To help achieve these goals, add greater intelligence to the control of the inverter, which converts the variable voltage output of the collector into a steady voltage that is used for running applications or charging batteries.
Although sun- and wind-powered systems are the obvious applications, intelligent inverters can also benefit other sources of power, such as fuel cells, in order to maximise output. For these applications, highly effective inverter control is available from digital signal controllers (DSC), which have been shown to cut conversion efficiency losses in half while significantly reducing costs. DSCs combine the high performance of digital signal processors (DSP) with the programming ease and integration of a microcontrol unit (MCU). Furthermore, DSCs with floating-point capabilities are now available, enhancing performance and making the job of programming complex algorithms easier.
The inverter's role and stages
The inverter is responsible for the control of electricity flow between the module, battery and loads. Its main function is to convert variable dc voltage input from the source into a clean sinusoidal 50- or 60-Hz current, depending on the region, for use by appliances and feeding back into the grid.
Other functions inversters perform include disconnecting the circuit to protect it from power surges, charging the battery, logging data on usage and performance and maximizing power point tracking to keep power generation as efficient as possible. Nominal power ranges between one and several hundred kilowatts peak (KWp), allowing inverters to be designed around sophisticated source topologies, either with or without transformers, and with the integration of multiple control processors. Fig. 1 shows where the inverter fits into an all-inclusive photovoltaic (PV) system that not only charges a battery and drives local ac loads, but also ties to the grid and has an alternate power source in the form of an ac generator.
The blocks of the essential inverter circuit are shown in the top portion of Fig. 2. First, dc/dc conversion raises or lowers the incoming voltage, adjusting its output for greatest efficiency. After some additional voltage buffering, MOSFETs in a bridge use a switching frequency, usually between 18 to 20 kHz, to convert the dc to an ac voltage. Finally, a low-pass filter smooths the switched ac to a sinusoidal signal for use in generating a grid-frequency ac output. (The figure does not show the dc/dc conversion and regulation that are required for battery charging.)
Transformers and protection
At some point, the dc has to be at a higher voltage level than the ac output, but the source input is usually not that high. The system can therefore either step up the voltage with a transformer on the ac side or boost it in the dc/dc conversion stage. Just as an ac transformer inherently provides galvanic isolation, so does a phase-shifted full-bridge dc/dc converter with zero voltage switching, thus making the latter equivalently a transformer. Figure 3 shows a commonly used dc/ac circuit with transformer for single-phase inversion, based on an H-bridge configuration controlled by four pulse-width modulated (PWM) signals.
Requirements for transformers vary by region. Transformers add weight, bulk and cost, and they also cause a about a two percent reduction in efficiency. However, they increase circuit protection and human safety by isolating the two sides of the circuit electrically, preventing a dc fault from flowing to the ac side, and an ac leakage current from developing a potential issue between PV panels and ground. The design may include a residual current protection device (RCD) that monitors the currents of all phases, and trips the relay if the current exceeds a certain value. RCDs are especially important for safety in transformerless systems due to the risk of current leakage.
Protection of the system mandates inclusion of a relay to protect the conversion and charging circuitry against voltage surges and spikes on the grid. If a power line is damaged or the utility has to shut down, the inverter needs to stop feeding out electricity to the utility. A 'non-islanding' inverter senses that the line has been de-energised, is under- or over-voltage, or has a significant disturbance for whatever reason. When this happens, the inverter automatically disconnects from the utility grid, thereby not becoming an electricity generating 'island'.
Obtaining maximum charging power
Just as the efficiency of the dc/ac conversion depends on the input voltage, so does the battery charging. But the input voltage is variable, depending on wind conditions for a turbine, or season, cloud cover and time of day for PV panels; and the battery conditions vary, too, depending on the charge state. Maximum power output to the battery occurs when the product of voltage and current is at its peak, the maximum power point (MPP). MPP tracking (MPPT) is designed to determine this point and adjust the dc/dc voltage conversion in order to maximize the charging output. Fig. 4 shows how the determination of MPP can vary with different conditions.
The most common algorithm for determining MPP is for the controller to perturb the panel's operating voltage with every MPPT cycle and observe the output. The algorithm continues oscillating around the MPP over a wide enough range to avoid local but misleading peaks in the power curve caused by movement in cloud cover or a brief wind lull. Alternatively, the incremental inductance algorithm solves the derivative of the power curve for 0, which is by definition a peak, then settles at the resolved voltage level. While this approach does not have the inefficiency caused by oscillation, it risks other inefficiencies because it may settle at a local peak instead of the MPP.
A combined approach maintains the level determined by the incremental inductance algorithm, but scans at intervals over a wider range to avoid selecting local peaks. Although this approach is the most efficient, it requires the greatest amount of performance on the part of the controller.
Control design requirements
A certain number of real-time processing challenges must be met by the control processor to effectively execute the precise algorithms required for efficient dc/ac conversion and circuit protection. For non-islanding requirements, the accurate measurement of voltages and currents is necessary to determine the flow of power, enabling fast disengagement.
MPPT and battery charge control, while only needing near-real time response, also involve algorithms with a high level of processing.
A single device that can control all stages and has sufficient performance for multiple algorithms allows the designer to deal with these issues.
DSCs offer a good solution for real-time control of inverters, batteries and protective mechanisms in renewable energy systems. These DSP-based devices inherently support high-speed mathematical calculations for use in real-time control algorithms.
A single DCS can control multiple conversion stages in the same inverter and have overhead remaining for performing additional functions such as MPPT, battery charge monitoring, surge protection, data logging and communications.
New floating-point controllers extend these advantages, making programming and debugging easier and less error-prone. The greater range inherent in floating-point operations makes saturation less likely and allows dynamic algorithm adjustment across all load conditions.
Furthermore, floating-point code is more compact in math operations and requires fewer cycles than fixed-point.
A floating-point DSP controller
An example of a DSC with floating-point capabilities is the Texas Instruments (TI) TMS320F2833x DSC. This 32-bit device operates at frequencies up to 150 MHz and provides up to 300 MFLOPS (mega floating-point operations per second). Integrated features include on-chip direct memory access (DMA), fast interrupt handling, a 32-bit enhanced memory interface (EMIF), ultra-fast 12-bit ADCs supporting up to 16 input channels, multiple timers, standard communication ports, and 12 individually controlled enhanced PWM (ePWM) channels, each with its own timer and phase register. Fig. 5 shows the F2833x architecture.
The F2833x differs from its predecessors in the F28x generation because it's based on a floating-point architecture.
With a single sign bit, 8-bit exponent and 23-bit mantissa, the device handles a normalized value range of approximately +/-~1.7e-38to+/-~3.4e+38. For control tasks in particular, the extensive range is valuable in that it deals with scaling and saturation more efficiently than fixed-point does.
With integrated flash memories, F2833x controllers enable straightforward program development and updates while helping to protect intellectual property.
To simplify design upgrades, the device runs fixed-point code ported from other F28x devices, and the C compiler can compile either floating- or fixed-point code with a simple switch. Programmers can enjoy the faster development that comes with floating-point, then let the compiler make the adjustments necessary to change the code to fixed-point.
Inverter control design
Fig. 6 shows the F2833x DSCs used to control the power stage inverter in a solar system. Inputs from sensors in the panel array are fed to the controller's ADCs to provide data on the instantaneous voltage and current available from the array for conversion.
Inputs may also provide information such as cell and ambient temperature, used to protect the panel, and feedback that meters power output from the cells, used to track MPP. All sensing inputs must be scaled so that peaks and spikes do not exceed the 3.0V level of the ADCs.
The data is first fed into a power control loop, and there may be more than one loop, depending on the design.
Other real-time tasks that are being performed also provide inputs to the power control loop.
Among these tasks are metering power returned to the grid, monitoring grid power levels for protection, regulating battery charging, tracking MPP, and communicating with parallel controllers handling other systems.
Real-time fault management is an important aspect of the design. A fault that occurs relatively slowly can be detected and managed using a dedicated ADC input, which monitors the temperature and initiates the appropriate system response.
By contrast, a critical fault such as over- or under-voltage or over-current requires immediate response to avoid severe system damage.
To address these critical faults, the F2833x provides dedicated fault lines called trip zones, which deactivates mapped PWM outputs within a couple of DSP cycles after receiving a fault signal.
A key to renewable energy
Renewable energy systems are continually being improved to achieve greater efficiency that will lower the cost per kW.
Variability in regulatory and operational requirements makes it important to select the right controller for the inverter, a controller that provides high performance, integration and flexibility.
DSCs have demonstrated their value in inverter systems, and new floating-point controllers will continue to help improve the efficiency and lower the cost of renewable energy.
Arefeen Mohammed is a C2000 Applications Engineer Texas Instruments, Houston, Texas, USA. www.ti.com.
