A low cost monitoring system of PV modules in nanogrid

Smart grids require to monitor a lot of parameters and variables in order to automate more and more control and management procedures. Failures and degradations that can be automatically identified in a smart system like a nanogrid are welcome if we wish to implement an efficient maintenance service to ensure the maximum PV plant duty. Here a low cost monitoring system of PV modules used for a DC nanogrid (Figure 1) is presented. The idea is to promptly identify modules failures and soiling for maintenance optimization [1]. The goal is to verify if it is possible to monitor the I-V curve of a reference module kept always clean and others modules spending less than increase of cash flow due to the reduced plant downtime and maintenance costs. Looking at the state of the art [2], we choose to use the capacitive load method (Figure 2) because it needs a short time to obtain the curve. This is almost mandatory for outdoor operating modules because the irradiance can change during the curve acquisition. Furthermore, shorter is the time needed to measure the I-V curve, shorter is the PV plant downtime. To reduce the impact of these measurements on the energy production, they can be carried out when the irradiance is low, at moments close to sunrise or sunset. Figure 3 shows the block diagram and a photo of the prototype made for this work. The core of the system is an STM32 Nucleo F334R8 board which is equipped with two ADCs to acquire at the same sampling times the current and the voltage of the module during the charge of the capacitor. The current is measured by a shunt resistor. Its resistance has been chosen to get a minimum voltage very close to zero. To measure the temperatures of modules and ambient we used three PT100, while the irradiance is measured with an analog direct sensor by SolarEdge. Some signal conditioning circuits have been made to meet the requirements of the Nucleo’s analog inputs. The cost of the whole prototype is around 100 €. Figure 4 shows some screenshots of the developed applications for PC, tablets and smartphones to set the sampling and to monitor the acquired curves. The ongoing tests, both with controlled and natural dirt, are demonstrating the validity of the developed system. It is particularly useful to optimize the cleaning of the modules, even in the case of not very dusty places. In this case, the cleaning is carried out when the efficiency of the PV array is reduced of more the 1% and the weather statistics do not provide rainfalls of a certain intensity (those able to clean the modules) for more than a month. Considering the climate and the place where the PV plant of the prototype is installed, we estimate that with this will increase the yearly production of more than 2%, compared with the same plant without a planned cleaning, gaining more than what was spent for the monitoring system and what will be spent for cleaning.