producing the difference or almost the difference of the two and providing the power to SCE during peak summer months the kilowatt demand we managed to drop it down from 144 215 so that was accomplished through the micro grid controller or essentially tap real time we are managing this maximum demand so that it's comfortable temperature but we do not reach a maximum demand of 144 because there's in the GS 2 schedule and again it varies from utility utility you will find that there is a time of use charge on the kilowatt hour and you'll find that there's a demand charge on the peak demand so these two are the main charges that add up together that make up this $7,800 obviously there's other pieces or components to the bill but these two are the main contributors so if you can lower both of them you can lower the bottom line or the overall bill that you received so 2013 we saw a drop down to 6300 2014 be tweaked and optimized a little bit more and we went down to about fifty nine hundred and then this past year even though our consumption went up our demand went up but you can see that our bill was still lower and this is what this was done by us negotiating with the local utility and that's something again as part of the lessons learned to make sure that you're on the correct schedule we switched RTO time of use rate we switched over from GS to 2g s to R which is general scheduled rate for renewables which was offered again in 2015 so that essentially lowered the time of use and demand charge so again reduced the overall bill so if he add up the savings put together we are not exactly in line with what we had estimated in people in our original estimate but we are close and hopefully in the next webinar we can go through a yearly analysis of the cost as well and and the ROI of putting such systems with the current utility rate schedules that are being offered here's a quick event that I just wanted to share with with you which is maintenance shutdown event so these are essentially we are tracking instantaneous values and this is from back from 2013 and we can see what we're tracking is actually the demand in the system the negative values are correct that's essentially where we were actually injecting power back into the network but as we can see over here at this particular date on August 19th we were somewhere in the range of 100 to 210 kilowatts everything was fine we were generating power and we had a outage on the inverter so as soon as we had an outage be the maximum demand for that month and obviously we were built the maximum demand charge for that month but you can see overall for the month we were doing very well till that outage event happened so this is again just another example that you have to make sure that any maintenance or any outages are prevented during peak operating time because once you have an outage you will hit the kilowatt demand and the corresponding charge and that applies for the whole month and it kind of negates the benefit of having the PV array system so this again justifies the need to actually have a controller in the system that can quickly correct for this situation and not allow the peak demand to continue to a higher level because there could be additional load coming on at this time that could have taken this peak demand further into the 130 140 kilowatt range so in conclusion from the the design side you can actually have considerable cost savings by reducing the kilowatt demand as I mentioned earlier you shut down the inverter for maintenance during off-peak hours and installation of any peak demand reduction system like ETA real-time will essentially give you the direct return on investment because you will see a reduction in the peak demand charge so with that let me quickly show you what we're seeing on the real-time side and let's take an example off right now and let's see what we get from our simulation so what we're seeing right now is an actual view into what's happening in the PV system we're actually producing about 43 kilowatts or so and the power that we're buying from the utility is about 20 kilowatts our load is about 65 and you can actually see right here around the time our somewhere in the middle when the webinar began Lord essentially went up while while we were going to our other discussions and that's essentially the air conditioning that's coming on because we are approaching 10:00 to 11:00 a.m.
The outside temperature is past the set point of the thermostat so that's actually triggering certain air conditioning within the building now that same single line diagram that I showed you in on the e type side with these strings has been replicated actually exported and replicated here onto this dashboard so it's essentially a live view or the main actual single line diagram as you can see those individual strings with amps connected to the combiner box they're shown here so you can quickly build the design model inside eita export that model and have it available as part of the dashboard for operating the system so you can see from each section from each combiner v– we are seeing about nine to ten kilowatts and obviously the last string as as I mentioned was only the last group was only seven strings so you'll see about seven kilowatts or so coming from from that group going into the inverter here's the inverter output and here's the power that we're buying from the utility and we can see what's going on with the individual breakers within the system and as you can see the bulk of the load is air conditioning and the the computers in the system so that makes up the maximum load that we're seeing at this point in time I can actually drill down further and you can see more detail about what's going on within the data center down to the PDU level you can actually then also drill down into the air conditioning side and you can see what's going on with each and every single air conditioner in the system and we can see that there's actually two of them running right now thing where we essentially definitely want to keep them cool and the cafeteria where people I should probably enjoying breakfast so that kind of makes sense that's where the air conditioning is actually running right now and if I look further into other types of dashboards here I can see here's what's going on with my pbra itself I'm seeing about 612 watt per meter square of irradiance in my system the temperature is about 76 Fahrenheit and the colors I actually are telling me here's what's going on on the panel the the lighter the color the more power is being produced the darker the color that's approaching darkness the less power the panel's going to produce these panels actually color-coded – to tell you if if there is any panel producing less power because of a cloud that comes by or the Sun angle is is changed so you will see a different gradient as you approach extremes of the day or extreme weather conditions in the system there's a little bit more weather data that we're collecting from ETF real-time and I can actually be displayed in the ETF model as well not just this dashboard so if I take 613 watt per meter square and apply it to my model and that ambient temperature probably corresponds to about 22 degrees Celsius and if I run my power flow calculation I should have been seeing about 98 kilowatts from from these panels I should also change my load correspondingly the load was actually about 64 kilowatts or so in the system so I should have been selling about 34 but that's really not what's happening in my system what I'm actually doing is mine 15 so there's definite difference between the design model and the planning model and by actually having them in a tab itself you can actually now start tuning the model to make sure that what you get from your simulation and what you're seeing from the real system actually corresponds in jives together and actually and each app does have tools that allow you to do that and we're not actually going to do that in this model and hopefully have that available for you to see in the next webinar have that tuned as well as using forecasted weather values other things we can do in this dashboard include HVAC monitoring so we can see what's going on with temperature different parts of the building as you can see accounting is set to 71 and we know that because the air conditioning is running so it's measuring about 71 Fahrenheit here and there's air conditioning running in the cafeteria too so that's sitting at about 65 fahrenheit so that kind of confirms what what we're seeing with the electrical side as well there's a section for alarm so if there are any alarms defined they would show up here right now the system appears to be running perfectly and there's also a trending application and this is how I obtained the data that I showed you earlier so I can actually go into my real power for my inverter my main income and the consumption so I'm actually trending real data right now that I'm measuring from the system you can actually combine this together and see it in one chart and this is what we are seeing is live data or you can go to a user-defined date and if I go back to the fifteenth 12 a.m.
To the 15th let's say 11 p.m. here's essentially how I got my data so you can see very quickly in how fast I was able to pull up this information and see what went on that date but you also have the capability to go further back I'm going now 15 days or so back into the system and we can quickly see what happened here and everyday you can see the there's a little bit of extra generation excess generation which is the yellow line that goes over the red line and that's really where we are actually selling power back into the grid these two dates you will see that there's more power being sold because this happens to be weekend Saturday Sunday and that repeats again Saturday and Sunday so that kind of again giants with what actually happens within within the system other things you can do here is energy accounting and that's how I actually got my kilowatt hour numbers you can also see the communication diagram to make sure everything is operating correctly and we're using various protocols here like Modbus 61 850 and SNMP and amongst others and here's the the neat part I can also do my simulation from right here so here's how my system is running right now and what if I just wanted a quick controlled decision to be made well what if I shut down both all the air conditioners in the building right now so I can double click on those two breakers you'll notice that I originally had them in black I'm gonna go ahead and open them so it'll turn it to to green so I'm opening these two breakers and I'm gonna press run now I'm actually running a simulation inside each app and come bringing those results back into my dashboard and you can do this from anywhere from within your network or outside your network depending upon your security levels and you can see that if you were to shut down those two air conditioners yep you would be actually selling about 1.5 kilowatts or so back into the grid and as you cut more and more load into the system we obviously expect more power to flow back into the grid so this type of simulation allows you to even quickly come up with what-if scenarios and and these are essentially decision strategies you can make for a more complex I'm using a simple system to help to help illustrate the the capabilities of the program but it doesn't necessarily have to be a renewable system it could be any industrial commercial or utility distribution system and you can do the same type of what-if analysis straight on the dashboard so hopefully I've managed to convey from design all the way into operations in a short timeframe there's obviously a lot more details involved in the modeling side of things and depending upon your response and requirements we can actually organize additional webinars and perhaps go into more specific details on each of these items once again I appreciate your time and apologize if it took longer than expected thank you again for attending this webinar