The Application of Lessons Learned from DC Microgrid Technology for Spacecraft to Terrestrial Microgrids

Authors: Fred Steinhauser, OMICRON electronics GmbH, Austria, and Benton Vandiver, OMICRON electronics Corp. USA

Similarities of DC Microgrids on Spacecraft to Terrestrial AC Microgrids

DC microgrids are not new.  They have been powering spacecraft for over 55 years and have been powering manned craft operating in hostile environments for over 100 years (the first submarines utilized DC microgrids). When operating in deep space, the importance of 100% reliability makes it necessary to have extraordinarily robust hardware and software to monitor and control the spacecraft DC microgrids.  Previous manned space missions have involved “near earth orbit” travel to the Moon and the International Space Station (ISS) where telecommunication latency is manageable by thousands of ground personnel at Houston Space Center, Cape Canaveral and other tracking stations around the world.  The premise for manned space travel to MARS is that telecommunication latency of up to 45 minutes is not acceptable for the reliability of the mission.  Therefore, the management of the onboard DC microgrid power management and distribution (PMAD) system must be robust, autonomous and fault tolerant.    

Solar arrays, batteries and DC-DC converters make up the majority of components powering the DC microgrid on board spacecraft.  These components have non-linear characteristics which can lead to voltage collapse if there is a loss of a key component or if the solar arrays, batteries and converters are operated in an unstable region of the system curve. The electric loads on board the spacecraft are constant power loads. This type of loading exacerbates a component failure scenario where a single failure can cause a larger voltage drop and resultant higher losses that potentially cascade into a complete voltage collapse. Terrestrial AC grids and microgrids have similar constant powerloads as reflected back into the high voltage transmission grid and are susceptible to voltage collapse under peak load conditions particularly where the reactive (I2X) losses on the transmission grid are high.  Figure 1 shows a diagram of spacecraft power system architecture.

Power Flow Analysis of large and medium scale terrestrial AC transmission grids has been performed by various software programs the first being Gauss-Seidel (Ward-Hale, 1956) and the second full Newton-Raphson (Tinney-Hart, 1967) to analyze the state of the transmission grid on the terrestrial grid.   While these programs have provided excellent analysis over the years, they have had known problems solving power flows near the limits of voltage collapse because they rely on iterative methods and are dependent on the quality of the initial starting point.  In the past three decades terrestrial transmission grids have been moving operationally closer to the voltage collapse limit for a variety of reasons including the retirement of older fossil fuel power plants near load centers.  As the trend toward operating closer to the voltage collapse limit continues, the ability to analyze the grid and provide command and control signals in real-time has become extremely important.

This mirrors NASA’s concerns regarding voltage collapse limits for the DC microgrid.  As part of a Phase I project to support NASA’s Mission to MARS program, Gridquant Technologies and Battelle Memorial National Laboratories proposed using a holomorphic embedding load flow (HELMTM) algorithm invented by Dr. Trias in the late 1990s.  The HELMTM algorithm is an extremely robust deterministic power flow that can solve at the point of collapse regardless of the initial starting point which is crucial for autonomous control of DC microgrids.  The Gridquant/Battelle team demonstrated the capability of HELMTM by comparing the solution of a 300 bus IEEE AC model with HELMTM power flow versus a Newton-Raphson power flow.  The 300 bus IEEE AC model was chosen because it represented the complexity of a DC microgrid on board a spacecraft.  Figure 2 presents a Sigma curve from HELMTM showing when load is added to Bus 528 of the 300 bus IEEE model, there are several points near the boundary of the Sigma curve indicating conditions near voltage collapse.
While this is not an actual AC transmission grid, it demonstrates that many transmission buses are very close to voltage collapse and if additional load were added or loss of a single transmission element were to occur, the transmission grid would collapse.  The deterministic holomorphic embedding load flow demonstrated the capability to solve right up to the point of collapse and can be utilized in a proactive way to measure the electrical distance to collapse from any steady state point on the AC transmission grid. (Figure 3).

Similarly, voltage collapse is a major concern on a DC microgrid due to the nonlinearity of dc-dc converters, solar generators and batteries.  An additional complication on board spacecraft are solar eclipses. These may happen abruptly and when they occur the DC microgrid must be able to rely solely on storage devices such as batteries and flywheels to maintain voltage stability.  During these periods the DC microgrid must be prepared to shed load if a critical voltage limit is approached.  The loads in spacecraft may be categorized into the following types: essential; semi-essential and non-essential.  Non-essential loads such as crew off-duty loads will be shed first to maintain voltage stability.  Of note is that both AC and DC microgrids use special protection systems (SPS) to maintain stability.

The HELMTM algorithm was adapted to model the nonlinearity of DC microgrids on board spacecraft.  The unique non-linear characteristics of solar arrays and DC to DC Buck Converters presents challenges for modeling that are not present in the AC terrestrial grids.  While large scale solar PV projects are connected to terrestrial AC grids, the DC to AC inverters and interposing transformers mitigate the effect of the non-linearity.  Figure 4 illustrates the schematic of the solar-array-buck-converter power system for spacecraft:

The system equations for the spacecraft are defined by sets of differential algebraic equations where the equilibrium of the system can be found by setting the derivatives of all the state variables to zero.  Finding the equilibrium of the simplified spacecraft is nontrivial.  Figure 5 demonstrates there are a number of sets of equilibrium points of which one set is in the desirable and stable region.
The DC-based HELMTM power flow demonstrated on the simplified spacecraft power system the ability to find the desirable and stable region for equilibrium. 

Power. Flexible. Easergy.
BeijingSifang June 2016