Synthetic residential load models for smart city energy management simulations

Abstract

The ability to control tens of thousands of residential electricity customers in a coordinated manner has the potential to enact system-wide electric load changes, such as reduce congestion and peak demand, among other benefits. To quantify the potential benefits of demand-side management and other power system simulation studies (e.g. home energy management, large-scale residential demand response), synthetic load datasets that accurately characterise the system load are required. This study designs a combined top-down and bottom-up approach for modelling individual residential customers and their individual electric assets, each possessing their own characteristics, using time-varying queueing models. The aggregation of all customer loads created by the queueing models represents a known city-sized load curve to be used in simulation studies. The three presented residential queueing load models use only publicly available data. An open-source Python tool to allow researchers to generate residential load data for their studies is also provided. The simulation results presented consider the ComEd region (utility company from Chicago, IL) and demonstrate the characteristics of the three proposed residential queueing load models, the impact of the choice of model parameters, and scalability performance of the Python tool.

Day‐ahead continuous double auction‐based peer‐to‐peer energy trading platform incorporating trading losses and network utilisation fee

“…Because of the randomness of the Poisson process inter‐arrival times, a large number of appliances can arrive during a very short time interval that may result in unrealistic peaks. If necessary, an implementer of this P2P framework could limit this overshoot via the M t / G / C model, where C defines a hard limit on defined capacity for each household in the network [34].…”

“…The queueing model used in this research is developed in accordance with Ref. [34] and uses the characteristics as defined by M t / G / ∞ , where M t represents the time varying Poisson process to model the inter‐arrival time of the devices, G represents that the probability distribution used for the service time is considered to be general (electric load run times can follow any probability distribution or mass), and ∞ shows that the capacity for the individual customers is taken to be infinite. The time‐varying Poisson process λ ( t ) to model the appliance rate of the system is given as follows: $$$\lambda (t)=\frac{L(t+E[d])}{E[d]E[p]}$$$ …”

“…The queueing load model is used to model the electric demand curves for the market participants [33,34], which uses queueing theory initially created for the communication sector. The major components of a queueing model are the probability distributions defined for the service time and inter-arrival time, the queue capacity, the size of the population, and the number of servers.…”

“…The readers are encouraged to go through the mathematical details of the developed queueing load model presented in Ref. [34].…”

Overvoltage Prevention and Curtailment Reduction Using Adaptive Droop-Based Supplementary Control in Smart Inverters

“…The PV size was 5 kW p for all the twelve houses. The household daily load profiles were estimated using Poisson process and Queue theory and this method of estimation has been presented in [38][39][40]. These techniques model the energy consumption profile for each household appliances based on their consumption pattern which is summed up to get the load profile of each house.…”