Composite Load Model Research Articles

Parameter Reduction of Composite Load Model Using Active Subspace Method

Over the past decades, the increasing penetration of distributed energy resources (DERs) has dramatically changed the power load composition in the distribution networks. The traditional static and dynamic load models can hardly capture the dynamic behavior of modern loads especially for fault-induced delayed voltage recovery (FIDVR) events. Thus, a more comprehensive composite load model with combination of static load, different types of induction motors, single-phase A/C motor, electronic load and DERs has been proposed by Western Electricity Coordinating Council (WECC). However, due to the large number of parameters and model complexity, the WECC composite load model (WECC CMLD) raises new challenges to power system studies. To overcome these challenges, in this paper, a cutting-edge parameter reduction (PR) approach for WECC CMLD based on active subspace method (ASM) is proposed. Firstly, the WECC CMLD is parameterized in a discrete-time manner for the application of the proposed method. Then, parameter sensitivities are calculated by discovering the active subspace, which is a lower-dimensional linear subspace of the parameter space of WECC CMLD in which the dynamic response is most sensitive. The interdependency among parameters can be taken into consideration by our approach. Finally, the numerical experiments validate the effectiveness and advantages of the proposed approach for WECC CMLD model.

Examination of Composite Load and Variable Frequency Drive Air Conditioning Modeling on FIDVR

Fault induced delayed voltage recovery (FIDVR) is a credible concern in power systems with high penetration of single-phase air conditioner loads. Power system planning studies use dynamic load models, such as the Western Electricity Coordinating Council (WECC) composite load model CMPLDW to examine FIDVR in power systems. The behavior of modern variable frequency drive (VFD) air conditioners to voltage drops is not the same as traditional one- or two-speed air conditioning systems. This paper investigates the impact of increasing VFD air conditioner penetration within a power system on FIDVR events, by first modeling VFD air conditioners as a separate load and then as a power electronic component within the CMPLDW model. Simulation results show that increasing the penetration of VFD driven air conditioners decreases the post-fault voltage sag, recovery time, and number of customers interrupted by undervoltage load shedding.

An Investigative Study on Impact of Frequency Dynamics in Load Modeling

Load modeling plays a significant impact in assessing power system stability margin, control, and protection. Frequency in the power system is desired to be kept constant, but in a real sense, it is not constant as loads continually change with time. In much literature, frequency dynamics are ignored in the formulation of load models for the basic assumption that it does not affect the models. In this paper, the composite load model was formulated with Voltage-Frequency Dependency (V-FD) on real and reactive powers and applied to estimate the load model. 2- Area network 4- machines Kundur test network was used for testing the developed model. The model was trained with measurements from a low voltage distribution network supplying the Electrical Engineering department at Ahmadu Bello University, Zaria. Both training and testing data were captured under normal system operation (dynamics). To evaluate the V-FD model performance, Voltage-Dependent (VD) model was examined on the same measured data. The work makes use of the Feed Forward Neural Network (FFNN) as a nonlinear estimator. Results obtained indicate that including frequency dynamics in modeling active power reduces the accuracy of the model. While in modeling reactive power the model performance improves. Hence, it can be said that including frequency dynamics in load modeling depends on the intended application of the model.

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

With the increasing penetration of electronic loads and distributed energy resources, 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 non-linear system with multi-time-scale property, which poses challenges on power system studies with heavy computational burden. In order to reduce the model complexity, the authors firstly develop a large-signal order reduction (LSOR) method using singular perturbation theory. In this method, the fast dynamics are integrated into the slow ones to preserve transient characteristics of the former. Then, accuracy assessment conditions are proposed and embedded into the LSOR to improve and guarantee the accuracy of reduced-order model. Finally, the reduced-order WECC composite load model is derived by using the proposed algorithm. Simulation results show that the reduced-order large-signal model significantly alleviates the computational burden while maintaining similar dynamic responses as the original composite load model.

WECC Composite Load Model Parameter Identification Using Evolutionary Deep Reinforcement Learning

Due to the increasing penetration of distributed energy resources (DERs), the load composition in distribution grids has significantly changed. This inverter-based device has notably different behavior from traditional static and induction motor loads. To accurately represent the combination of static load, induction motor and the emerging inverter-based devices, the composite load model with distributed generation (CMPLDWG) has been developed by Western Electricity Coordinating Council (WECC). Due to the large number of parameters and model complexity, the CMPLDWG model brings new challenges to parameter identification, which is critical to power system studies. To address these challenges, in this paper, a cutting-edge approach inspired by the evolutionary deep reinforcement learning (EDRL) with an intelligent exploration mechanism is innovatively proposed to perform parameter identification for CMPLDWG. First, to extract parameters' contributions to dynamic power, parameter sensitivity analysis is conducted using a data-driven feature-wise kernelized Lasso (FWKL). Then, the EDRL with intelligent exploration, which can handle the natural high nonlinearity and nonconvexity of CMPLDWG, is employed to perform parameter identification. In the parameter identification process, the extracted parameter sensitivity weights are innovatively integrated into the EDRL with intelligent exploration to improve discovery efficiency. Finally, the proposed approach is validated using numerical experiments.

Modeling the Aggregated Response of Variable Frequency Drives (VFDs) for Power System Dynamic Studies

Accurate representation of system loads is critical for performing power system dynamic studies. As of date, in positive sequence bulk power system dynamic studies, the WECC composite load model is the state-of-the-art model for representing the aggregated behavior of loads at the substation bus. Although used widely, this model is continuously undergoing improvements to ensure that a reasonable representation of different load components can be obtained. One such refinement is the development of a model for representing the aggregated response of VFDs. This paper proposed a generic positive sequence model used to capture the aggregated response of a large number of VFDs connected at a load bus for bulk system dynamic studies. The model was developed and validated using laboratory test measurements of response different VFDs to voltage sags.

A two-stage framework for ambient signal based load model parameter identification

Load modeling is a crucial part of modeling a power system. Traditional load model identification methods are mainly based on post large disturbance response, which cannot be conducted without large disturbance events. In this paper, a new approach of ambient signals based load model parameter identification is proposed to solve this problem, through which the time-varying load model parameters can be identified more frequently. First, the composite load model structure and the measurement based identification approach are introduced. Then, a two-stage identification framework to identify the load model parameters from ambient signals is proposed. In the first stage, the electromagnetic parameters are identified using an optimization method, with the predicted error as the objective function and the differential evolution as the optimization algorithm. In the second stage, the mechanical parameters are identified using linear regression. Finally, the effectiveness of the proposed identification method is validated through the simulation results in Guangdong Power Grid with practically measured ambient signals data. The proposed two-stage identification framework has shown its ability to accurately identify the load model parameters from ambient signals and the advantage over the other identification approaches.

Reactive Power Cost Optimization Acquiring the Combined Properties of Static and Aggregate Dynamic Load as Composite Load Model

In electrical systems especially isolated wind and diesel based hybrid systems, voltage control has been achieved in several works from adequate supply of reactive power using STATCOM/SVC as dynamic compensators. However, in most of the existing works, studies are performed only with static loads which do not have any influence of previous system conditions. In such available studies, the two major concerns were (i) cost based studies, and (ii) control techniques of dynamic compensation. The electrical systems are always influenced by the unrealistic load modelling in such systems which is almost untouched in available works and therefore motivate the authors to work on it. Realistic loads are of composite nature that composes the blend of characteristics that acquire static as well as dynamic nature of the load. In present work, the authors are interested to investigate the lower cost of reactive power with composite load for best proportion of dynamic and static compensations. The study is done in MATLAB software with 10% step increment in reactive power demand and input wind power to analyse compensation effect for extreme voltage deviation of pu at load end as recommended by several energy policies and standard IEC 60038.

Probabilistic Time-Varying Parameter Identification for Load Modeling: A Deep Generative Approach

The uncertainty of power resources introduces significant challenges for classic load modeling approaches. Moreover, load parameter identification techniques are affected by various load components with highly nonlinear time-varying behaviors and dependencies. This article presents a new deep generative architecture (DGA) based on the long short-term memory (LSTM) network for probabilistic time-varying parameter identification (PTVPI). In contrast to previous methods that merely compute point estimations of load parameters, our objective is to learn the continuous probability density function (PDF) of load parameters for composite load modeling (CLM) with ZIP load and induction motor. The proposed DGA learns complex temporal patterns from the time-varying parameters/measurements to estimate load parameters in a probabilistic fashion. Leveraging the LSTM network, our DGA computes deep temporal states and state transitions of load parameters. An encoding neural network extracts useful latent variables from the captured temporal states that are further mapped by a decoding neural network into the observed load parameters; hence, learning the underlying PDF of these parameters. Numerical results on the 68-bus New England and New York Interconnect System with four CLMs show accurate results for PTVPI in terms of various probabilistic estimation metrics, including reliability, sharpness, and continuous ranked probability score.

Global Sensitivity Analysis in Load Modeling via Low-Rank Tensor

Growing model complexities in load modeling have created high dimensionality in parameter estimations, and thereby substantially increasing associated computational costs. In this paper, a tensor-based method is proposed for identifying composite load modeling (CLM) parameters and for conducting a global sensitivity analysis. Tensor format and Fokker-Planck equations are used to estimate the power output response of CLM in the context of simultaneously varying parameters under their full parameter distribution ranges. The proposed tensor structured is shown as effective for tackling high-dimensional parameter estimation and for improving computational performances in load modeling through global sensitivity analysis.

Mathematical Representation of WECC Composite Load Model

Composite load model of Western Electricity Coordinating Council (WECC) is a newly developed load model that has drawn great interest from the industry. To analyze its dynamic characteristics with both mathematical and engineering rigors, a detailed mathematical model is needed. Although composite load model of WECC is available in commercial software as a module and its detailed block diagrams can be found in several public reports, there is no complete mathematical representation of the full model in literature. This paper addresses a challenging problem of deriving detailed mathematical representation of composite load model of WECC from its block diagrams. In particular, we have derived the mathematical representation of the new DER_A model. The developed mathematical model is verified using both MATLAB and PSS/E to show its effectiveness in representing composite load model of WECC. The derived mathematical representation serves as an important foundation for parameter identification, order reduction and other dynamic analysis.

Observed behavior of distributed photovoltaic systems during major voltage disturbances and implications for power system security

As distributed photovoltaics (PV) levels increase around the world, it is becoming apparent that the aggregate behavior of many small-scale PV systems during major power system disturbances may pose a significant system security threat if unmanaged. Alternatively, appropriate coordination of these systems might greatly assist in managing such disturbances. A key issue is PV behavior under extreme voltage events. PV connection standards typically specify aspects of inverter voltage behavior. However unresolved questions remain regarding compliance, ambiguity and transition between versions of these standards. In addition, how major voltage disturbances manifest in the low voltage network is complex, and analysis of operational system data could be particularly useful for establishing the behavior of distributed PV in the field. Our study utilizes 30 s operational PV generation data from 376 sites during two major voltage disturbances in Australia. Australia has one of the highest penetrations of distributed PV worldwide, and as such provides a useful case study. Results show that an aggregate ~30–40% reduction in distributed PV generation occurred during these events, but individual inverter behavior varied markedly. To the authors’ knowledge, this is the first time the aggregate response of distributed small-scale PV to voltage disturbances originating in the transmission system has been demonstrated. Four novel techniques for analyzing events are proposed. Results show a potential increase in system security service requirements as distributed PV penetrations grow. Our findings would seem to have major implications for development of composite load models used by power system operators and for contingency management.

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

With the increasing complexity of modern power systems, conventional dynamic load modeling with ZIP and induction motors (ZIP + IM) is no longer adequate to address the current load characteristic transitions. In recent years, the WECC composite load model (WECC CLM) has shown to effectively capture the dynamic load responses over traditional load models in various stability studies and contingency analyses. However, a detailed WECC CLM model typically has a high degree of complexity, with over one hundred parameters, and no systematic approach to identifying and calibrating these parameters. Enabled by the wide deployment of PMUs and advanced deep learning algorithms, proposed here is a double deep Q-learning network (DDQN)-based, two-stage load modeling framework for the WECC CLM. This two-stage method decomposes the complicated WECC CLM for more efficient identification and does not require explicit model details. In the first stage, the DDQN agent determines an accurate load composition. In the second stage, the parameters of the WECC CLM are selected from a group of Monte-Carlo simulations. The set of selected load parameters is expected to best approximate the true transient responses. The proposed framework is verified using an IEEE 39-bus test system on commercial simulation platforms.

PMU-Based Monitoring and Mitigation of Delayed Voltage Recovery Using Admittances

This paper analyzes the delayed voltage recovery phenomenon by simplifying the Western Electricity Coordinating Council (WECC) composite load model using the load admittance and the thermal relay dynamics. From this analysis, a closed form expression approximating the recovery time is derived and the key load parameters impacting the behavior of voltage recovery are identified. A monitoring scheme based on the measured load admittance is then proposed to detect the onset of stalling, even in the presence of voltage oscillations, and estimate the duration of the delayed voltage recovery. A mitigation scheme utilizing smart thermostats and offline learning is also derived to ensure that the voltage recovers to the pre-contingency voltage within a specified time. Both the monitoring and mitigation schemes only need local measurements at a substation making them promising for online applications. Results for the monitoring and mitigation schemes are described in detail for the IEEE 162 bus system validating the various assumptions used for the analysis and establishing the connection between the delayed voltage recovery phenomenon and load admittance.

Deep Learning-Based Time-Varying Parameter Identification for System-Wide Load Modeling

The integration of uncertain power resources is causing more challenges for traditional load modeling research. Parameter identification of load modeling is impacted by a variety of load components with time-varying characteristics. This paper develops a deep learning-based time-varying parameter identification model for composite load modeling (CLM) with ZIP load and induction motor. A multi-modal long short-term memory (M-LSTM) deep learning method is used to estimate all the time-varying parameters of CLM considering system-wide measurements. It contains a multi-modal structure that makes use of different modalities of the input data to accurately estimate time-varying load parameters. An LSTM network with a flexible number of temporal states is defined to capture powerful temporal patterns from the load parameters and measurements time series. The extracted features are further fed to a shared representation layer to capture the joint representation of input time series data. This temporal representation is used in a linear regression model to estimate time-varying load parameters at the current time. Numerical simulations on the 23- and 68-bus systems verify the effectiveness and robustness of the proposed M-LSTM method. Also, the optimal lag values of parameters and measurements as input variables are solved.

Multi-Objective Demand Side Management at Distribution Network Level in Support of Transmission Network Operation

This paper proposes a methodology for advanced demand side management (DSM) in the distribution network (DN), catering at the same time for the requirements of the network operator, transmission or distribution, the available flexibility of the demand side, and the preservation of network performance. The basic premise of the study is that the distribution network operator is providing flexibility to the transmission system operator through load shaping. Different, one or more, indicators can be chosen to assess preservation of network performance. In this paper, for illustrative purposes of the methodology, the steady state voltage stability index, namely the load margin, is chosen as the network performance indicator. This is evaluated before and after a DSM action, in order to analyse the possible effect of DSM on network loadability. Load at each DN bus is represented using a realistic composite load model comprising controllable and uncontrollable loads. The DSM is initially performed based on optimal power flow calculation to ensure that the distribution system load profile at the grid supply point follows the required (pre-specified) load profile during a 24 hour period. Following this, a particle swarm optimisation is used to modify DSM program in those time steps of the planning horizon where load margin is reduced. The methodology is illustrated on a number of case studies using modified IEEE 33 bus distribution network model. The results illustrate the extent to which the available flexibility of demand can be used in support of transmission network operation when the preservation of network performance and appropriate load modeling are considered.

Aggregated dynamic model of active distribution networks for large voltage disturbances

This paper proposes a “grey-box” aggregated dynamic model for active distribution networks, taking into account a heterogeneous fleet of generation technologies alongside their expected behavior when taking into account the latest European grid codes requirements in terms of voltage support services. The main goal of the proposed model and underlying methodology for its identification is to represent the transient behavior of the active distribution system following large voltage disturbances occurring at the transmission side. The proposed aggregated model is composed by three main components: an equivalent power converter for generation and battery energy storage systems portfolio representation; an equivalent synchronous generation unit; and an equivalent composite load model. The model’s parameters are estimated by an evolutionary particle swarm optimization algorithm, by comparing a fully-detailed model of a distribution network with the aggregated model’s frequency domain’s responses of active and reactive power flows, at the boundary of transmission-distribution interface substation.

Analysis and study of two‐dimensional parameter bifurcation of wind power farms and composite loads

AbstractThis study focuses on the stability of power system based on codimension‐two bifurcation theory. In this paper, we investigate the impact of load modeling on permissible wind power generation margins in distribution networks. The study considers codimension‐two bifurcations of equilibria and limit cycles in wind power systems depending on varying two parameters simultaneously. The principle parameter is the wind power generation, and the other parameter depends on the different types of loads. The types of loads are ZIP, exponential recovery, dynamic induction loads, and composite load models. To study the effects of the induction motor loads, the proportion of the static component in the motor load is changed and assessed with respect to their mechanical loads. Wind generation margin boundaries are traced, and saddle‐node, Hopf, and limit‐induced bifurcation branches are obtained, delimiting the stable and unstable operating regions in the parameter space. The analysis presented in this paper can pave the way for determining methods for improving and monitoring these margins with consideration to the system parameters and load composition.

A generic modeling and development approach for WECC composite load model

The composite load model (CLM) proposed by the Western Electricity Coordinating Council (WECC) is gaining increasing traction in industry, particularly in North America. At the same time, it has been recognized that further improvements in structure, initialization and aggregation methods are needed to enhance model accuracy and flexibility. However, the lack of a generic modeling and open-source implementation of the WECC CLM has become a roadblock for many researchers for further improvement. To bridge this gap, this paper presents a generic CLM modeling approach and the first open reference implementation of the WECC CLM. Individual load components and the CLM are first developed and tested in MATLAB, then translated to the high-performance computing (HPC) based, parallel simulation framework – GridPACK™. The main contributions of the paper include: (1) presenting important yet undocumented details of modeling and initializing the CLM, particularly for a parallel simulation framework like GridPACK™; (2) implementation details of the load components such as the single-phase air conditioner motor; (3) an improved initialization method; (4) implementing the CLM in a modular and extensible manner. The implementation has been tested at both the component and system levels, and benchmarked against commercial simulation programs with satisfactory accuracy.

Bayesian Estimation on Load Model Coefficients of ZIP and Induction Motor Model

Parameter identification in load models is a critical factor for power system computation, simulation, and prediction, as well as stability and reliability analysis. Conventional point estimation based composite load modeling approaches suffer from disturbances and noises, and provide limited information of the system dynamics. In this work, a statistics (Bayesian Estimation) based distribution estimation approach is proposed for both static and dynamic load models. When dealing with multiple parameters, Gibbs sampling method is employed. The proposed method samples all parameters in each iteration and updates one parameter while others remain fixed. The proposed method provides a distribution estimation for load model coefficients and is robust for measuring errors. The proposed parameter identification approach is generic and can be applied to both transmission and distribution networks. Simulations using a 33-feeder system illustrated the efficiency and robustness of the proposal.