For the past five years, I have been involved in research to determine whether hybrid propulsion technology represents a viable alternative to conventional marine propulsion systems. I’ve come to the conclusion that the answer is a more or less unqualified “yes”—but not necessarily for the efficiency reasons cited by most proponents.
Click Here To Read Hybrid Science: Part 2
In this first part of a two-part series on the topic, we will define a framework for gauging the efficiency of hybrid propulsion. In the next issue we will see how this plays out in the context of hybrid propulsion systems, and then we’ll broaden the discussion to look at other reasons for installing a hybrid system.
First, we need a metric for assessing efficiency. I am going to use specific fuel consumption (SFC), a measure of how much fuel it takes to create each unit of energy delivered by an engine. SFC typically is expressed in terms of grams per kilowatt-hour (g/kWh) or pounds per horsepower-hour (lb/hph). It is analogous to what you see on your electricity bill every month, except that instead of the cost per kilowatt-hour of energy you have used, we are measuring the fuel consumed in producing each kilowatt-hour.
To measure the SFC at all speeds and loads, several conditions must be met. The engine must be operated from idle speed to full speed, and from no load to the full load it will support at any given speed. Fuel consumption must be measured at all times, and then, if you take the fuel consumption at any speed and load and divide it by the load, you will derive the SFC at all speeds and loads. This can be displayed in the form of a fuel map, with torque or power on one axis of the graph and engine speed on the other.
Much of my testing was done using a Volvo Penta D2-75 engine. The fuel map for this engine is shown in Figure 1. Peak efficiency of 230 g/kWh occurs at around 1800 rpm and 28kW of load (as measured at the flywheel). Total fuel consumption at any point on the fuel map equals the load at that point (e.g., 28kW) times the SFC (e.g., 230 g/kWh): 28 x 230 = 6,440 grams per hour. The standard weight of diesel is 840 grams per liter; this converts to a burn rate of 6,440 grams per hour/840 grams per liter which equals 7.67 liters per hour (around 1.9gph).
What happens when we use this engine to turn a propeller? We can plot the power required to spin the propeller at any given speed in the form of a curve on the fuel map. From this we can derive the SFC at any point on the propeller curve. The manner in which a propeller absorbs energy is such that these curves are almost always concave and never mimic the full-load curve for an engine. As a result, the engine full-load curve and propeller curve can only be made to come together at one point. Typically, a propeller is sized such that this concurrence occurs at, or close to, full engine speed and load (the matched propeller in Figure 1).
If the propeller is undersized, the engine will reach its full speed before the propeller fully loads it—the engine will never be loaded to its full rated power. If the propeller is oversized, it will fully load the engine before the engine reaches its full rated speed. The propeller curve will cross the engine’s full-power curve before the engine reaches full speed, and the engine (and transmission) will be overloaded. Sustained operation at this point on the propeller and power curves is likely to result in damage.
It is apparent that at no point on any of the propeller curves is the engine operating at its peak efficiency of 230 g/kWh. It is this failure to load the engine to peak efficiency that defines the primary window of opportunity for a hybrid system. Note that at any engine speed or power level, the oversized propeller is operating in a more efficient part of the fuel map than the matched propeller, which in turn is always operating in a more efficient part of the fuel map than the undersized propeller.
Finding The Crossover Speed
During the course of our project, we ran a series of tests with multiple propellers on a 48-foot, 18-ton sailboat. We recorded SFC versus boat speed (Figure 2). As expected, in general the lowest SFC for a given boat speed (that is, the most efficient engine operation) occurred with the oversized propellers, and the highest SFC for a given boat speed occurred with the undersized propellers. The matched propeller curves were in the middle of the pack. All of the propellers came closest to loading the engine to peak efficiency at higher boat speeds (higher loads), but none loaded the engine to its peak efficiency of 230 g/kWh.
If we assume a hybrid system can be designed such that the engine always runs at peak efficiency, the fuel efficiency gains are represented by the area under the propeller curves down to a line drawn at peak efficiency (in this case, 230 g/kWh). This defines the principal window of opportunity for a hybrid propulsion system.
In practice, as long as the power for the hybrid system is derived from running an engine—in Part II we will look at situations in which this is not the case—even if the engine in the hybrid system is run at peak efficiency, there are additional losses that must be taken into account. In both serial and parallel systems (see Two Flavors: Serial & Parallel) the engine drives a generator that provides power to an electric motor, either indirectly via the batteries or directly. There are losses in both the generator and electric motor. If the energy is stored in batteries before use, there are additional losses during the charge and discharge cycles.
For a number of reasons, we theorized that the engine would, in practice, deviate from peak efficiency at times but that we would be able to hold it to within 5 percent of peak efficiency. This raises the SFC for the peak efficiency in our test system to 230/0.95 = 242 g/kWh. We surmised that we could build a generator with an electrical efficiency of 90 percent over the necessary power range for the system. This raises the SFC to 242/0.90 = 269 g/kWh.
We similarly surmised that we could build an electric motor with an electrical efficiency of 90 percent over the necessary power range for the system. This raises the SFC to 269/0.90 = 299 g/kWh. All this represents system operation in diesel-electric mode. We used thin plate, pure lead (TPPL) batteries, a variant of AGM batteries capable of supporting higher charge and discharge rates than conventional AGMs, and we surmised combined in and out (charge and discharge) losses of 15 percent. This raises the SFC, when in battery-powered mode, to 299/0.85 = 352 g/kWh.
Given these numbers, we discover that regardless of the propeller used, the hybrid systems are always more efficient at low boat speeds, but at some point the conventional system becomes more efficient. We defined this point as the crossover speed (see Figure 3). The more efficiently a propeller loads the engine in a conventional system, the lower the speed at which the crossover occurs; conversely, the less efficiently a propeller loads the engine in a conventional system, the higher the speed at which the crossover occurs.
If the energy in a serial system comes primarily from the generator, and thus from the engine driving the generator, the only way the system will be more efficient than a conventional installation is if the boat is operated most of the time below the crossover speed. A parallel system is different because the engine is still connected to the propeller shaft. Given a sufficient knowledge of the operating characteristics, and a sufficiently sophisticated control mechanism, the system can be designed in a way that uses electric power only at those times when this mode of operation is more efficient than conventional operation. In this case, the system will never be less efficient than a conventional installation.
In Part II of this series, we will see whether hybrid systems are more efficient than conventional ones in the real world, and we’ll explore some other critical components of the whole-boat energy equation that come into play in hybrid propulsion.
The Two Flavors of Hybrid: Serial and Parallel
There are two architectures for marine hybrid propulsion systems. One is known as a serial system, the other as a parallel system.
At the present time, both types include an internal combustion engine. The difference lies in the relationship between the engine and the propeller that moves the boat. In a serial system, the engine drives a generator, which powers an electric motor connected to the driveshaft—there is no mechanical connection between the engine and the driveshaft.
A parallel system, has a direct mechanical connection between the engine and the driveshaft (as in a conventional installation), with an additional electric motor operating on the same driveshaft. The two propulsion systems operate in parallel on the same shaft. The electric motor in a parallel system can also be driven by the engine as a generator.
Almost all automotive hybrids are parallel systems, the best known being the Toyota Prius. The sole exception, more or less, is the Chevy Volt, which has a serial system. On the Volt, an electric motor powers the wheels until the batteries are discharged, at which point the engine fires up and powers a generator that maintains the supply of electricity to the electric motor. In marine applications, the engine is almost always a diesel; hence, this mode of operation is known as diesel-electric.
The primary goal in both systems is never to run the engine at anything other than peak efficiency, and at all other times to use stored electrical energy (which, in practical terms, means batteries, until we get effective fuel cells). The batteries perform another critical function: acting as an energy balancer. If, for some reason, the engine has to be run at less than optimal load, the batteries are charged at a rate that applies whatever additional load is necessary to maintain the engine at an optimal load.
For example, if a serial system is running in diesel-electric mode with a propulsion load that is not high enough to load the generator to peak efficiency, then the batteries are charged at a rate that loads the engine to peak efficiency. When the batteries are fully charged and can no longer absorb this level of charging current, the generator is shut down, and the batteries are used for power until their charge acceptance rate is high enough to once again enable the generator to run at peak efficiency.
In a parallel system, if the engine is being used for light propulsion loads, just as in the serial system, the batteries are simultaneously charged at a rate that maintains the engine at peak efficiency. If the batteries cannot accept this level of charging current, the engine is shut down, and the boat is run under electric power. Note that whereas the serial system can sometimes run in diesel-electric mode, eliminating the energy losses inherent in using battery power, when under electric power the parallel system must always use the batteries and pay the associated efficiency losses.
Because the engine in a serial system is not connected to the propeller shaft, the generator and electric motor have to be powerful enough to deal with the highest propulsion loads anticipated. In a parallel system, the engine is still connected to the propeller shaft and as such can handle high propulsion loads (at which point it is reasonably efficient), with the electric motor downsized to handle light propulsion loads.
The batteries in both serial and parallel systems are used to store energy from other sources (shorepower, solar power, wind, etc.) and to absorb regenerative energy, if available—for example, the energy created by a freewheeling propeller on a sailboat under sail.
