PaperlayerView exact revision on arXiv

Source-linked AI summary

A Control Philosophy for Multiplexed Power Converters in Active Distribution Systems

Matthew Deakin, Mahmood Jamali, Nail Tosun, Shafiq Odhano

arXiv:2609.06621v1eess.SYcs.SY

TL;DR

M-PED operation is complex, and time-domain control approaches are lacking despite the topology’s potential flexibility. The paper proposes a state-machine control philosophy to coordinate current-control and selector signals. Simulations illustrate the approaches, with the M-SOP example offering doubled per-unit reactive injection and the two approaches trading simplicity against transient mitigation.

  • Problem

    Time-domain operational studies of changing M-PED topologies are scarce, despite their added operational complexity and potential to address network constraints.

    Source: §I · p. 1, B. Characterising M-PED Control… · p. 3

  • Method

    A state machine converts network-level M-PED setpoints into PWM and multiplexer signals while preventing mutually exclusive reconfiguration and current-control signals from occurring simultaneously.

    Source: §III · p. 4, §V · p. 6

  • Results

    The M-SOP example doubles per-unit reactive injection versus a conventional SOP, and simulations show that both approaches meet switching requirements in a stiff-grid case.

    Source: B. Example of Off-Load and Hot Swap… · p. 5, §IV · p. 6

  • Takeaways & Limitations

    Off-Load control is simpler but risks voltage sag or swell during reconfiguration, whereas Hot Swap continues injecting current during transitions and can mitigate voltage changes at greater complexity.

    Source: §V · p. 6, §IV · p. 6

    The simulation used a stiff grid, making voltage requirements trivial; feeder impedances remain to be tested to demonstrate Hot Swap avoidance of transient sags and swells.

    Source: §IV · p. 6

Abstract

from arXiv · show

Reconfigurable, multiplexed power electronic devices (M-PEDs) can provide more effective, flexible operation for applications in active distribution networks, enabling the mitigation of thermal constraints and voltage violations. To-date, operational approaches of these M-PEDs have not been described, a significant issue as a result of their relative complexity as compared to conventional PEDs. This work proposes a state machine-based control philosophy that can be used to manage the mutually exclusive selector- and current-control signals for M-PEDs. Two control approaches are described using this philosophy: an Off-Load approach, suitable for systems without time-critical network constraints; and a Hot-Swap approach, implemented when the impact of transients must be minimized. Simulations illustrate the evolution of the state machine for a back-to-back voltage source converter topology. It is concluded that the proposed state machine-based control philosophy enables M-PED operation whilst managing the additional complexity of this promising converter topology.

I. INTRODUCTION

M-PEDs add output reconfiguration through power multiplexers, increasing converter flexibility. This paper addresses their operational complexity with a control philosophy linking network-level setpoints to PWM and selector signals.

I. INTRODUCTION

  • M-PEDs add a power multiplexer at each module output and can increase peak power transfer per port by a factor of two or greater, depending on topology and application,,.The paper identifies back-to-back soft open points and four-wire static compensators as PED types that can benefit from this structure.Source: §I · p. 1
  • An M-PED’s operating condition depends on both PWM, which defines each module’s active and reactive power, and selector signals, which determine its output port.The paper’s example is a three-module, two-feeder M-SOP.Source: §I · p. 1, Figure 1 · p. 1
  • Prior M-PED work focuses on quasi-steady-state network-level operation, while time-domain studies of changing converter topology are scarce,,.The passage notes one time-domain study of a PED bypass switch.Source: §I · p. 1
  • The proposed philosophy covers network-level operation through generation of PWM and multiplexer selector signals, addressing the complexity of M-PED operation.The paper presents simulations for an M-PED prototype.Source: §I · p. 2

II. PED AND M-PED CONTROL REQUIREMENTS FOR OPERATION IN ACTIVE DISTRIBUTION NETWORKS

PEDs and M-PEDs are intended to address network constraints without compromising power quality. The paper frames their operation around voltage sag, swell, and rapid voltage change requirements, using Great Britain as a representative setting.

II. PED AND M-PED CONTROL REQUIREMENTS FOR OPERATION IN ACTIVE DISTRIBUTION NETWORKS

  • PEDs and M-PEDs are installed to manage thermal constraints or voltage violations while maintaining the quality of power supplied to end customers.When mitigating voltage violations, a PED can significantly affect voltage at its point of common coupling, making transient and steady-state voltage quality relevant.Source: §II · p. 2
  • The paper uses Great Britain’s EMC requirements as a representative operational setting, while noting that thresholds for voltage sags, swells, and rapid voltage change vary.The typical rapid-voltage-change threshold cited is 3% pu.Source: §II · p. 2
  • Sag and swell are assessed against rms-voltage thresholds, but classification by excursion duration is not considered in this work.The full-cycle rms voltage is updated every half-cycle (10 ms).Source: §II · p. 2
  • When a setpoint change would alter PCC voltage by more than 3%, sufficiently slow PED output can avoid an RVC constraint; the cited steady-state criterion is 1 s with no more than 0.5% voltage change.The paper notes that this steady-state definition can allow the stated RVC constraint to be neglected when output is sufficiently slow.Source: §II · p. 2

A. Conventional PED Control to Meet EMC Requirements

Conventional PED control uses a hierarchy from network-level setpoints to real-time PWM generation. Under the stated assumptions, lower-level PI control can account for voltage sag and swell, while loop timing can address rapid voltage change.

A. Conventional PED Control to Meet EMC Requirements

  • A network-level controller supplies active and reactive power setpoints to a real-time controller that generates PWM signals, typically using PI controllers in cascaded voltage and current loops.The current loops track reference d-q currents.Source: A. Conventional PED Control to Meet… · p. 2
  • For conventional PEDs, lower-level PI controllers can implicitly satisfy sag and swell constraints when the operating region is convex and PCC voltage varies linearly with d-q currents.Under these assumptions, operating points between the old and new setpoints remain feasible.Source: A. Conventional PED Control to Meet… · p. 2
  • When rapid voltage change is relevant, choosing a sufficiently large control-loop time constant can implement slew-rate constraints.The passage says this allows the RVC inequality to be neglected.Source: A. Conventional PED Control to Meet… · p. 2

B. Characterising M-PED Control Philosophy Challenges

M-PED setpoint changes combine current-control changes with discrete multiplexer reconfiguration, and non-convex operating regions can leave no continuous feasible path between setpoints. Fast transitions and switching constraints create practical challenges, motivating an approach that seeks to minimize transient impact while respecting operational requirements.

B. Characterising M-PED Control Philosophy Challenges

  • M-PED operation changes both converter PWM signals and multiplexer selector signals, which determine where power is transferred.The operating configuration therefore changes alongside the converter current control.Source: B. Characterising M-PED Control… · p. 2
  • Non-convex operating regions mean Old and New setpoints generally lack a continuous operating trajectory, so transitions may cause voltage sags or swells.The transition also requires changing discrete selector signals.Source: B. Characterising M-PED Control… · pp. 2–3
  • Completing a transition within 10 ms rules out electromechanical multiplexer switches, whose settling times are tens of milliseconds.The paper gives 25 ms as an example operating time for an 80 A contactor.Source: B. Characterising M-PED Control… · p. 3
  • On-load multiplexer switching requires precise zero-current switching or substantial snubbers to avoid large voltage transients.Source: B. Characterising M-PED Control… · p. 3
  • Feeder angle differences can create shoot-through risk during switching, while PI current control would require very high gains for rapid current changes.A deadband is proposed between opening one feeder switch and closing another.Source: B. Characterising M-PED Control… · p. 3
  • The slower-transition approach assumes off-load switching, at most one closed multiplexer switch, and aims to minimize transient sags and swells while respecting ramp-rate constraints.These constraints can make it impossible to avoid all transient sags and swells when mitigating voltage violations; simple, understandable operation is a heuristic requirement.Source: B. Characterising M-PED Control… · p. 3

III. AN M-PED CONTROL PHILOSOPHY

The proposed control philosophy uses a state machine to manage mutually exclusive PWM and selector controls. Its hierarchy converts network setpoints into transition plans, then uses a follower controller to safely enable the required signals.

III. AN M-PED CONTROL PHILOSOPHY

  • Because multiplexer switches must be switched off-load, PWM current control and selector changes are mutually exclusive and suit state-machine control.Source: §III · pp. 3–4
  • State-machine logic checks ensure that no more than one switch in a multiplexer is closed at a time.Source: §III · p. 4
  • The three-stage architecture has a network-level system set per-feeder powers, a lead controller plan currents and configuration transitions, and a follower controller manage sel and PWM.The follower also implements conventional PI module current and voltage loops.Source: §III · p. 4, Figure 4 · p. 4

A. M-PED Follower Control State Machine

The follower controller uses normal current-control states and additional states that suspend PWM so selector signals can change. Off-Load and Hot-Swap approaches differ in how many modules may remain conducting during reconfiguration.

A. M-PED Follower Control State Machine

1) Implementation of an M-PED lead controller:

The lead controller translates network-level setpoints into current and multiplexer signals through sequenced states. Off-Load simplifies reconfiguration, while Hot Swap updates modules sequentially and requires more controller decisions.

1) Implementation of an M-PED lead controller:

  • Off-Load sets current setpoints to zero, disables PWM, changes multiplexer configurations, then re-enables PWM and sets currents for the new operating point.This simple controller is considered appropriate when voltage sag or swell is not a concern.Source: 1) Implementation of an M-PED lead… · p. 4
  • Hot Swap updates each module’s multiplexer in turn, setting that module’s current to zero while other injections support operational objectives.The chosen operating point and state order affect the capability chart and can influence sag or swell risk.Source: 1) Implementation of an M-PED lead… · pp. 4–5
  • Hot Swap requires a more complex lead controller to choose active and reactive setpoints that meet sag, swell, and state-transition guard conditions.The operating point and order of state transitions must also be selected.Source: 1) Implementation of an M-PED lead… · pp. 4–5

B. Example of Off-Load and Hot Swap Control Approaches

The M-SOP example contrasts Off-Load and Hot Swap reconfiguration with a conventional SOP. M-PED flexibility increases reactive injection capability, but its control transitions are more complex and Off-Load can violate the sag and swell constraint.

B. Example of Off-Load and Hot Swap Control Approaches

  • The example compares an M-SOP setpoint change from 1/3 pu active injection between F1 and F2 to 1 pu reactive injection on F2 with a conventional SOP change to 1/2 pu reactive injection on F2.The M-SOP multiplexer configuration changes to increase the module capacity connected to F2 from 1/3 pu to 1 pu.Source: B. Example of Off-Load and Hot Swap… · p. 5
  • The M-SOP example doubles the per-unit reactive injection possible versus the conventional SOP, while making setpoint changes more complex.The comparison uses an M-SOP change to 1 pu reactive injection on F2 and an SOP change to 1/2 pu on F2.Source: B. Example of Off-Load and Hot Swap… · p. 5
  • Hot Swap changes state four times, compared with twice for Off-Load, while Off-Load’s required zero-load point causes a transient overvoltage that fails the sag and swell constraint.Hot Swap also requires selecting setpoints that satisfy guard conditions and sag and swell constraints, and determining the state-transition order.Source: B. Example of Off-Load and Hot Swap… · p. 5

IV. SIMULATION RESULTS

Time-domain simulations illustrate state-machine operation for a three-module M-SOP. Both approaches meet switching requirements in the modeled stiff-grid case; Hot Swap continues injecting current during reconfiguration, while Off-Load confines current injection to outside its switching interval.

IV. SIMULATION RESULTS

  • The simulation models a three-module M-SOP on two feeders with a three-phase 50 V rms source and 10 Arms switch ratings, using dq-axis PI current control in PLECS.The per-unit current base is 30 Arms.Source: §IV · p. 5
  • Off-Load changes multiplexer configuration only between its trigger signals, when the M-SOP is in S1, and injects current only outside that interval, in S0.The simulation displays multiplexer coil energization and feeder RMS current.Source: §IV · p. 6, Figure 8 · p. 6
  • Hot Swap continues injecting current during the transition to a new setpoint, while each module’s multiplexer changes only when that module is not conducting.The simulation shows reconfiguration between the trigger pair for that module, in its corresponding state.Source: §IV · p. 6, Figure 9 · p. 6
  • The initial feeder RMS-current difference occurs in both control examples because DC-link voltage control supplies active power to balance DC-link energy, not because of reconfiguration.Source: §IV · p. 6
  • Both approaches meet off-load switching and single-selector-output requirements, but the stiff simulated grid makes the voltage constraints trivial to satisfy.Feeder impedances are proposed for future experiments to test whether Hot Swap avoids transient sags and swells.Source: §IV · p. 6

V. CONCLUSIONS

The proposed state machine coordinates M-PED setpoints with PWM and multiplexer signals, preventing simultaneous reconfiguration and current-control signals. The conclusions describe distinct Off-Load and Hot Swap tradeoffs and identify controller optimization as future work.

V. CONCLUSIONS

  • The state machine uses network-level setpoints to determine PWM and multiplexer signals, ensuring reconfiguration and current-control signals cannot occur simultaneously.The architecture’s state count grows only linearly with M-PED modules in the worst case.Source: §V · p. 6
  • Off-Load is simple but risks transient voltage sags or swells during no-load reconfiguration; Hot Swap sequentially reconfigures multiplexers and may reduce voltage-change severity and duration, with added complexity.Hot Swap does not provide the continuously feasible operation of a conventional PED.Source: §V · p. 6
  • Future work could explore lead-controller implementation and systematic optimization to choose transitions that mitigate sag and swell risk.Source: §V · p. 7
  • The authors conclude that M-PEDs have potential as an effective and efficient solution for multiport mission profiles in distribution power-flow control applications.Source: §V · p. 7
Loading 2609.06621v1…