AC /DC Power Flow Benchmarking
In our previous blog post, we shared how eRoots has been pioneering a generalized power flow methodology tailored for modern HVDC-inclusive grids. Today, we’re excited to showcase how our method stands up to benchmarking against the widely used MatACDC sequential method.
Key Insights from Benchmarking
1. Accuracy of Voltage Solutions: The voltage magnitude and angle differences between the two approaches show minimal mismatches, with deviations clustering near zero. This confirms the high accuracy and consistency of our generalized approach.
For a medium-sized system (39 bus):
For a large system (3120 bus):
2. Computational Efficiency: While the sequential method appears to solve faster within each iteration, our generalized methodology converges in 12 fewer overall iterations, showcasing a marked improvement in computational efficiency.
System | MatACDC Mean Time (ms) | GridCal Mean Time (ms) |
5-bus | 29.22 | 1.90 |
39-bus | 29.48 | 3.12 |
3120-bus | 245.13 | 56.63 |
3. Scalability and Robust Convergence: From small (5-bus) to large systems, our methodology scales effectively. The quadratic convergence, characteristic of the Newton-Raphson method, remains robust across all test cases. In contrast, the sequential method exhibits oscillatory behavior during convergence, often requiring additional re-computations due to fragmented AC/DC interactions.
Our unified approach
Sequential Method
What Sets Our Approach Apart
By unifying the formulation of the AC and DC subsystems, our generalized methodology ensures steady error evaluation and more robust convergence. This contrasts with the sequential method’s reliance on iterative recalculations, which can hinder efficiency and accuracy.
Why This Matters
As power grids evolve to include HVDC systems, accurate, scalable, and efficient methodologies are critical for planning and operations. At eRoots, we’re not just addressing today’s challenges; we’re building the tools to future-proof power grids for tomorrow.