Leonardo Da Vinci believed that simplicity was the ultimate sophistication. Well, if one were to look at the principles of a traditional grid-connected DC solar system, the parts that make up the system could be innocently simple, while its design could be deceptively complex due to the problems associated with DC voltage in general. By virtue of being cheaper, a single wall-mounted inverter (called the Central or String inverter that can be easily replaced should it turn defective), would tempt anyone to disregard their serious disadvantages:
When a series of panels are connected together as a “string”, the 300 to 600 volts of DC voltage that they carry can be dangerous (just 200 mA can paralyze your respiratory muscles and snuff out breathing, while 100 mA can cause ventricular fibrillation and stop the heart).
And the other bizarre fact is that if one goes down, they all go down. So, if one solar panel fails, everybody in class fails…And that’s not all. If a single panel produces only 50% power, then the power produced by the entire bank of panels would come down to 50% as the system is only as strong as its weakest…you know the rest.
This was when someone shockingly chanced upon an intelligent solution “Eureka” …and module-level power electronics were born. Module Level Power Electronics (MLPE) are microinverters and optimizers, typically mounted on solar modules. MLPEs are placed either 1:1 or 1:2 (MLPE: solar module ratios).
While power electronics are said to have made their mark as early as 1902, the use of MLPE came into prominence only around 2009. The main advantages of using MLPE are (1) better shading control hence better energy generation, (2) better system generation visualization, and (3) integrated fire safety thru rapid shutdown.
In this blog, we will discuss the impact of fail rate on-site visits and the cost of a site visit.
When time is used as the measure over which the rate of product failure is calculated, then it gets described as the failure rate. Components can fail by either time or cause, and these should not be confused with reliability.
It is exceedingly difficult to get the exact fail rate for MLPEs from each manufacturer. The widespread industry speculation is that MLPE fail rates are between 1% and 2%. However, in lieu of extensive experience in installing both MLPEs and string inverters, it is our understanding that 1% to 2% seems fairly accurate when it comes to MLPEs.
We can model the failure rate and the number of homes based on a certain module power (watts). As the module power increases, the number of modules for a given system size will decrease therefore fewer MLPEs would be needed. For string inverters, the module power or system size is inconsequential since the number of components for a string inverter does not change with module power or system size.
Two charts are made, one for MLPEs and the other for string inverters. The factors considered are (1) system size, (2) module power, and (3) fail rate.
The insight that we can harvest from this model: as the module power increases, the number of homes for each fail increases correspondingly. Similarly, if the failure rate decreases, the number of homes to fail decreases as well, meaning more site visits and site calls.
Therefore, for a 1% fail rate and 400W module using MLPE, the solar installer can anticipate failures for 2.6 homes that have a nominal 15 kW system size or 7.7 homes for a nominal 5 kW system. On the other hand, for a string inverter, at a 1% fail rate, the number of homes an installer has to visit before a failure occurs is 100.
The mathematics then adds up. If Installer Adam had achieved a target of 100 kWs in a month and had used 400W modules for 5 kW of power in each home, he would have carried out a total of 20 installations. In these 20 homes, he can expect to have a failure in every 7.7 homes if the failure rate is 1%.
If one were to extrapolate the same 1% fail rate on a string inverter, Installer Adam will not see any fails in the 20 homes that add up to 100 kW. Only when he installs in 100 homes (500 kW) would he see his first fail.
When an object obstructs sunlight from hitting the solar panel directly by causing shade, the output of the entire string of panels is severely curtailed. The reason for the diminished performance is because the panel that is so affected can only absorb a limited number of photons. MLPEs decrease the impact of shading especially when there are a lot of trees or vents that partially shade the array during the day. The more shading the better the energy generation using MLPE.
The table below illustrates the relationship between shade and power generation.
When shading increases from 1% to 30% during a day, the sun hours lost also increases proportionately. For example, if 30% shade falls on one solar module only, then the sun hours lost is 30.9. Whereas, if two modules have shade for 30% of the time, the sun hours lost every year is 61.8 hours.
If shading is significant, it is best to use MLPE over string inverters. However, when there is no shading or very minimal shading, there is extraordinarily little gain from the use of MLPEs.
When the MLPE fail rate increases, a module’s downtime also increases. This happens when the installer is unable to get to the site instantly, to remove and replace or reinstall the failed microinverter. Thus, the energy production is down for that single module till the problem is rectified.
In the case of string inverters, since fail is few the service calls are few.
If an installer takes 60 days or more to remove and reinstall a defective MPLE, the total power generation loss is negligible, whether it is an MLPE or a string inverter system. Simply put, it is better to use a string inverter if the installer were to take 60 days or more to rectify a single failed MLPE.
The table below illustrates the sun hours lost in solar production and the number of days it takes for an installer to rectify a failed MLPE.
For an installer, the cost to replace the MLPE is generally higher as they are installed on the roof – this requires the technician to climb the roof and use safety equipment to secure himself. Such additional costs are incurred only in the case of an MLPE system and not on string inverters. Replacing a string inverter is simple and inexpensive, as most of these are on the ground requiring either the swap of a board or removal and reinstallation.
Microinverts have solar module-level monitoring which gives much better granularity and visualization. If one MLPE is down, the map clearly points to the failure on the roof (mostly MLPEs are used on roofs) hence provides a way for the installer to access and fix.
String inverters, on the other hand, have a system-level generation; which means it indicates what the entire system is producing at any given time and does not have module-level granularity. But if the string inverter is down, it is monitored on the cloud and will send you an alert to troubleshoot. And in almost all cases, the failure of a string inverter is on the inverter itself and not on the solar module.