3 Accuracy

Current measurement accuracy is critical to the solar inverter system, because it determines the control accuracy of the power stage and further affects the energy harvest efficiency. SBOA624 discusses common solar application scenarios with different current measurement accuracy requirements in different current sampling places.

Table 3-1 summarizes the typical design target of current measurement accuracy in solar inverters.

Generally, string current sampling is mandatory for string current monitoring and display. Additionally, string current sampling is also widely used for I-V curve scanning and diagnosis for smart maintaining work, especially for high power string inverters or central inverters deployed for grid or commercial-industrial application scenarios. For monitoring and display purposes, usually there are not strict accuracy requirements, 3% is enough. However, for I-V curves scanning and for diagnosis, the accuracy of string current is one of the key factors to determine the final failure diagnosis accuracy and indirectly determine the power generation efficiency. This is very important to a commercial-industrial PV plant and utility PV plant which cares a lot about output efficiency.

Figure 3-1 An 320kW String Inverter Example With 6 Strings per MPPT

Usually, a PV array consists of 2 or more PV strings connected in parallel to one MPPT. While there is a trend that more strings are tied together to one MPPT to cost down in high power string inverters. Figure 3-1. shows a 320kW string inverter example with 6 strings per MPPT. Note that the current sampling for string 6 is optional, because it can be calculated from MPPT boost average current subtracting other string currents. So, with more than 2 strings tied to one MPPT and using this subtraction method, the error of the last string current can be very large if errors of other string current deviate in the same positive or negative direction. That's why the last string current sampling is optional to add.

For MPPT boost current sampling, average inductor current is usually sampled and MPPT control frequency is much lower than the switching frequency. The accuracy of MPPT boost current sampling is also critical because this determines the MPPT accuracy which ultimately affects the power generation efficiency.

For inverter 3-phase current sampling, this includes AC current and the corresponding DC component. For grid-connected inverters, theoretically, only AC current is allowed to be injected into the grid. But in fact, inverter output current inevitably contains some DC component which does harm to the grid, grid load and grid equipment. Therefore, it is impractical to completely remove the DC componentof the inverter but it does need to be controlled under a certain low range. Standards such as IEEE 1547-2018 have defined the limit for DC component in the grid-side AC current. For example, below 0.5% of the rated output current.

So, the accuracy of phase current sampling is important for inverter power stage control, power generation statistics and DC component suppression. Especially for DC component excess issue, using hall-effect current sensor with high accuracy and low drift can help to solve the issue at the beginning.

Another issue regarding the accuracy of current sensors is reactive power generation. For active power generation, the reference of current loop is generated by the voltage loop. The error of current sensor is greatly alleviated by the current controller, in this case, the accuracy of DC bus voltage sensing is important. But for reactive power generation, the reference of reactive current is generated directly by the MCU. So, if the current sensor is not accurate, the output current of the inverter is not the set value. Using a high accurate TI Hall-effect current sensor helps with this problem.

For neutral current sampling in hybrid inverter, the fourth leg (neutral) current sampling is used to actively control the midpoint voltage that allows the inverter to support unbalanced output. Though the neutral current is not large as the phase current, the same current sensor is usually selected from a simple design perspective.

For BDC current sampling, it's not only used for control and protection purposes. In some designs, it's also used for battery power statistics to align with BMS (Battery Management System) data. A 1% accuracy target is required.

For EPS (Emergency Power Supply) 3-phase current sampling, this is different from inverter 3-phase current sampling. Theoretically, EPS 3-phase current sampling is not used for power stage control, and current sampling does not need to consider DC component suppression because this is for the backup loads and doesn't do harm to the grid, grid load and grid equipment even out of range. Typically, 3% accuracy is sufficient. However, if this is also used for backup loads power statistics, using Hall-effect current sensors with high accuracy and low drift benefits the metering accuracy and reliability.