Improving the Accuracy of DC Optimal Power Flow Formulations via Parameter Optimization

Babak Taheri,Daniel K. Molzahn

2024-10-15

Abstract:DC Optimal Power Flow (DC-OPF) problems optimize the generators' active power setpoints while satisfying constraints based on the DC power flow linearization. The computational tractability advantages of DC-OPF problems come at the expense of inaccuracies relative to AC Optimal Power Flow (AC-OPF) problems which accurately model the nonlinear steady-state behavior of power grids. This paper proposes an algorithm that significantly improves the accuracy of the generators' active power setpoints from DC-OPF problems with respect to the corresponding AC-OPF problems over a specified range of operating conditions. Using sensitivity information in a machine learning-inspired methodology, this algorithm tunes coefficient and bias parameters in the DC power flow approximation to improve the accuracy of the resulting DC-OPF solutions. Employing the Truncated Newton Conjugate-Gradient (TNC) method -- a Quasi-Newton optimization technique -- this parameter tuning occurs during an offline training phase, with the resulting parameters then used in online computations. Numerical results underscore the algorithm's efficacy with accuracy improvements in squared two-norm and $\infty$-norm losses of up to $90\%$ and $79\%$, respectively, relative to traditional DC-OPF formulations.

Optimization and Control,Systems and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to significantly improve the accuracy of the active power set - points of generators in the direct - current optimal power flow (DC - OPF) problem through parameter optimization under specific operating conditions, making it closer to the solution of the alternating - current optimal power flow (AC - OPF) problem. Specifically, the paper proposes an algorithm that utilizes sensitivity information and machine - learning - inspired methods to adjust the coefficients and bias parameters in DC - OPF, thereby improving the accuracy of the DC - OPF solution. Through parameter tuning using the Truncated Newton Conjugate Gradient (TNC) method in the offline training phase and applying these optimized parameters in online calculations, the paper demonstrates the effectiveness of this algorithm, especially in terms of the squared two - norm loss and the infinity - norm loss, with the accuracy improved by up to 90% and 79% respectively compared to the traditional DC - OPF formula.

### Background and Problem Description of the Paper

- **Background**:
- Optimal power flow (OPF) is a fundamental tool in power system design and operation, aiming to optimize system performance (such as minimizing generation cost or reducing voltage deviation) while satisfying equality constraints (based on power flow models) and inequality constraints (based on operating limits).
- Although alternating - current optimal power flow (AC - OPF) can accurately model the nonlinear steady - state behavior of power systems, it is very computationally challenging because these problems are non - convex and NP - hard in the worst case.
- Direct - current optimal power flow (DC - OPF) simplifies the problem by linearizing the nonlinear equations in AC - OPF, improving computational tractability but sacrificing accuracy.

- **Problem**:
- How to improve the accuracy of its solution while maintaining the computational efficiency of DC - OPF, making it closer to the solution of AC - OPF?
- The specific objective is to reduce the difference between the DC - OPF solution and the AC - OPF solution by optimizing the coefficients and bias parameters in DC - OPF.

### Solution

- **Method**:
- An algorithm is proposed, which utilizes machine - learning techniques to optimize the coefficients and bias parameters in DC - OPF through the TNC method in the offline training phase.
- By calculating the gradients of the DC - OPF solution with respect to the parameters and using these gradients for iterative adjustment to minimize the difference between the DC - OPF solution and the AC - OPF solution.
- The optimized parameters are used in online calculations to improve the accuracy of the DC - OPF solution.

- **Results**:
- Numerical results show that, compared to the traditional DC - OPF formula, the accuracy of this algorithm is improved by up to 90% and 79% respectively in terms of the squared two - norm loss and the infinity - norm loss.

### Main Contributions

1. **Introduced an optimization algorithm** for adaptively selecting the coefficients and bias parameters in DC - OPF, which are crucial for modeling line losses and considering factors such as shunts, HVDC feed - in, phase shift, and line losses.
2. **Calculated the gradients of the DC - OPF solution with respect to the coefficients and bias parameters** to obtain the sensitivities of the generator set - points and their defined loss functions.
3. **Utilized the quasi - Newton method (TNC)** to enable the proposed algorithm to be extended to large - scale power systems.
4. **Provided numerical results** to prove the superior accuracy of the proposed algorithm under various operating conditions.

### Conclusion

The paper significantly improves the accuracy of the DC - OPF solution by optimizing the parameters in DC - OPF, making it closer to the AC - OPF solution while maintaining computational efficiency. This provides strong support for the optimal operation and planning of power systems.