Yung-Tien Chen

Measurement-based dynamic load models: derivation, comparison, and validation

Accurate load modeling is essential for power system dynamic simulation. In this paper, four dynamic load models are identified based on multiple online measurement data from the Taiwan Power System. The performances in modeling real and reactive power behaviors by dynamic and selected static load models are evaluated. Parameter variation with respect to different loading conditions is analyzed. A simple and efficient method is presented to estimate a representative parameter set for different loading conditions. The cross-validation technique is applied to validate the four dynamic load models in order to obtain a better estimate of their performance. Numerical studies indicate that linear dynamic load models studied in this paper give better results than two nonlinear dynamic load models in modeling reactive power behaviors during disturbance while they are comparable in modeling real power behaviors

Comparison of simulated power system dynamics applying various load models with actual recorded data

Time domain stability simulation results using derived load models were compared with corresponding data obtained from transient disturbance recorders to validate the accuracy of the load models. Observations of real and reactive power load responses at primary substations and real and reactive power flows on major trunk lines indicated that using a composite (dynamic and static) load model provides a more accurate representation than other models, and closely matches the actual recorded values from the Taipower System. A severe fault, during which enormous voltage and frequency excursions occurred in the system, was extensively examined considering load rejection due to large voltage drops. >

Development of a dynamic ZIP-motor load model from on-line field measurements

Abstract We developed, using on-line field measurements, a dynamic ZIP-motor load model which attempts to capture the dynamic behavior of loads. The dynamic ZIP-motor load model consists of two parts: a static part and a dynamic part, where the static load is represented by a combination of constant impedance, constant current and constant power while the dynamic load is represented by an induction motor model. A procedure for building the model structure of the dyanmic ZIP-motor load model is developed. An effective solution algorithm for estimating the associated model parameters is presented. A notable feature of the algorithm is that the model parameters are estimated based upon pseudo-gradient information instead of exact gradient calculation. The proposed algorithm is easy to implement and can be used to identify nonlinearities associated with ZIP-motor models, such as models accounting for magnetizing and leakage reactances. The results developed in this paper are tested and evaluated using real field measurements from a real power system. It is shown numerically in the case study that the developed dynamic load model can accurately capture the dynamic behaviors of loads.

On-line measurement-based model parameter estimation for synchronous generators: model development and identification schemes

Accurate generator modeling allows for more precise calculations of power system control and stability limits. In this paper, a procedure using a set of measured data from an online plant transient recording and analysis system to develop the synchronous generator model for the Taipower system is described. A continuous-time transfer function matrix is derived for a popular sixth-order synchronous generator model. In order to accommodate the nature of online digital measurements, the transfer function matrix is transformed into a simple discrete-time linear regression model. A measure of discrepancy between the generator model outputs and the online measurements from generators is employed. A modified conjugate gradient method suitable for identifying generator parameters is developed to minimize the measure of discrepancy, from which a set of accurate generator parameter values can be obtained. The merits of the modified conjugate gradient method include its computational efficiency and numerical reliability. The proposed procedure allows simultaneous estimation of all generator parameter values. >

Development of composite load models of power systems using on-line measurement data

Load representation has a significant impact on power system analysis and control results. In this paper, composite load models are developed based on on-line measurement data from a practical power system. Three types of static-dynamic load models are derived: general ZIP-induction motor model, exponential-induction motor model and Z-induction motor model. For dynamic induction motor model, two different third-order induction motor models are studied. The performances in modeling real and reactive power behaviors by composite load models are compared with other dynamic load models in terms of relative mismatch error. In addition, numerical consideration of ill-conditioned parameters is addressed based on trajectory sensitivity. Numerical studies indicate that the developed composite load models can accurately capture the dynamic behaviors of loads during disturbance

Development of a micro-processor-based transient data recording system for load behavior analysis

A microprocessor-based transient data recording system for the investigation of actual load behavior during system disturbances is described. The system is currently installed at the primary substations and distribution substations in the Taipower system. Analytical tools for computing the corresponding voltage and current phasors from the instantaneous three-phase voltages and currents measured by the system as well as the corresponding real and reactive powers are presented. A procedure for applying the recording system to develop an accurate load model is detailed. >

Experience with the identification and tuning of excitation system parameters at the second nuclear power plant of Taiwan power company

Experience with the identification and tuning of exciter constants for a generating unit at the second nuclear power plant of the Taiwan Power Company is reported. A field test is first performed on the excitation system with the generator open-circuited. Since the field test results differ from the computer simulation results using manufacturers constants, the authors first modify the manufacturers constants based on their previous experience to reach a preliminary set of parameters for the excitation system. Then a hybrid nonlinear simulation-sensitivity matrix method is developed to further refine the excitation system parameters. The exciter constants are tuned in order to give better dynamic response before a power system stabilizer is applied to the generator. Field tests are then performed in order to compare the dynamic response of the generator without and with power system stabilizer.

On-line measurement-based model parameter estimation for synchronous generators: solution algorithm and numerical studies

In this paper, a sixth-order synchronous generator model identification technique from online measurements is considered. An algorithm is devised to identify the generator model. A constrained conjugate gradient method is incorporated into the algorithm to guarantee rapid convergence to the final solution. Using the algorithm, a complete generator model is derived from online measurements recorded by a plant transient recording system during a system disturbance. In addition, the algorithm does not greatly rely upon the accuracy of the initial estimates, allowing the initial estimates to deviate reasonably far from the true parameters. Detailed numerical studies of the Taipower system using raw and filtered data are included. >

Power system load ranking for voltage stability analysis

Load modeling is well known to have a significant impact on voltage stability analysis while it is not clear the degree of influence each load has on voltage stability analysis. In this paper the problems associated with quantifying the degree of influence and load ranking on voltage stability analysis are addressed. A computational method based on load margin for this numerical quantification is presented. A load ranking scheme is developed and applied to Taiwan power system to rank the degree of influence each nodal load of Taipower transmission network on load margin of voltage stability. Physical insights of top-ranked nodal loads are provided. Furthermore, the robustness of load ranking under different loading conditions, different contingencies and different power transfer patterns is investigated. Some applications of load ranking results are described

APPLYING BIFURCATION ANALYSIS TO DETERMINE OPTIMAL PLACEMENTS OF MEASUREMENT DEVICES FOR POWER SYSTEM LOAD MODELING

Load modeling is well known to have a significant impact on power system analysis. The task of load modeling, however, is time-consuming and expensive. Accurate load models should be developed for loads at critical locations. In this paper, applying bifurcation analysis, the problem of optimal placements of measurement devices for load model development from the viewpoint of voltage stability analysis is investigated. Voltage instability/collapse is modeled using bifurcation theory first. An optimal placement problem is formulated. An optimal placement identification scheme is proposed and applied to Taiwan power system. Optimal placements of measurement devices are identified. Validation of the selected optimal placements is performed. The robustness of optimal placements under different power transfer patterns is also examined.