Is Solar the Next Wind?

Continuing our series on integrating renewable energy into today’s power systems, we now turn our attention to solar energy.  Solar is another growing renewable resource for electric power.  Total installed capacity worldwide powered by either solar panels or solar reflector installations is estimated at 10,000 MW as of 2005, with the largest capacities in Japan, Germany and the United States.  About 80% of this capacity is grid-connected.

There are numerous conversion technologies currently available or under development.  The primary ones used for bulk power are:

  • photovoltaics (PV)- use the photovoltaic effect of semiconductors to generate electricity directly from sunlight.  A variant, concentrated photovoltaics, uses lenses to concentrate sunlight into a smaller panel of high efficiency PV
  • thermal – use solar heat to run a steam turbine or an induction generator.  For further discussion, we’ll refer to the steam turbine system by its commercial name – SEGS for the solar energy genrating system developed by Solel, and to the induction generator as SES for Stirling Energy Systems, developer of that design.

The largest existing or under construction PV installations are on the order of 10-11 MW (Bavaria Solarpark in Germany, GE solar plant in Serpa, Portugal).  The largest thermal installations of SEGS are located in the Mojave Desert in the United States where plants ranging in size from 14-80 MW provide a net capacity of 345 MW.   Two installations of thermal solar using SES have recently been proposed in Southern California with sizes of 500 and 300 MW.

With this type of growth and advancement, one wonder if solar is the next wind?

What is unique about solar fields?

Solar energy conversion technologies have many similarities with wind power (see “The Renewables: Part 1 – Wind Farms,” Techblog of May 2006).  Like wind, solar increases the delivered electric energy by terrain and acreage.  In the case of solar, deserts are an ideal location.  The hotter the better.  And like wind, to obtain power levels above 50 MW requires a farm or park of many individual collectors interlinked by power lines and cables.  Ultimately, the lines connect to substations which step up the voltage to transmission levels and interconnect with the power grid.  Now, the actual energy conversion from solar to electric energy is what makes this resource unique.

  • With PV, solar panels convert the sun’s rays into direct current (DC).  The DC is passed through inverters to obtain alternating current (AC) than can be stepped to higher voltages and shipped over long distances [See Ref 1].
  • A Stirling DishSEGS design has been in operation since 1984 and involves a hot fluid loop to collect thermal energy and drive a conventional steam turbine [See Ref 2].
  • SES is more recent, although the technology for the Stirling engine is nearly 200-years old.  The system uses dishes to concentrate sunlight onto a fluid loop.   The same fluid then operates an engine to produce electric power.  The design provides for silent, no-emission operation [See Ref 3].

A summary of solar energy characteristics:

  • Output Predictability: The output from a solar field depends primarily on the prevailing solar output.  This would be ideal support for day peaking systems, and a stark contrast to wind power which peaks at night and can be still when the day is at its hottest.  Collectors take into account the position of the sun over the course of the day and through the seasons.  Output is only available during daytime.  This might be extended a short extent into nighttime with the help of batteries or energy storage systems.  Solar presents a planning challenge in how to account for the energy available for this resource when looking ahead 2-5 years medium term or 5-10 years long-term.

  • Variability: Cloud cover and weather patterns add a second layer of complexity to predicting delivered power.  This presents an operating challenge to integrate the solar output with the overall energy portfolio for electric supply.  More and more, storage energy options are seen as a means to provide operators greater flexibility in meeting cyclical load demand.

  • Voltage Control.  If inverters or induction generators are used, there is generally limited reactive power control.  For low voltage ride-through (LVRT) , inverters can be provisioned with battery energy systems (BES), rotating energy from flywheels, super magnetic energy storage (SMES) or fast reactive sources such as static var compensators (SVC).  Special design induction generators can be equipped to withstand the mandated LVRT capability, with additional protection from voltage and frequency relays.  A steam turbine moving a generator with sufficiently capable excitation control will not generally have a problem with respect to voltage control.

System Impact Concerns

The technical concerns with integrating solar fields are:

  • Steady-state – the solar field must meet reliability standards for thermal loading and voltage during normal operations and under credible contingencies.  The analysis may include any impacts on thermally or voltage limited interfaces.  Mitigation may include transmission reinforcement, capacitor addition, remedial action schemes and other transmission planning solutions.
  • Fault levels – the solar field should demonstrate mitigated impact on short circuit levels.  Inverter technology does not contribute significantly to fault levels.  Thermal solar conversion with wither steam or induction generators introduce increases to the fault levels that have to be checked against existing and planned withstand capacity.
  • Stability – the solar field must meet reliability standards for acceptable levels of oscillations, damping, transient stability and long-term stability when integrated with an existing grid.  The tests cover normal, emergency and and extreme contingencies with varying levels of acceptable response.  The plant will be required to stay online during close-in normal faults and meet the mandated LVRT capability.
  • Switching – If mitigation includes switched capacitors, the size should not exceed allowable delta-V at the load and transmission buses.
  • Power Quality – The operations of inverters may introduce voltage flicker and harmonic distortion into the system.  The level should be kept within acceptable criteria defined by standards.  Mitigation may include filters and static var devices.
  • Subsynchronous Resonance – Inverters may introduce excitation currents that may excite resonant frequencies in closeby steam generators.  Torsional analysis may be needed to demonstrate potential impacts from torsional stresses.

Modeling Issues

Solar fields using PV or SES designs would electrically operate as distribution systems with line voltage in the range of 4-69 kV for connections between collector stations.  For bulk power studies, the collectors can themselves be aggregated into equivalent collectors in terms of how the transmission grid would view them electrically.  The modeling challenge is to figure out the characteristics of the various aggregates for each of the analytical tests – steady-state power flow, short-circuit, switching, stability, etc.

Solar fields based on SEGS would appear as small to medium-sized steam generators to the power grid.  Steam generators are familiar ground to most planners and analysts, and for the purposes of this discussion, SEGS are excluded from the Modeling Issues coverage.

Power Flow Modeling

For power flow modeling, each collector aggregate is represented by the combined real power generation of the individual inverters or induction generators.  Reactive power is modeled depending on the reactive control strategy, the typical being application of static reactive compensation to bring the net power factor to unity.  The impact of energy storage would be to provide some range in the reactive power in the steady-state.

Like wind farms, solar fields are also susceptible to voltage drop on certain line out contingencies.  Power flow solution algorithms may diverge in these cases.  Among techniques to confirm that the issue is a physical rather than a numerical issue are:

  • Adding a fictitious synchronous condenser
  • Changing solution method – a full Newton solution that recalculates the Jacobian every iteration would work best in this case
  • Changing the solution acceleration factor (in PSLF, the reactive solution factor may also be changed)
  • Changing location of the swing bus
Modeling for Short Circuit and Switching Analyses

If inverters are used, there is no contribution to short circuit.  Induction generators, on the other hand, would inject fault currents in accordance with the magnitude of the machine reactance.  This could be either the subtransient (X”d) or transient (X’d) reactance.  Since the generator may represent an aggregate, the reactance used would be the same value if all the generators have the same characteristics on a machine base which is the sum of the MVAs of the individual machines.

For switching analysis, a more detailed model is required.  Avoiding aggregation is the first consideration.  The impedance of overhead lines and cables, and transformers would need to be modeled explicitly.

Modeling for Dynamic Simulation

First, let us deal with solar fields that use inverters.  The inverters have controls that may be able to adjust the real and reactive power output of the solar farm depending on voltage and frequency at the collector terminals.  In addition, there may be additional equipment, noted earlier, that may be included in the installation to meet LVRT requirements, provided extended operating life or to condition voltage.  The aggregate response of the inverters and additional equipment would need to be modeled specifically for each design.  (For further information on special model development, see “Developing My Dynamic Model,” by M. Gutierrez, Techblog of March, 2006.

On the other hand, for solar fields that use induction generators, the model for dynamic simulation is based on an equivalent induction generator which represents all the induction generators in a collector bus.  If all the generators are the same , then the electrical characteristics of one unit can be used with an MVA base equal to the sum of MVAs of the individual units.  The controls of the solar dishes may show some dependence on the voltage and frequency of the collector terminals, and these would have to be included in a detailed special model.

Since the PV and SES conversion technology for solar power results in an asynchronous device (inverter or induction generator) conventional aspects of generator performance related to internal angle, excitation voltage, and synchronism are not applicable.  We would have to infer the impact of the solar farm from the rotor swing curves and damping simulations of neighboring synchronous machines.  Another indicator of solar farm dynamic performance would be the voltage and frequency at the terminals, especially the LVRT capability.   Note that though it would be tempting to represent the solar field as a standard induction machine, its inherent controls would shift response away from this assumption.

Conclusions

Solar fields are a growing component of today’s power systems.  Conversion technologies are evolving rapidly.  Planners, operators and designers would need to keep abreast of developments to anticipate integration and operating issues for these new resources.  Specifically, new analytical methods and models may be required to ensure that their specific response characteristics are captured accurately for simulation and study.

The main challenge is whether we can be pro-active to the changes that solar energy in its many forms of conversion technology and design will most certainly bring to the table.

For questions, comments and further discussion, contact us at mailto:info@pterra.us

References

  1. Bavaria Solarpark“, Powerlight website
  2. Solar Fields for Utility Power Plants,” Solel website
  3. What is a Stirling Engine?” Stirling Energy Systems website.
  4. The Basics of Solar Power for Producing Electricity,” Advanced Energy Group website