Search

Numerical models play important roles in power system operation. They are widely used for planning studies to identify and mitigate issues, determine transfer capability, and develop transmission reinforcement plans. These models need to be accurate and updated regularly to serve these purposes faithfully over time. In this paper, we formulate the problem of parameter calibration for machine models in a power system into the framework of reinforcement learning and demonstrate the feasibility of applying Deep Deterministic Policy Gradient (DDPG) for a two-parameter generator model calibration on a 4-bus system. To improve the efficiency and accuracy of DDPG, we introduce memory forgetting mechanism and dynamic range adjustment (DRA) into the original DDPG, i.e., DRA-DDPG. To reduce the parameter estimation errors due to partially observable disturbance states in the power system, we introduce the concept of maximal K-Nearest-Neighbor (KNN) reward to enable our reinforcement learning algorithm to accommodate a finite set (K) of unknown disturbance states in the system. Our experimental results show that the proposed DRA-DDPG outperforms the baseline DDPG in terms of accuracy and efficiency and the proposed maximal KNN reward is well-suited for resolving the uncertainties from partially observable system states.

In power systems, having accurate device models is crucial for grid reliability, availability, and resiliency. Existing model calibration methods based on mathematical approaches often lead to multiple solutions due to the ill-posed nature of the problem, which would require further interventions from the field engineers in order to select the optimal solution. In this paper, we present a novel deep-learning-based approach for model parameter calibration in power systems. Our study focused on the generator model as an example. We studied several deep-learning-based approaches including 1-D Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU), which were trained to estimate model parameters using simulated Phasor Measurement Unit (PMU) data. Quantitative evaluations showed that our proposed methods can achieve high accuracy in estimating the model parameters, i.e., achieved a 0.0079 MSE on the testing dataset. We consider these promising results to be the basis for further exploration and development of advanced tools for model validation and calibration.

Air-conditioning (AC) systems have the highest power consumption among the appliances and consumer devices used at residential homes and buildings. Reducing their energy use will lower peak time usage and lower CO2 emissions. Recently, employment of the Information and Communications Technologies (ICT) to the power grid has smartened the grid. In the smart grid new opportunities emerge for AC energy consumption control. The aim of this paper is to reduce the air conditioning energy consumption of residential customers. It proposes an architecture that provides easy management and control using sensor network web services. A simulation thermal model of a house considers house data and outside temperature is presented. Simulation results showed a proposed temperature control method can have significant energy saving while maintaining customer comfort.

Smart Grid laboratory (SGL) has been established at Buffalo State College as a collaborative effort of two academic institutions, University at Buffalo and Buffalo State College (UB-BSC). It is equipped with state-of-the-art equipment and serves as a hands-on teaching tool for undergraduates, as well as a research lab for graduate students. This paper discusses the development of a Static VAR Compensator (SVC) for use with the SGL at Buffalo State College. An overview of the UB-BSC SGL is presented first. The SVC circuit and characteristics are discussed in brief. Next, the determination of SVC parameters is discussed, thoroughly. With these parameters, the modes of operation of the SVC have been simulated and the system has been implemented using UB-BSC SGL equipment, the procedures of this experiment are documented. Finally, the hardware results are compared with those of simulated results.

Preisach technique is a powerful modeling tool for analyzing the hysteresis phenomenon in magnetic materials. Therefore, it has been used to investigate the effect of magnetizing current harmonics upon hysteresis losses in a solid core. The simulation and analysis show that the effect of harmonics on hysteresis losses can be neglected.

This paper presents a thermal model presented by the author in previous work. The model is established to predict the top oil and hot spot temperatures in a transformer. The model is examined using data from three different transformers equipped with sensors to collect measured data. A non linear square is used to estimate some of the model parameters which are not known accurately. The model showed good results

This paper presents a thermal model to predict the top oil and hot spot temperatures of a transformer. The model is applied to 2500 kVA ONAN (without external cooling) distribution transformer equipped with thermo-couples and tested at varying load. A non linear least square algorithm is used based on measurements to determine the model parameters. The model showed good results.

The use of non-linear (solid state) loads in power system has increased the awareness of the potential reduction of a transformer life due to the increased losses. Manufacturers and users have been applying IEEE C57.110-1989 which gives a conservative estimate of loading capability of transformers. In this paper, the effect of harmonics on transformers is discussed and a thermal model to predict a transformer hot spot temperature is presented. The model considers a time varying harmonic load cycle and ambient temperature. The model is applied on 31.5 MVA, 115/spl times/9/spl times/1.67%/6.3 kV ONAF transformer.

This paper presents a thermal model of a power transformer in the form of simple equivalent circuit. The model is established to calculate the top oil and hot spot temperatures. The model parameters are obtained from manufacturer test data and an electromagnetic analysis using finite element method (FEM). The model results are compared with measured data obtained from 250 MVA, 230/spl plusmn/8/spl times/1.5%/118/21 kV ONAF transformer equipped with fiber optic probes and tested according to a load cycle comprising a short time load up to 2.1 pu. The results are additionally compared with the conventional equations of the IEEE clause 7.

In this paper a transformer thermal dynamic model for use in an on-line monitoring and diagnostic system is examined. The model is established to calculate the real time top oil and hot spot temperatures in the form of simple equivalent circuit, based on the fundamental heat transfer theory. A substation transformer was chosen and equipped with sensors to gather data. A non-linear least squares algorithm is used to determine the top oil model parameters that best fit the measurements for different cooling modes. The hot spot temperature is simulated. The model showed good results

The assumption that the transformer windings eddy loss rise in proportion to the square of the harmonic order is reasonable for small conductors and low harmonics with errors on the conservative side only for a combination of large conductors and higher harmonics. This paper presents a corrected harmonic loss factor by taking into account the depth of penetration, which leads to more accurate estimation of the increased transformer winding losses under non-sinusoidal load currents.