Disturbance-Observer-Based Grid-Forming Control for Unbalanced Grids
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.
Rosa: We're looking at the paper: "Disturbance-Observer-Based Grid-Forming Control for Unbalanced Grids".
Dev: Here's the gist of the paper. This article proposes a grid-forming control method for operation under unbalanced grid-voltage conditionsThe method regulates the positive-sequence active power delivered to the grid and actively suppresses the negative-sequence converter voltage, controlling the voltage magnitude to be constant also during unbalanced faults when within the physical limits of the converter.
Rosa: Control Objectives and Methodology The proposed method aims to regulate positive-sequence active power and converter-voltage magnitude in grid-forming modeIt employs a disturbance observer to unify synchronization and estimation of both the positive and negative sequences of the grid voltage, avoiding conventional phase-locked loops and AC-side voltage measurement.
Dev: The control structure is shown in Fig.
Rosa: 1The main contributions are as follows "The proposed grid-forming control method uses, unlike
13: –
22: , a disturbance observer to unify synchronization and estimation of both the positive and negative sequences of the grid voltage." "In contrast to
22: , the disturbance observer is shared between the grid-forming and current-control modes, thus, reducing the order of the control system so that it requires only seven (real-valued) state variables." "The method includes a current-limitation strategy for both symmetrical and asymmetrical faults." System Modeling and Sequence Separation The system model utilizes complex space vectors and per-unit quantities to represent the voltage-source converter connected to an inductive gridThe dynamics of the converter current are governed by Ldic dt = uc − ug − jωcLic (1a)For control synthesis, quasi-static positive-sequence and negative-sequence currents are defined using a complex low-pass filterThe quasi-static positive-sequence and negative-sequence currents are defined as i+c = ic / (1 + j + jif) and i−c = ic / (1 - j - jif).
Paper summary: Rosa: So we're looking at this paper called "Disturbance-Observer-Based Grid-Forming Control for Unbalanced Grids". Essentially, they propose a new way to make the converter operate well even when the grid voltage gets unbalanced. It claims this method can keep the positive power delivered to the grid steady and actively manage those negative sequence voltages, keeping their magnitude constant if everything is within what the hardware can handle during faults.
Dev: That sounds like a significant step because dealing with unbalanced conditions usually requires complex synchronization methods or measuring the AC side voltage directly. The paper suggests they use a disturbance observer to figure out both the positive and negative sequences of the grid voltage, which avoids those conventional phase-locked loops and direct measurements on the AC side, which is pretty elegant for real-time operation.
Taro: And what I find interesting is that this disturbance observer isn't just for synchronization; they use it to estimate both sequence components simultaneously, which they say reduces the complexity of the whole control system down to just seven real variables instead of more. That makes sense if you're trying to keep the loop rate high and minimize latency, which is crucial for fault response.
Rosa: Right, so we're talking about how this unified observer lets them handle both positive and negative sequence estimation without needing separate loops for each sequence component, which simplifies the structure significantly. But they also include a current limitation strategy that works for both symmetrical and asymmetrical faults, so it's built to be robust across different types of problems.
Dev: I see that current limitation is important because it acts as a safety net when things go wrong, providing a concrete boundary for the converter currents during those unbalanced events they are testing. They’re also defining quasi-static positive and negative sequence currents using these complex filter definitions, which then become the controlled outputs of their augmented model.
Taro: Those sequence current definitions you mentioned are what allow them to define things like the quasi-static converter voltage and its components, v plus c and v minus c, which are really important for the grid-forming control perspective. The active power calculation is also derived from these quantities, specifically using the real part of the product of these sequence currents and a specific term.
Rosa: So they're taking those definitions—those sequence currents—and turning them into something physically meaningful, like the positive-sequence active power and converter voltage magnitude, which then feed back into the control law. The control law itself seems to use an error term from the estimated negative sequence voltage to regulate that positive power and magnitude.
Paper summary: Dev: That feedback mechanism is key because it lets them decouple the channels; they can fully separate the positive-sequence active power and voltage control using specific complex gains, which helps maintain stability while controlling those two things independently. It’s a different approach than what some other methods do, like thirteen to twenty-two, which rely on different synchronization techniques <ref:2608.11857#pg2>.
Taro: If you're thinking about the real world, this means the system is designed to maintain stability even when the grid conditions are messy, and it does that by using this shared observer structure to get a better handle on what's actually happening on both sides of that unbalanced connection. How fast they test it is also worth looking at.
Rosa: Yeah, they did experimental verification using a twelve point five-kVA converter, and the results show it can operate during both strong and very weak grids when there are severe balanced or unbalanced faults, which speaks to its broad applicability outside of just clean lab conditions <ref:2608.11857#pg1,using a 12.5-kVA converter>.
Dev: The experimental data also showed that the inductance estimate stays consistent across those different grid conditions, which suggests a good level of robustness against errors in how much inductance the system thinks it has. That’s a solid finding when you’re tuning control parameters for real hardware.
Taro: And the paper itself flags a stability condition where the gain k has an upper limit, specifically that k must be less than four alpha plus alpha v sigma minus alpha plus two alpha v, which means you have to be careful about how high you set that gain if you’re operating in inverter mode with high active power.
Rosa: So while the method is quite powerful for handling unbalanced conditions, the stability analysis shows there are tuning constraints on the gains to keep things from going unstable under certain high-power scenarios. It’s a trade-off between performance and guaranteed stability.
Dev: I agree, it sets clear guidelines for tuning, which is what an engineer needs when deploying this kind of control scheme in a real power system application. The paper provides comprehensive stability analysis and tuning guidelines to help implement this in practice.
Taro: What I’m thinking about moving forward is how well this handles the transient response when the grid suddenly shifts from balanced to unbalanced, because that's where most hardware struggles. The paper addresses that by providing a structure that inherently improves transient stability through its use of realizable references, which they link to work like thirty-four and thirty-five <ref:2608.11857#pg2>.
Rosa: That’s a good point about transients; the method seems designed to provide those references in a way that helps stabilize things quickly when the environment misbehaves. It moves away from just reacting to errors toward proactively shaping the system's behavior based on those sequence estimates.
Paper summary: Dev: So, if you think about this in terms of deployment, this paper suggests a more integrated control system where the disturbance observer handles both estimation and synchronization, which cuts down on system complexity considerably compared to older methods. That reduction in state variables is a tangible benefit for hardware implementation.
Taro: It’s about making sure the control loop can maintain its performance even when the fundamental input—the grid voltage—is changing its characteristics rapidly due to an imbalance or a fault. The way they define those quasi-static quantities lets the system react based on what it *thinks* is happening right now.
Rosa: So, to wrap up this section, the paper lays out a method that uses a disturbance observer to estimate both positive and negative sequence voltages without using AC measurements or conventional PLLs, and it ties that estimation into an active power regulation scheme with current limits. This sets the stage for understanding how grid-forming inverters can handle real-world voltage disturbances.
Dev: And looking at the overall scope of "Disturbance-Observer-Based Grid-Forming Control for Unbalanced Grids", it’s clearly focused on extending existing control techniques to make them work robustly when the grid isn't perfectly balanced. It tackles the synchronization and estimation problem in a way that simplifies the overall system requirements.
Taro: The implication for autonomy research, which is where I see this paper connecting, is that if you can design a controller that handles these specific sequence estimations reliably under fault conditions, it opens up more possibilities for systems operating in fluctuating or imperfect power environments. It’s about graceful recovery when the world misbehaves.
Rosa: That seems like the big picture—moving from just stable operation to actively managing performance during instability. The combination of the disturbance observer and sequence separation is a strong tool for that goal, especially given the experimental success they showed with various grid conditions.
Dev: Exactly, it’s about realizing a control structure that requires fewer state variables while still achieving the necessary control objectives for both active power regulation and voltage magnitude keeping during unbalanced faults. It’s practical engineering focused on reducing system overhead.
Taro: I just hope the real-world deployment proves that those stability limits they derived are achievable with the hardware they tested, because theory is one thing, but surviving a severe fault requires knowing exactly where you stand on that stability margin.
Rosa: Yeah, so the next step for this research is definitely looking at how this performs under more extreme or novel types of faults than what was covered in their twelve point five-kVA converter tests <ref:2608.11857#pg1>. We need to see if that robustness holds up when things get really tough outside the lab setting.
Conclusion: Rosa: So, basically, this paper proposes a new way to control these power converters so they can handle unbalanced grid voltages without needing those complicated phase locked loops.
Dev: Right, they use something called a disturbance observer to figure out both the positive and negative sequence voltages at the same time. It simplifies things because it avoids measuring the AC side voltage directly.
Taro: That means they don't need two separate synchronization systems running in parallel, which cuts down on complexity for real-time operation.
Rosa: And they show that this approach lets them keep the active power delivered to the grid steady while also controlling the converter voltage magnitude, even when there are faults.
Dev: They also have a current limitation strategy built in that works for both symmetrical and asymmetrical faults, which is important for safety because it caps what the hardware can handle.
Taro: I'm interested in how robust this is outside of a perfect lab setup; they tested it on a twelve point five kilo volt converter during severe unbalanced faults.
Rosa: And those results showed good active power tracking even in weak grids, and the inductance estimate stayed consistent, which is a nice sign for real-world deployment.
Dev: But remember, there are stability conditions they found, like this condition where the gain has an upper limit depending on whether you're running in inverter mode with high active power.
Taro: So it’s not just about making it work; it’s about knowing the limits—knowing when the system might become unstable if you push the settings too far.
Rosa: Exactly, this isn't just a theoretical exercise; it sets up a framework for how we can design inverters that react gracefully when the grid itself gets messy.
Dev: It moves us toward control systems that are less reliant on perfect measurements and more focused on estimating what's actually happening in real time.
Taro: If we can get this kind of reliable sequence estimation under fault conditions, it opens up possibilities for more resilient power systems overall.
Rosa: And the next thing we need to look at is how this method handles those really extreme or novel types of faults that might not be covered in their initial twelve point five kilo volt tests.
Aalto University School of Electrical Engineering Department of Electrical Engineering and Automation Aalto University School of Electrical Engineering ABB Oy Drives ABB Corporate Research Aix Marseille University CNRS LIS
eess.SY, cs.SY
Submitted: 2026-08-12
Updated: 2026-11-11
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 83/100
The gist: This article proposes a grid-forming control method for operation under unbalanced grid-voltage conditionsThe method regulates the positive-sequence active power delivered to the grid and actively…
Key concepts
- Disturbance Observer
- This technique is used to unify the synchronization and estimation of both the positive and negative sequences of the grid voltage. It helps in understanding disturbances affecting the system without needing conventional phase-locked loops or direct AC-side measurements, simplifying complex control.
- Sequence Separation
- The system model separates the converter's dynamics into quasi-static positive and negative sequence currents using a complex low-pass filter. These separated quantities are then used as controlled outputs to design the control law, allowing for independent regulation of power and voltage components.
- Grid-Forming Mode Control
- This mode aims to make the converter act like a stable voltage source connected to an inductive grid. The control law regulates positive-sequence active power and ensures the converter's voltage magnitude remains constant, even when dealing with unbalanced faults within physical limits.
Terminology
Summary
This article proposes a grid-forming control method for operation under unbalanced grid-voltage conditionsThe method regulates the positive-sequence active power delivered to the grid and actively suppresses the negative-sequence converter voltage, controlling the voltage magnitude to be constant also during unbalanced faults when within the physical limits of the converter. p. 0
Control Objectives and Methodology
The proposed method aims to regulate positive-sequence active power and converter-voltage magnitude in grid-forming modeIt employs a disturbance observer to unify synchronization and estimation of both the positive and negative sequences of the grid voltage, avoiding conventional phase-locked loops and AC-side voltage measurement. The control structure is shown in Fig. p. 1 1The main contributions are as follows
The proposed grid-forming control method uses, unlike [13]–[22], a disturbance observer to unify synchronization and estimation of both the positive and negative sequences of the grid voltage.
In contrast to [22], the disturbance observer is shared between the grid-forming and current-control modes, thus, reducing the order of the control system so that it requires only seven (real-valued) state variables.
The method includes a current-limitation strategy for both symmetrical and asymmetrical faults.
System Modeling and Sequence Separation
The system model utilizes complex space vectors and per-unit quantities to represent the voltage-source converter connected to an inductive gridThe dynamics of the converter current are governed by Ldic dt = uc − ug − jωcLic (1a)For control synthesis, quasi-static positive-sequence and negative-sequence currents are defined using a complex low-pass filterThe quasi-static positive-sequence and negative-sequence currents are defined as i+c = ic / (1 + j + jif) and i−c = ic / (1 - j - jif). These quantities are used as controlled outputs of the augmented model. p. 2
Control Law Implementation
The grid-forming mode control law is defined as uc,ref = vˆc + e+c + e−cThe term e+c is used to regulate the positive-sequence active power and converter-voltage magnitudeThe term e−c provides feedback from the estimated negative-sequence converter voltage vˆ−c, which is zero when the converter voltage contains no negative-sequence component. The positive-sequence active-power and voltage channels can be fully decoupled with complex gains kp = kp(1 + j)vˆ+cvˆ+ckv = kv(1 − j)uˆ+guˆ+g. p. 2
Stability Analysis and Tuning
The stability analysis involves linearizing the closed-loop system around an operating pointThe stability conditions are obtained by applying the Routh–Hurwitz criterion to the characteristic polynomial (17)The stability condition kn < 4(αp + αv)σ − αp − 2αv means that the gain kn has a strict upper limit only when operating in the inverter mode with a high active power.
Experimental Verification
Experimental results using a 12.5-kVA converter demonstrate that the proposed method can operate during severe balanced and unbalanced faultsThe results verify the performance under balanced and unbalanced faults in both strong and very weak grids. p. 1 The method achieves good active-power reference tracking in both strong and weak grids while the inductance estimate remains the same in both cases, showing excellent robustness against severe inductance estimation errors. p. 6
Implementation Details
Reference limitation is used together with transparent current control to limit the converter current to allowable limitsThe limited active-power reference is given by p¯g,ref = sign(pg,ref) min(pg,ref, pmax). The elliptical trajectory of the current-reference vector is preserved by scaling both sequence components with the same dynamic factor. p. 5 The proposed control method comprises the control system (4)–(9) and the current limitation (20). p. 5 The disturbance observer is implemented according to (23). p. 5
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Unified Synchronization and Estimation via Disturbance Observer: The AI system can synchronize with unbalanced grids and estimate both positive-sequence and negative-sequence grid voltages using a disturbance observer, thereby avoiding the need for AC-side voltage measurement.
-
Grid-Forming Control for Unbalanced Conditions: The control system can regulate the positive sequence active power and converter voltage magnitude while actively suppressing the negative sequence converter voltage to maintain purely positive-sequence voltage source characteristics under unbalanced conditions.
-
Dynamic Current Limitation during Faults: The AI system can implement a current limitation strategy for both symmetrical and asymmetrical faults by employing dynamic sequence separation to directly control the positive- and negative-sequence current components in the current-control mode.
-
Reduced System Order: By sharing the disturbance observer between grid-forming and current-control modes, the resulting control system order is reduced, requiring only seven real valued state variables instead of ten or twenty for alternative methods.
-
Robust Stability Analysis: The AI system can operate within a wide stable operating region by deriving explicit stability conditions and tuning guidelines based on a linearized small-signal model that accounts for estimation errors and the negative sequence feedback gain.
-
Active-Power Reference Tracking: The system can achieve good active-power reference tracking in both strong and weak grids, showing excellent robustness against severe inductance estimation errors when using the proposed method.
Related papers
- Industrial Load Modeling and Optimization for Market-Based Interaction with Power Systems
- Exact and Approximate Convex Reformulation of Linear Stochastic Optimal Control with Chance Constraints
- Certified Set Convergence for Piecewise Affine Systems
- Symbolic Control for Stochastic Systems via Finite Parity Games
- Multiobjective optimization-based design and dispatch of islanded, hybrid microgrids for remote, off-grid communities in sub-Saharan Africa
- Physiology-informed layered sensing for intelligent human-exoskeleton interaction