Motor Mode Analysis of Exterior-Rotor PM Machine with Gramme ’ s Winding

This paper presents the PM electrical machine with the novel stator geometry, which is intended for a small unmanned aircraft. The PM machine is a part of a hybrid propulsion system operating in conjunction with an internal combustion engine. The main study involves operations in the motor mode, where the PM electric machine performs as the main propulsion, as well as the additional power source or a starter in the hybrid drive. The purpose of the investigation is experimental evaluation of the electric parameters of the PM electric machine in the motor mode; in particular, achievable torque at different speeds on several loads, to compare the results with the initial requirements. DOI: http://dx.doi.org/10.5755/j01.eie.23.1.17579


I. INTRODUCTION
Today small unmanned aerial electric vehicles (UAV) are used for a wide range of purposes, like wildlife tracking, traffic monitoring, border patrols, environmental research, geographical mapping, numerous military applications and several more [1].The reason for the increasing trend in the use of UAVs is their simple construction, reliability, easy operation, relative safety and environmental friendliness.The market of electric UAV market has grown into a multibillion dollar industry [1].
Electric machines (EM) are also used in the UAVs that use an internal combustion engine (ICE) as their main source of power.In the named combination EMs are used mainly as generators.Lately hybrid propulsion systems (HPS), which have proven itself in the land and in the maritime transport [2] have started to expand into the field of UAVs.An HPS is generally an integrated single system, which uses two or more different power sources to move the vehicle.ICE's, also electric motors, have certain strengths and weaknesses.An HPS combines the strengths of both types of motors [2].
Nowadays interior, exterior and axial rotors are mainly used in electric-powered vehicles (Fig. 1) [3].In this paper, the exterior-rotor geometry is selected.The novelty of the research lies in implementation of a new type of stator geometry in the exterior-rotor permanent magnet synchronous machine (ERPMSM) in combination with Gramme's or toroidal winding, also in usage of this kind of ERPMSM in a motor, as well as in a generator operation in an HPS for a small-size UAV.
The principle scheme of the novel HPS is shown in Fig. 2.
Fig. 2. Principle scheme of the UAV hybrid propulsion and energy storage system [4].
components into the internal opening of the stator, thus obtaining more compact solutions.For ERPMSM it is the mechanical clutch, which connects and disconnects ICE from it, which is rather complicated with the axial and interior-rotor design [4].
The other advantages of the exterior-rotor topology comparing for example with the interior-rotor design are higher torque density, better cooling possibilities, reduced copper loss, lower weight and cost [1].
ERPMSM solution is novel in its geometry and functionality as the same EM can be used as a motor and a generator.The operation in the generator mode was discussed already in the previous phase of the research [5].In this study the main focus is on the ability of ERPMSM to work as a motor.This includes its capability to function as a primary engine which directly rotates a propeller, also operates in parallel with ICE as a starter or giving extra power during UAV's take off.
During the research the prototype of ERPMSM was also constructed (Fig. 3).Fig. 3.The section geometry of the exterior-rotor motor-generator with the toroidal winding.

II. INITIAL PARAMETERS
One of the first initial design parameters for the HPS was the range of rotational speed between 3000 rpm to 4000 rpm, which meets the efficient operational regime of the propeller at UAVs stable flight situation.
An additional important nuance considered when ERPMSM functions as a motor, was the ability to output enough torque at required speeds of rotation while driving the propeller whilst maintaining the desired thermal parameters.
Various supplementary design parameters were given from ICE technical data, including the starting torque that ERPMSM had to overcome.In addition, manufacturer's requirement for the UAV is to be able to maintain stable flight for minimum of 15 minutes on electric power.

III. LOAD TESTS
During load tests of ERPMSM in motor regime, UAVs actual wooden propeller with the diameter of 508 mm and pitch of 355.6 mm was used as a load [7].
In Fig. 4, the counter torque of the propeller Tpr characteristic in dependence of rotational speeds is shown.ERPMSM load tests in motor operation were conducted in delta configuration to achieve more power at lower rotational speeds.Three primary supply voltages were used, UDC = 14.8, 18.5 and 20.9 V. UDC = 14.8 V was considered as the minimal voltage at which the rotational speed n = 3000 rpm was guaranteed by the specified propeller.
Voltage UDC = 18.5 V was chosen due to the nominal voltage of 5S-type batteries, which was achieved at the above-mentioned rotational speed in ERPMSMs generator mode in star connection [4].The chosen voltage level of UDC = 20.9V is equivalent to the voltage output of the battery at fully charged state, at which rotational speed n = 4000 rpm was reachable.
While loading ERPMSM with a propeller, using nominal battery voltage UDC = 18.5 V the rotational speed of nearly n = 3500 rpm was reached.Due to rapid rise from n = 2500 rpm the counter torque of the propeller was increasing with the supply current from the batteries.At n = 3500 rpm the measured supply current was IDC = 35 A at the torque level of Tpr = 1.46 Nm.
The measured phase current values in the machine after rectification were considerably smaller than the supply currents.As previously noted, the supply voltage from the batteries was IDC = 35.0 A. In this case the measured phase current in the machine was I13 = 16.4A, also the phase voltage was only U1 = 12.2 VAC.In Fig. 5, the dependency of phase and supply voltages and phase and supply currents at different rotational speed is represented.The ratios between the incoming power to inverter PDC and mechanical power P2 measured by a torque sensor which was connected to the ERPMSM's rotor is characterized in Fig. 6.The difference between PDC and P2 is an expression of the total exponential rise in power loss Ploss of the machine while consisting of magnetic, mechanical and residual losses.Losses in semiconductors of rectifying circuits also have a part in those losses.The most significant portion, though, is formed by ERPMSM electrical, in other words, copper losses Pcu, which constitute 55 % of the total loss in conditions n = 3500 rpm, UDC = 18.5 VDC and IDC = 35.0 A. Almost 30 % of Ploss value resulted from the higher harmonics by the inverter during switching.Other losses compose of iron losses Pfe = 8.0 W, which have been previously determined [5], [6], mechanical losses Pmech, induced by friction in bearings and aerodynamic losses, generated by aerodynamic drag dependant on rotational speed of rotor.
Experimental documentation of ERPMSM mechanical characteristics were conducted with the of a specific test rig with a servo drive for loading the ERPMSM (Fig. 7), enabling torque and rotational speed measurements.During the load tests, servo drive was configured to be equivalent to the counter torque characteristics of the propeller.
In Fig. 8 based on propeller, the values of the dependence of the mechanical load Tm values against rotational speeds are given.At rated voltage UDC = 18.5 V ERPMSM was set to rotate at the corresponding no-load speed n0 = 4260 rpm.Further, the mechanically coupled servo machine was automatically increasing the load by the preliminary specified time intervals.As Tm was gradually growing during the load test, the ERPMSM's rotational speed was decreasing accordingly.This lasted until ERPMSM's torque Tm became equivalent with the counter torque resulted by the load machine.In this method ERPMSM's maximum torque Tmax = 2.2 Nm at UDC = 18.5 V and n = 670 rpm was measured.At rotational speed n < 500 rpm ERPMSM's rotor didn't have enough inertia and power anymore.This caused also the torque reduction in this part of Tm curve.
To identify the load points the counter torque curve of the propeller was included to the ERPMSM's mechanical characteristic.At rotational speed of n = 3000 rpm, the counter torque of the propeller was measured Tpr = 1.13 Nm.On the graph, the crossing point of Tm and Tpr can be noted.Correspondingly, the mechanical load torque Tm is sufficient for maintaining a stable flight regime of the UAV.
The motor operation of ERPMSM also consists of functioning as a starter, during which torque level and rotational speed must be achieved that should be sufficient for starting an ICE connected with the propeller.
According to previous measurements, the starting torque of ICE can be specified as TICE = 0.8 Nm [8], [9].Additional to starting torque, propeller counter torque Tpr, = 1.13 Nm at rotational speed of n = 3000 rpm must be taken in account.Therefore, the optimal starting torque is the sum of Tpr and TICE and gives a result of Tsum = 1.9 Nm.Achieving this starting torque at nominal battery voltage of UDC = 18.5 V, the reduction in rotational speed at the starting moment was expected.
Corresponding to Fig. 8, crossing of Tm and Tsum curves can be observed on rotational speed of n = 2500 rpm.Nevertheless, for the reason that the starting interval does not exceed 3 seconds, loss in flight speed caused by loss of rotational speed can be interpreted as marginal.
In addition, in case of fully charged batteries, it would be possible to exploit higher voltages than nominal UDC for starting.
In Fig. 9, on torque characteristic at UDC = 20.9V, at rotational speed of n = 4000 rpm is shown.
The rise in phase voltages (U1 = 14.6 VAC), phase currents (I13 = 25.0A) and rotational speed (n = 4000 rpm) induces the rise in the ERPMSMs mechanical torque up to Tm = 2.6 Nm.At given conditions, the crossing of Tsum = Tpr + TICE with Tm can be marked at rotational speed of n = 3500 rpm, which exceeds the minimal rotational speed of the propeller, required for a stable flight by 500 rpm.Therefore, the necessary speed is ensured even during the ICE starting period.
At ERPMSM operation as a motor, it is also used for take-offs of the UAV.Ensuring the required rate of increase in flight height, the flight speed of 80 km/h at normal atmospheric pressure must be achieved.Therefore, enough thrust must be generated by the propeller i.e. rotational speed n = 3800 rpm reached, without considering the UAVs aerodynamic drag.
At take-off, ICE and ERPMSM work in parallel, being mechanically coupled.Summation of both of the machines output torques Tm and TICE at various rotational speeds, based on ERPMSM maximal supply voltage UDC = 20.9 and using counter torque of the propeller Tpr as a contrast, total output torque of the HPS Tsum = 2.95 Nm and maximal rotational speed of n = 4860 rpm at given conditions could be observed (Fig 10).

IV. THERMAL CHARACTERISATION
The temperature characteristic of a PMSM expresses the nature of the temperature change during a certain amount of time and at a constant current in the stator windings.The objective of temperature tests is to determine the value of the maximum current where a PMSM reaches the maximum stabilized temperature, which is permissible to the winding insulation and to the permanent magnets [4].
The aim of the current thermal experiments was to determine the maximal temperature of ERPMSMs active parts on steady load.During the experiments, a propeller was used as a load and as a cooling element.Assuming intensive cooling by the propeller, ERPMSM could handle substantially higher supply currents than expected.The first load experiment was made using the propeller at the rotational speed of 4000 rpm and at the supply voltage of UDC = 20.9V, that was near to 5S batteries voltage of UDC = 21 V at their fully charged state.
The measured supply current during the experiment reached IDC = 46.17A, while the phase voltage was measured U1 = 14.6 VAC and the phase current I13 = 20.39 A. Based on that the resulting ERPMSMs apparent power S = 890 VA.
At set parameters, the constant temperature of windings was stabilizing after 8 minutes i.As the maximum temperature of the chosen magnets and winding insulations was substantially higher [10], [11], the additional experiment was made at the rotational speed of n = 5000 rpm, supply voltage of UDC = 27.7 V and current of IDC = 65.6 A. At given parameters the phase voltage U1 = 19.4VAC and phase current I13 = 27.6 A, was measured, resulting the apparent power S = 1600 VA.
Although, at higher power, the constant temperature of ϑp2 = 95 ºC was measured.Similarly to previous experiment the winding temperature rose to constant value during 480 seconds and gave the same time constant  ≈ 120 s.The thermal time constant τ characterizes the approaching velocity rate to the final temperature, as in (1) [12] ( ) 0 0 ( ) 1 , where ϑ0 is the initial temperature (ºC) and t is time (s).The time constant τ depends mainly on the heat capacity factors and weight of ERPMSM's active materials.In practice, the final temperature is reached within 4 times of the time constant values.
The machine can be operated for an indefinite period of time only if the temperatures do not exceed specifications for each part or component of the machine [13].Therefore, a thermal camera was used for capturing temperatures of various parts of ERPMSM, enabling comparison with results collected by use of a temperature probe.The maximum temperature in windings registered on scale was 97.3 ºC, being close to the results achieved with the temperature probe.
Considering the measured temperature in windings as maximum, the temperatures of other components in the machine are lower, which ensures that in previous load conditions the allowed maximum temperatures are not exceeded either in permanent magnets.
In previous phase of the research the performed generator tests with ERPMSM were made with the maximal phase current of If = 23.0A at rotational speed of n = 5500 rpm, therefore the current density in winding conductor reached up to J = 19.3A/mm 2 .The rise in temperature during the experiment stabilized at 106 ºC.From the perspectives of current density and temperature, the given values are notably high.In the current case though, copper wire with the insulation rating of 180 ºC (class H) is used, which has the maximum allowable current density of 23.7 A/mm 2 [14], resulting in even higher allowable rate.According to experiment results, the apparent power of the generator was S = 890 VA [4].During the generator tests, the cooling was only provided by the impeller on the ERPMSMs external rotor.
During the experiments covered with this stage, the forced cooling by using a propeller allowed even higher phase currents I13 as high as 27.6A at rotational speed of 5000 rpm, resulting in the current density of J = 23.2A/mm 2 and reaching to the apparent power of S = 1600 VA, while the winding temperature was stabilizing already at 95 ºC.
The performed thermal experiments hence indicated significantly higher power and current density handling capabilities of ERPMSM with forced cooling provided by the propeller while keeping other machine active parts at notably lower constant temperatures.

V. CONCLUSIONS
Based on this experimental work it can be concluded that ERPMSM fulfils the initial requirements set as a primary propulsion element, either working in parallel with ICE or in starter mode in UAVs stable flying or in take-off regimes at the provided propeller pitch.Experimental work was performed in laboratory conditions at normal atmospheric pressure.Therefore, air mass dynamics in open air, which make an impact on UAV flight speed, were not considered.
It was important to follow UAVs power and mass ratio or the specific power, during the research.As designed HPS exploits the electric machine with its batteries and rectifier circuits in parallel with ICE, it results in additional 1.7 kg i.e. 11 % of weight increase.These parameters were added to UAVs standard weight.Using both ERPMSM and ICE in parallel operation during the take-off, the specific power increased nearly by 71 %.On the other hand in separate performance of both ICE and ERPMSM the specific power decreased by 10 % due to the increased weight.The situation could be improved by increasing the power and torque of both machines, but as for ERPMSM this would result in larger and therefore heavier, accumulator system, which is the main reason of increase in weight of the UAV.Accordingly, the problem must be seen as complex optimization problem with objectives of keeping the mass of every element in the system as low as possible and when possible, research of advanced battery technologies, accumulation devices (super capacitors) or alternative energy resources (solar energy).
The next stage of the research includes also the reliability tests of the HPS in real life situation, mounted on the UAV.Results must be compared to theoretical calculations and with obtained experimental data.
Manuscript received 30 March, 2016; accepted 25 September, 2016.The research was performed in cooperation with Helicam Service Ltd and ELI Airborne Solutions.

Fig. 5 .
Fig. 5. Current and voltage dependencies of the rotational speed with propeller as the load.

Fig. 6 .
Fig.6.The relations between the input PDC and output power P2 in dependence of the rotational speed.

Figure 9
Figure 9 depicts also the rise of counter torque of the propeller (Tpr = 1.97 Nm) at this rotational speed.Due to that the torque required to start ICE increases to Tsum = 2.77 Nm, which exceeds the maximum mechanical torque Tm.

Fig. 10 .
Fig. 10.Torque-speed characteristics during take-off.At such rotational speed, theoretical flight speed in relation to earth of 103 km/h could be reached.Although, taking into account the increasing aerodynamic drag of the UAV due to increase in flight speed, the real flight speed e. 480 seconds to the temperature of ϑp1 = 64 ºC, resulting as the time constant of  ≈ 120 s.The characteristics in Fig. 11 indicate the thermal time constant τ.