Two-Stage WECC Composite Load Modeling: A Double Deep Q-Learning Networks Approach

Wang, Xinan; Wang, Yishen; Shi, Di; Wang, Jianhui; Wang, Zhiwei

doi:10.1109/tsg.2020.2988171

Cited by 51 publications

(29 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Refer to an EPRI white paper [15, 21], this paper tests a voltage sag benchmarking bus voltage input that is generated by (73). The parameters are set as Table 2 referring to [22].

\begin{matrix} right left right left right left right left right left right left0.278em 2em 0.278em 2em 0.278em 2em 0.278em 2em 0.278em 2em 0.278em3pt \\ V (t) = \{\begin{matrix} left left left1em4pt \\ a & if 1 ⩽ t < (1 + b / 60) \\ \frac{false(1 - d false) false(t - 1 - c false)}{b / 60 - c} + 1 & if (1 + b / 60) ⩽ t < 1 + c \\ 1 & otherwise \end{matrix} \end{matrix}

…”

Section: Model Validation Via Simulationmentioning

confidence: 99%

High‐fidelity large‐signal order reduction approach for composite load model

Ma¹,

Wang²,

Zhao

et al. 2020

IET Generation, Transmission & Distribution

View full text Add to dashboard Cite

With the increasing penetration of electronic loads and distributed energy resources (DERs), conventional load models cannot capture their dynamics. Therefore, a new comprehensive composite load model is developed by Western Electricity Coordinating Council (WECC). However, this model is a complex high-order nonlinear system with multi-time-scale property, which poses challenges on stability analysis and computational burden in large-scale simulations. In order to reduce the computational burden while preserving the accuracy of the original model, this paper proposes a generic high-fidelity order reduction approach and then apply it to WECC composite load model. First, we develop a large-signal order reduction (LSOR) method using singular perturbation theory. In this method, the fast dynamics are integrated into the slow dynamics to preserve the transient characteristics of fast dynamics. Then, we propose the necessary conditions for accurate order reduction and embed them into the LSOR to improve and guarantee the accuracy of reduced-order model. Finally, we develop the reduced-order WECC composite load model using the proposed algorithm. Simulation results show the reduced-order large signal model significantly alleviates the computational burden while maintaining similar dynamic responses as the original composite load model.

show abstract

“…Refer to an EPRI white paper [15, 21], this paper tests a voltage sag benchmarking bus voltage input that is generated by (73). The parameters are set as Table 2 referring to [22].

\begin{matrix} right left right left right left right left right left right left0.278em 2em 0.278em 2em 0.278em 2em 0.278em 2em 0.278em 2em 0.278em3pt \\ V (t) = \{\begin{matrix} left left left1em4pt \\ a & if 1 ⩽ t < (1 + b / 60) \\ \frac{false(1 - d false) false(t - 1 - c false)}{b / 60 - c} + 1 & if (1 + b / 60) ⩽ t < 1 + c \\ 1 & otherwise \end{matrix} \end{matrix}

…”

Section: Model Validation Via Simulationmentioning

confidence: 99%

High‐fidelity large‐signal order reduction approach for composite load model

Ma¹,

Wang²,

Zhao

et al. 2020

IET Generation, Transmission & Distribution

View full text Add to dashboard Cite

show abstract

“…A recent survey conducted by CIGRE [6] shows that a majority of 50 utilities on 5 continents adopt measurement-based approaches. Currently, there are two predominant types of measurementbased methods: static load modeling method [7]- [8], which estimates load parameters based on a ZIP (constant impedance, constant current and constant power) model, and composite load modeling method [9]- [10], which expands beyond the ZIP model by including more dynamic load components such as motor models. All these approaches deal with root mean square (RMS) dynamic load models that uses RMS measurements and This work was supported by the Power System Engineering Research Center.…”

Section: Introductionmentioning

confidence: 99%

Synthesis of Load and Feeder Models Using Point on Wave Measurement Data

Nekkalapu

Vittal

Undrill

et al. 2021

IEEE Open J. Power Energy

View full text Add to dashboard Cite

This paper describes a novel method to synthesize load and feeder models, by estimating the load composition and the load parameters, using instantaneous voltage and current measurements, obtained at the distribution level, from disturbance events. A novel feeder model structure has been proposed in this work comprising of single-phase loads, threephase loads, impedance loads, distribution transformers and distribution line segments. The load composition of the proposed feeder models and the motor load parameters are determined using a non-linear least squares algorithm integrated with an electromagnetic transient (EMT) analysis tool PSCAD. The efficacy of the proposed algorithm is examined for feeders at different physical locations and with distinctly varied load characteristics. Using this approach, parameters of the motor loads have been obtained within a fixed set of bounded values for each corresponding parameter. The ability of the proposed feeder and load models to capture severe fault induced delayed voltage recovery events (FIDVR) has also been discussed here.

show abstract

“…In recent years, with the development of power systems, measurement techniques and load structures have also continued to change. For the composite load model with distributed generation (CMPLDWG) model, scholars have adopted Levenberg-Marquardt algorithm [11], deep learning method [12], double-deep Q-learning method [13], the support vector machine (SVM) approach [14], and so on. Nowadays, many novel swarm intelligence algorithms continue to appear.…”

Section: Introductionmentioning

confidence: 99%

Novel method for power system synthesis load identification

et al. 2020

IET Generation Trans & Dist

View full text Add to dashboard Cite

Power load parameters are one of the most important parameters for power system model. The load has the characteristics of time‐varying, diversity and non‐linearity. Therefore, obtaining the load parameters accurately is crucial to the safe and stable operation of the power system. The research in this paper is based on the synthesis load model which takes into account the impedance of distribution network. The multi‐fault curves identification method and chaos‐whale algorithm are adopted to improve the accuracy and practicability of load parameter identification. First, the algorithm proposed in this paper applies chaos algorithm to whale algorithm, and uses the randomness and ergodicity of chaos algorithm to make up for the shortcomings of the whale algorithm that the solution accuracy is not high and it is easy to fall into local optimum. Secondly, the identification method of multi‐fault curves is to select several different faults for parameter identification, and the average value of each group of results obtained is regarded as the final result. Then compared with single curve identification, the multi‐fault curves method is more superior. Finally, the accuracy of the results is verified by other faults. Simulation results show that the method is reliable to a certain extent.

show abstract

Two-Stage WECC Composite Load Modeling: A Double Deep Q-Learning Networks Approach

Cited by 51 publications

References 30 publications

High‐fidelity large‐signal order reduction approach for composite load model

High‐fidelity large‐signal order reduction approach for composite load model

Synthesis of Load and Feeder Models Using Point on Wave Measurement Data

Novel method for power system synthesis load identification

Contact Info

Product

Resources

About