Patent Application
20250150007
This application claims the benefit of European Patent Application No. 23206276.0 filed Oct. 26, 2023, the disclosure of which is incorporated herein by reference in its entirety.
This invention relates generally to the field of multi-channel drives, and the control thereof.
Electronic control architectures are provided with power converters (e.g., motor drives) that form an electrical circuit which accepts an input (e.g., AC or DC) and generates an output (e.g., AC or DC) of a different voltage. The development of such power converters is a significant part of the development of a new control architecture, program or system. It is envisioned that standardised power converters could significantly reduce development costs, by providing the ability for one or more loads to be controlled using multiple power converters.
To facilitate wide power requirement ranges of the one or more loads, it may be necessary to use a plurality of power converters operating in parallel. Mutualised or multiplexed power converter architectures become more relevant as the number of electronic components increases. This is the case, for example, in the aircraft industry as more electronically controlled components are used.
In such systems, a plurality of power converters may be interfaced to one or more electrical loads through a switch or switch matrix that can reconfigure the connection between electrical loads and power converters. In the case of an aircraft, certain loads may not be used simultaneously during a flight or at the same point in time, such as a landing gear actuator and thrust reverser actuator. As such, these components can in theory be powered by the same power converter, with the switch reconfiguring the connection accordingly.
The benefit of a combined or multiplexed approach is the potential reduction of weight and increase of drive availability compared to conventional architectures.
To take the maximum advantage of the combined or multiplexing concept (in particular the latter), paralleling control architectures for motor drives is a critical technology, which enables large loads to be powered by the parallel connection of multiple drives. This avoids having to dimension one motor drive for maximum load power, for example.
Solutions for providing parallel motor drives should be aimed at reducing weight, for example of the various components used within the architecture. It has also been found that the use of scalable and modular solutions may be important in enabling motor drive inverters to operate in standalone or a parallel configuration to a variable number of motor drives (e.g., two or more drives in parallel).
One example architecture that is utilised so as to allow for multiplexing of various power converters is a multi-channel drive (MCD) system, such as the one shown in
Such architectures may use synchronous reference frame control, which is known in the art (also called “D-Q control” or Field Oriented Control). Using such a control mechanism, the voltage and current can be converted into a reference frame that rotates synchronously with a voltage vector (e.g., by Park Transformation) so that three-phase time-varying signals are transformed into DC signals. Using “D-Q currents” (direct and quadrature currents) provides a convenient way of representing the output current of the motor drive, and analysing the system. By utilising such control, it is possible to align the current vector with the Q-axis of a rotating DQ reference frame, in order to ensure maximum motor torque production for a given current.
In such MCD systems, it is desirable to improve the dynamic speed response in MCD configuration (e.g. the ability for the system to react to a change in demand from the load(s)), minimise steady state error of the MCD in all load conditions, and to ensure that the provided current is equally shared between the paralleled motor drives.
According to a first aspect, there is provided a system for providing power to one or more loads. The system comprises a central controller configured to output a speed demand for one or more loads; and a plurality of power converters arranged in a parallel configuration with each other and configured to provide power to the one or more loads. Each power converter comprises a speed control loop configured to calculate a local current demand based at least in part on an output speed error between the speed demand of the one or more loads and a measured speed of the one or more loads; a torque/current control loop configured to calculate an output voltage demand based at least in part on the local current demand of its respective speed control loop; and a module configured to modulate the output voltage of an inverter based on the output voltage demand from the torque/current control loop.
The central controller may be configured to calculate a global current demand based upon the local current demands calculated by each of the power converters, and transmit the global current demand to each of the power, and each speed control loop may be further configured to calculate its local current demand based at least in part on a proportional feedback signal of the global current demand.
In this way, there is provided an architecture providing “nudge control” in the speed control loop, thereby enabling improved dynamic speed response in the multi-channel drive configurations, minimized steady-state speed error of the paralleled multi-channel drives in all load conditions, and is very effective against the low-speed high-torque load condition. Furthermore, this is achieved whilst ensuring balanced current sharing between drives, and increasing robustness to communication failures. If the communication lines are down and the average Q-axis current demand is unavailable, the controller can revert back to be like a discharge-term architecture.
Each torque/current control loop may be configured to calculate its output voltage demand based on the global current demand from the central controller. In doing so, the architecture may make use of a similar “nudge” control in the torque/current control loop, improving performance in multi-channel drive configurations.
In this way, according to a first example, there may be provided a system for providing power to one or more loads. The system comprises a central controller configured to output a speed demand for one or more loads; and a plurality of power converters arranged in a parallel configuration with each other and configured to provide power to the one or more loads. Each power converter comprises a speed control loop configured to calculate a local current demand based at least in part on an output speed error between the speed demand of the one or more loads and a measured speed of the one or more loads; a torque/current control loop configured to calculate an output voltage demand based at least in part on the local current demand of its respective speed control loop; and a module configured to modulate the output voltage of an inverter based on the output voltage demand from the torque/current control loop; wherein the central controller is configured to calculate a global current demand based upon the local current demands calculated by each of the power converters, and transmit the global current demand to each of the power, and each speed control loop is further configured to calculate its local current demand based at least in part on a proportional feedback signal of the global current demand.
Each local current demand may comprise a local Q-axis current demand and a local D-axis current demand, and the global current demand may be a global Q-axis current demand, and wherein the local Q-axis current demand may be calculated based at least in part on the output speed error between the speed demand of the one or more loads and the measured speed of the one or more loads, and the proportional feedback signal of the global Q-axis current demand. In this way, the architecture may advantageously make use of DQ control.
Each local D-axis current demand may be calculated at least in part from the measured motor speed, and may be calculated in order to take into account field weakening.
Each output voltage demand may comprise an output Q-axis voltage demand and an output D-axis voltage demand, the output Q-axis voltage demand may be calculated at least in part based on the respective local Q-axis current demand of its respective speed control loop, and the output D-axis voltage demand may be calculated at least in part based on the respective local D-axis current demand of its respective speed control loop.
The system may further comprise a module configured to convert the voltage demand from the DQ domain to a three phase voltage domain for the module configured to modulate the output voltage of the inverter.
Each output voltage demand may be further calculated based on a feedforward term that is proportional to the measured speed of the one or more loads. The measured motor speed of the one or more loads may be scaled by an associated gain that is set equal to motor back EMF gain in order to provide the feedforward term. Such a system may further improve the MCD control performance during configuration changes by implementing the proportional feedforward voltage demand term, which is independent of the global D-axis and Q-axis voltage demands.
The central controller may configured to instruct one or more of the plurality of power converters to come online in order to meet the speed demand.
In this way, according to a second example, there may be provided a system for providing power to one or more loads. The system comprises a central controller configured to output a speed demand for one or more loads; and a plurality of power converters arranged in a parallel configuration with each other and configured to provide power to the one or more loads. Each power converter comprises a speed control loop configured to calculate a local current demand based at least in part on an output speed error between the speed demand of the one or more loads and a measured speed of the one or more loads; a torque/current control loop configured to calculate an output voltage demand based at least in part on the local current demand of its respective speed control loop and a feedforward signal that is proportional to the measured speed of the one or more loads; and a module configured to modulate the output voltage of an inverter (350) based on the output voltage demand from the torque/current control loop. The central controller may be configured to instruct one or more of the plurality of power converters to come online in order to meet the speed demand.
Such a system may further improve the MCD control performance during configuration changes by implementing the proportional feedforward voltage demand term, which is independent of the global D-axis and Q-axis voltage demands and also independent of the “nudge-term”.
As would be appreciated, the system of the first aspect may be combined with the system of the second aspect, so as to provide the architecture of the first aspect, with the feedforward term of the second aspect. Such a system may therefore not only provide improved dynamic speed response in the multi-channel drive configurations, minimized steady-state speed error of the paralleled multi-channel drives in all load conditions, and is very effective against the low-speed high-torque load condition, but also be robust to configuration changes when one or more of the power converters come online.
The measured speed of the one or more loads may be scaled by an associated gain that is set equal to motor back EMF gain in order to provide the feedforward term.
The central controller may be configured to calculate a global current demand based upon the local current demands calculated by each of the power converters, and transmit the global current demand to each of the power converters, and each speed control loop may be further configured to calculate its local current demand based on a proportional feedback signal of the global current demand.
Each torque/current control loop may be configured to calculate its output voltage demand based on the global current demand from the central controller.
The central controller may be configured to calculate a global voltage demand based upon the local voltage demands calculated by each of the power converters (, and transmit the global voltage demand to each of the power converters, and each torque/current control loop may be further configured to calculate its local voltage demand based at least in part on a proportional feedback signal of the global current demand.
According to a second aspect, there is provided a method of providing power to one or more loads. The method comprises receiving a speed demand for the one or more loads, calculating, by each of at least one power converters, a respective local current demand based at least in part on an output speed error between the speed demand of the one or more loads and a measured speed of the one or more loads; calculating, by each of the at least one power converters, a respective output voltage demand based at least in part on the local current demand of the speed control loop; and modulating, by each of the at least one power converters, the output voltage of a respective inverter, based on the respective output voltage demand.
Each respective local current demand may be calculated based at least in part on a proportional feedback signal of a global current demand, the global current being calculated based upon the local current demands.
The respective output voltage demand may be calculated based on a feedforward term that is proportional to the measured speed of the one or more loads.
One or more further power converters may be instructed to come online in order to meet the speed demand of the one or more loads. Each of the further power converters may be configured to perform the same steps as the power converter(s) that are already online.
Certain examples of the disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which:
Whilst the description herein may refer to only a specific number of paralleled power converters (e.g. motor drives) configured to be able to drive one or more common load(s), it would be appreciated that the teaching herein could be applied to any number of loads, and their motor drives. For example, there could be three or more paralleled powered converters (e.g. motor drives) configured to drive (e.g. control or actuate) a common load.
It will be appreciated that references to motor drives in accordance with various embodiments herein could refer to any type of power converter, and the broadest aspects of the present disclosure are not intended to be limited to using motor drives, or providing a load in the form of a motor.
To enhance the paralleled drive operation, a distributed control architecture has been proposed which utilise the global voltage demand as described in EP3916990A1 and EP4024649A1, both of which are incorporated herein by reference and/or utilise a discharge term within the speed control loop as described in EP4243274A1, which is incorporated herein by reference. Such a control architecture utilising these methods has been implemented, for example as shown in
As can be seen in
Each power converter may include an inverter 250. The inverter 250 may be considered to convert DC power received by each power converter into an output voltage (e.g., a three-phase AC voltage) for powering/actuating one or more load(s). Whilst two motor drives 201, 202 are shown in
As discussed above, the plurality of motor drives 201, 202 typically need some sort of common or centralised operation (e.g., control system) for example to synchronise their operation. Accordingly, the motor drives 201, 202 may be controlled centrally using the DSC 260 (e.g., processor or circuitry). The central controller 260 may be configured to control, for example, which motor drive(s) 201, 202 are to be used (e.g., selected as discussed above) at a particular time. The DSC 260 may achieve this by switching between the motor drives (e.g., the power inverters thereof) using, e.g., a modulation scheme or technique.
Each motor drive 201, 202 may include components or circuitry that form a speed loop 210 having as an input a reference speed (illustrated as a speed demand and speed ramp in the Figures), for example an angular speed demand ωref, and also a motor speed ωm (which may be measured or calculated).
The reference speed ωref would typically be given to all power converters operating in parallel, and could be set by the central controller (e.g. DSC) 260 or elsewhere (e.g., a flight control computer, or other component or module, e.g., having another control loop). For example, a flight controller may be commanding that a component moves by actuating to a certain position at a given speed. The speed loop 210 is configured to generate suitable current reference(s) (e.g. a local Q-axis current demand, and/or local D-axis current demand), which is then input into a torque/current loop 220 (described below) to regulate the output current of the motor drive, in effect regulating the speed of rotation of the motor. Although speed loops such as this are known in the art, the technology disclosed herein means we can retain this control while integrating a parallel system.
Each motor drive 201, 202 comprises a (local) torque/current control loop 220 that is configured to calculate a local voltage reference (e.g. local Q-axis voltage demand and/or local D-axis voltage demand) for the motor drive 201, 202 as discussed above. The torque/current loop 220 is configured to regulate the output current of the respective motor drive 201, 202, so as to ensure the motor drives 201, 202 operating in parallel work together adequately and effectively. That is, to generate the correct currents, the output voltage of the motor drive 201, 202 needs to be set correctly. The current loop 220 (and speed loop 210) are used to calculate the local voltage reference(s) (such as local Q-axis voltage demand, and/or local D-axis voltage demand). These voltage reference(s) are then converted from the DQ domain to the ABC (i.e. three phase voltage) domain by a conversion module 230, which may then be used as part of the modulation (e.g., to generate suitable modulation duty cycles). This is then used to generate the correct switching sequence of the motor drives 201, 202 (e.g., via the modulation technique or scheme) to facilitate this.
Any suitable method or type of modulation may be employed in the present disclosure. In particular, the modulation may comprise pulse width modulation (“PWM”) such as space vector modulation (“SVM”). Each motor drive 201, 202 may include a modulation module 240 configured to control the synchronisation of the motor drive 201, 202, for example the synchronisation of a switching frequency of the modulation, e.g., using PWM or SVM.
The modulation (e.g., PWM or SVM) module 240 is configured to generate the appropriate signals (e.g., logic and/or switching signals) for the inverter 250 of the motor drive 201, 202 (e.g., gate drivers thereof) to provide a controlled commutation of the output voltages thereof, whilst regulating the output voltages/currents of each inverter 250 to maintain consistency between the motor drives 201, 202. Thus, the modulation module 240 may be configured to send suitable such signals to the inverter 250 that control the timings associated with the respective motor drive 201, 202 (e.g., the output voltages thereof) in use. Generation of the switching signals for the inverters 250 by each respective modulation (e.g., SVM) module 240 is facilitated by the use of the global voltage references.
Each motor drive 201, 202 may be a single power converter and the circuitry/components thereof may be located on a single circuit board (e.g., a single printed circuit board or “PCB”). The circuit boards (or “blades”) holding the motor drives 201, 202 could be similar (or substantially identical), such that they comprise the same interface for communicating with the DSC 260. In this manner, the motor drives 201, 202 could be easily scaled (e.g., from using two motor drives 201, 202 to using three or more) without any modification of the motor drives 201, 202 or central controller 260.
A rack or similar structure could be used to connect a variable number of (e.g., between 2 and 20) motor drives 201, 202 to the central controller 160, in a similar manner to server racks, into which server blades can be inserted and removed easily depending on system requirements.
The current values of the three phases (i.e., the instantaneous inverter 250 output currents) may be measured and/or determined for use in the local current control loop 220 (see “Q-axis current”/“D-axis current” in the Figures). Suitable components or circuitry may be configured to determine the output currents. The output currents (e.g., phase currents) could be measured and/or determined using a current sensor, although any suitable technique could be used, with optionally a current sensor provided for each of the three phases.
The control architecture 200 supplements the PI control loops (such as those discussed above) with two additional control methods.
Firstly, the DSC 260 is configured to determine or calculate a global voltage reference for each of the q-axis and d-axis based on the local output voltage references from each of the paralleled motor drives 201, 202. The global voltage reference may be an average or sum of the local voltage references. The global voltage reference is then transmitted back to each local motor drive 201, 202, to be converted from the DQ domain to the ABC (i.e. three phase voltage) domain by conversion module 230, which may then be used as part of the modulation (e.g., to generate suitable modulation duty cycles). For example, the modulation elements may generate a suitable switching pattern for the inverters 250 based on the global voltage reference.
The control architecture 200 of
The calculated local voltage demands may be transmitted to an averaging unit 231 of DSC 231, which calculates respective global average demands (i.e. the respective average of each of the local voltage demands from each separate power converter). The global voltage demands are then fed back to their respective inputs of the torque/current control loop 220, and subtracted from their respective current error.
In this way, the “nudge term” is fed by the difference between the global (average) from the averaging unit 261 of the DSC 260, and the local voltage demand values, as can be seen in
The architecture 200 also utilizes a “discharge” algorithm 211 which is a proportional feedback control inserted in the speed control loop 210. In this way, the Q-axis current demand of the motor is calculated based on a speed error between the desired speed/speed demand indicated by the DSC and the measured speed of the motor, with a further feedback signal of the local output current demand subtracted from the speed error. The discharge term is fed by the local demand value, and is aimed at improving current sharing between paralleled motor drives.
The proportional nudge gain is normally set as a relatively smaller value compared to the discharge gain. Even so, the discharge-term control architecture produces a steady-state speed error particularly in low-speed high-torque load conditions. This mainly error occurs due to the discharge term affecting the speed error calculation at the input of the speed PI controller. The discharge term is used to balance the Q-axis current demands of the paralleled drives.
This can comprise producing a local Q-axis current demand as shown in
In this way, the DSC 360 may be configured to determine or calculate a global current demand for the Q-axis current, based on the local output current references from each of the paralleled motor drives 301, 302. The global current reference may be an average or sum of the local voltage references.
In order to implement a “nudge” algorithm 311 in the speed control loop 310, an additional proportional feedback control inserted in each current control loop for their respective Q axes current demand controllers.
The speed control loop 310 receives, at its input, the reference speed (e.g. speed demand) and the measured/calculated motor speed from the DSC 360, and calculates a respective local Q-axis axis current demand based on their respective speed error. The local D-axis current demand may be calculated based on the measured/calculated motor speed, optionally taking into account a certain degree of field weakening.
The calculated local Q-axis current demands may be transmitted to an averaging unit 362 of DSC 360, which calculates a global average current demand (i.e. the respective average of each of the local current demands from each separate power converter). The local current demand is then subtracted from the average global current demand, and the resulting difference is then scaled by a gain and feedback to the input of the speed control loop, and subtracted from the speed error from which the Q-axis local current demand is calculated.
In this way, each drive sends its locally calculated Q-axis current demand value to the DSC 360. Then the DSC 360, in averaging unit 362, calculates an averaged value of current demands, and sends it back to each drive as a global Q-axis current demand. The difference between the global and the local Q-axis current demands is fed to the nudge proportional controller 311, while the local Q-axis current demand is fed to the current control loop.
A similar control architecture 400 is provided as seen in
As such, if the communication with DSC 430 fails and the global Q-axis demand is unavailable, the controller simply reverts to the discharge-term architecture, similar to that seen in
While the description herein is described with reference to two-level drives, it would be appreciated that the teaching herein is applicable for other multi-level power converter topologies because in case of multi-level drives, the multi-level capacitor voltages are controlled locally by each drive controller.
Control architectures with the speed control discharge term (as seen in
It can be seen that, under the no-load condition, the discharge-term controller produces about 5% speed error during the speed ramp-up period, and <1% during the steady-state period as shown in plots 501 and 502.
The results of a similar test using a “nudge term” architecture, such as the architecture 300 seen in
Whilst the nudge-term controller produces momentary 5% speed error during the speed ramp-up period and the speed error reduces toward 0% during the steady-state period as shown in plots 602 and 603 of
The discharge-term controller can act only as a balancer, which helps minimize instability between the two drives by damping the oscillation of the Q-axis current demand as shown in plots 503 of
The results of a similar test using a “nudge term” architecture, such as the architecture 300 seen in
Under the load condition, the discharge-term controller produces about 10% speed error during the speed ramp-up period and the error keeps increasing to about 25% during the steady-state period as shown in plots 701 and 702 in
As such, it has been show that speed error problems become more pronounced with a discharge-term controller under high load conditions. Whilst, in a discharge-term control strategy, it is possible to implement a look-up table to boost the speed demand and compensate for the speed drops due to the increasing load torque, such a method can be is inaccurate and time consuming to design as it will requires the users to pre-map speed-torque characteristic of the load. By implementing “nudge control” in the speed control loop, it is possible to achieve the following results:
As such, an architecture providing “nudge control” in the speed control loop enables improved dynamic speed response in the multi-channel drive configurations, minimized steady-state speed error of the paralleled multi-channel drives in all load conditions, and is very effective against the low-speed high-torque load condition. Furthermore, this is achieved whilst ensuring balanced current sharing between drives, and increasing robustness to communication failures. If the communication lines are down and the average Q-axis current demand is unavailable, the controller reverts back to be like a discharge-term architecture.
With such “nudge control” in the speed loop and/or the torque/current control loop (for example, in the configurations of any of
When more drives come online during the MCD operation (e.g. as instructed by the DSC, or other central controller), the DSC is configured to calculate a new average value for both the DQ voltage demands (referred to as Global D-axis and Q-axis voltage demands in
If the reduced Q-axis voltage demand is lower than the motor's back EMF (BEMF), unintentional transition from motoring to regenerative mode can occur within the MCD. This problem becomes more pronounced at higher mechanical speeds, particularly in the field-weakening region, where the motor's BEMF voltage surpasses the DC-link voltage. The immediate regenerative effect may cause the DC-link voltage to overshoot, increasing system instability and posing a risk to the drive's power devices, which have limited tolerance to overvoltage conditions. The steady-state regenerative power could be controlled if a braking resistor unit is present on the DC bus. However, a basic MCD, such as the one shown in
The performance of an architecture, such as architecture 300, has been tested experimentally using the two-paralleled drives connected to a motor load similar to the MCD in
As can be seen in plots 902 and 1002, each drive provides the motor Q-axis with about 6A of current in the DD mode. A single drive provides about 12A in the SD mode.
As illustrated in plots 903 and 1003, when observing the DQ-axis voltage demands, the Q-axis voltage demand increases from 43V to 50V (an increase of around 16%) when changing from the DD to SD mode and reduces back when the mode changes back. The D-axis voltage demand remains at the same level of around-5V.
From plots 904 and 1004, there is no DC-link voltage overshoot during the DD-to-SD mode changes, because the voltage demands from the two drives are equal to each other and equal to the average voltage demand from DSC. However, as shown in plots 1002 and 1003 of
As the load torque is fixed at 5 Nm, as seen in plots 1102, 1202, 1302, and 1402 the Q-axis current demand remains the same as the value recorded during corresponding tests performed at 1 krpm/5 Nm. As seen in plots 1103, 123, 1303 and 1403, the Q-axis voltage demand increases slightly from 100 to 107V (7%) at 3 krpm and from 180V to 203V (12%) at 6 krpm when changing from the DD to SD mode and reduces back when the mode changes back. The D-axis voltage demand oscillates during the mode transition but remains at the same level after (around-18.6V for 3 krpm and −50V for 6 krpm).
However, as can be seen in plots 1204 and 1404, there are the DC-link voltage overshoots of about 30V for 3 kprm (i.e. from 540V to 570V) and about 81V for 6 kprm (i.e. from 540V to 621V) during the single-mode to dual-mode transition. The DC-link overshoots are in-line with the occurrences of the negative Q-axis current responses of −5 A, for 3 kprm (plot 1203) and −10 A, for 6 krpm (plot 1403).
In conclusion, the MCD suffers from the DC-link voltage overshoot problems when the number of electrically-paralleled drive channels is increased. Without the speed feedforward term, the voltage rating of the power devices must be increased which may lead to physically larger power converters.
An architecture 1500 that can manage overvoltage levels may be seen in
Such an architecture may further improve the MCD control performance during configuration changes by implementing the proportional feedforward voltage demand term, which is independent of the global D-axis and Q-axis voltage demands, and also independent of the “nudge-term” as seen in
As can be seen, by utilising the feedforward term, the architecture 1500 calculates the total Q-axis voltage demand as the sum between the locally calculated feed-forward term and the globally calculated (i.e. average of the local demands) Q voltage demand received from DSC 360. This solution stabilises the operation of the drives, when the number of active drives is increased on-the-fly during the MCD operation.
The feedforward term may be proportional to the mechanical motor speed and the associated gain may be set to be equal to motor back EMF gain, Ke. As a result, the total Q-axis voltage demand may then be the average or sum of the global demands from the DSC 360, and the local feedforward term. This voltage demand is then converted from the DQ domain to the ABC (i.e. three phase voltage) domain by conversion module 330, which may then be used as part of the modulation (e.g., to generate suitable modulation duty cycles using modulator 340). This is then used to generate the correct switching sequence of the motor drives 301, 302 (e.g., via the modulation technique or scheme) to facilitate this.
Again, such a feedforward term is applicable to two-level drives and multi-level drives because the feedforward term operates independently of the capacitor voltage balancing algorithm of a multi-level drive.
It will be appreciated by those skilled in the art that the disclosure has been illustrated by describing one or more specific examples, but is not limited to these examples; many variations and modifications are possible within the scope of the accompanying claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 23206276.0 | Oct 2023 | EP | regional |
