G.A.N. Mbamalu

The generalized Kalman filter approach to adaptive underwater target tracking

The generalized Kalman filtering (GKF) method is applied to underwater target tracking. The proposed GKF is based on the formulation developed by J.C. Lagarias and F. Aminzadeh (1983) establishing a tradeoff between the cost associated with estimation error and the cost related to the lateral discontinuity of the estimates. By assigning proper weights for accuracy and stability in the objective function, the desired balance between accuracy of estimates and lateral continuity is achieved. Computational results illustrate the performance of the technique. Conclusions as to the effects of the accuracy and stability weights are drawn. >

Stochastic optimal load flow using Newton-Raphson iterative technique

ABSTRACT In this paper we consider the problem of stochastic optimal power flow for an all thermal electric power system. The formulation includes the effects of variables uncertainty modeled as stochastic and normally distributed. The system performance is represented by the load flow equations in polar form, reformulated to include the uncertainty aspect. Inequality constraints are handled using Powells penalty function method. We employ a variational approach to derive the optimality conditions resulting in a set of nonlinear equations. Newtons method is employed, to actually implement the optimum strategy by solving the nonlinear set of optimality conditions iteratively. An application example involving the IEEE standard 14 bus system is presented and the stochastic solution is compared with a deterministic solution for the same system in order to demonstrate the applicability of the proposed method.

Stochastic optimal load flow using a combined quasi-Newton and conjugate gradient technique

Abstract This paper treats the problem of optimal power flow in electric power systems using a method which includes the effects of uncertain variables in formulating the problem for an all thermal electric power system. The method assumes that the system power demand is random and is normally distributed with zero mean and unit variance. The equality constraints associated with the formulation are the power flow equations in polar form utilizing the nodal admittance matrix approach. The non-linear programming technique of Powells penalty function is used to allow the incorporation of inequality constraints in the formulation. The resulting stochastic optimal power flow problem reduces to one of solving a set of non-linear equations. Here we use a technique that combines the quasi-Newton method and the conjugate gradient method in what is referred to as the CONMIN algorithm. Computational implementation results involving four IEEE standard test networks are given to demonstrate the validity of this method.

Dynamic heave compensation using robust estimation techniques

Abstract Compensating for underwater motion effects (heave component) arises in a number of activities. Earlier treatments relied on frequency response methods to model the phenomenon and subsequently applied Kaiman filtering for the compensation task. The success of the least squares estimation procedures depends on assuming that the estimation errors resulting from fitting a model to a set of data follow a Gaussian distribution. For non-Gaussian errors, the method of least squares may not perform satisfactorily, and robust regression procedures appear to perform better in this case. An alternative model of the heave process based on an autoregressive integrated moving average representation of the underlying time series is proposed. Robust estimation of the models parameters is conducted using the iteratively reweighted least squares technique employing four weighting functions. A computational comparison of the performance of the four resulting estimators is conducted. Illustrative results are shown for a base case of a real heave record and synthetically derived records in Monte Carlo simulations. We use both time and frequency domain measures of estimation accuracy to rank the proposed estimators. We conclude that an ARIMA (3,1,2) model with parameters estimated using Andrews weighting function gives better results.

Effects of prioritizing demand on optimal load shedding policy

Demand priorities in the power system optimal load shedding policy are major decision variables used to regulate the amount of load shedding at specific buses during contingencies. Previous work on optimal load shedding does not explicitly consider demand priorities. This paper investigates the effects of demand priorities on the performance of the power system during emergencies. A nonlinear programming approach is used to determine the optimal values of demand priorities for two test systems. Optimizing demand priorities is shown to enhance overall system performance measured by total supplied load, bus voltages and system losses during the contingency. Results are offered demonstrating additional enhancements by optimizing the demand priority at the bus with the second largest load.

MINIMUM EMISSIONS HYDRO-THERMAL POWER FLOW

ABSTRACT Minimum NOx emission optimal power flow scheduling for a hydro-thermal power system is considered. Computational results using Newtons method and Powells penalty function approach arc obtained for the 57 bus and 118 bus hypothesized IEEE hydro-thermal power systems. Based on a comparison with the minimum cost solution, it is concluded that regardless of which scheduling strategy is used, increasing the hydro generation results in decreasing both emissions and thermal fuel costs. The results obtained indicate that system N O x emissions are influenced by bus voltage lower limits, reactive power generation and the hydro unit water conversion factor.

NOx emission modelling using the iteratively reweighted least-square procedures

Abstract NO x emission dispatching is a method to reduce the amount of oxide of nitrogen produced by thermal generating units to meet given power demands. This requires adequate models relating NO x emissions to the active power generation of the unit. Least-squares estimation procedures work best when measurement errors are Gaussian. For the non-Gaussi an errors and Gaussian errors, the iteratively reweighted least-squares (IRWLS) procedure with optimized weight functions and their associated weight constants can be used to estimate, refine and fine tune parameters of identified models. The form of the emissions model is restricted to those with monotonically increasing derivatives due to subsequent minimization process requirements. We explore the application of some transformations to the form of model chosen. Moreover, emission models are developed using the IRWLS procedure. The results obtained using the iteratively reweighted least-square procedures are compared with the least squares.

Effects of load modeling on minimum loss, minimum emission, and multiple-objective optimal hydrothermal power flow

Abstract The effects of incorporating load models in three formulations of hydrothermal optimal power flow are considered in this paper. The formulations are transmission loss minimization, NO x minimization, and the multiple objectives of minimum NO x emission and minimum cost. The conventional algorithms and those incorporating load models were tested using a 14-bus and a 30-bus test system. Solutions are obtained using the MINOS optimization package. The computation time requirements of the conventional algorithm are lower than those incorporating load models. In all cases, there are measurable differences in the optimal voltage magnitudes for the systems tested. Incorporating load models yields lower active power generations and transmission losses, lower thermal fuel costs and reduced environmental impact than the conventional formulations.

A PSEUDO-INVERSE-BASED PROBABILISTIC POWER FLOW APPROACH

Abstract A probabilistic formulation of the power flow problem that includes the effects of uncertainty is given. The random variations of active and reactive loads are assumed to be normally distributed with zero mean μΔPD= 0 and some variance σΔPD≠0 The resulting probabilistic active and reactive power flow equations differ from the conventional active and reactive power flow equations due to the presence of some augmented terms. As a result, the number of equations resulting from this formulation is less than the number of the unknowns. The structure of the traditional Jacobian matrix encountered in the conventional power flow problem is altered, and the resulting problem is precisely a minimization problem. A minimal least squares based minimization technique and the Newtons method of successive approximation are proposed to obtain solutions. Test results for the IEEE 5 bus and 14 bus test systems are documented

Computational Experience With The Generalized Kalman Filter For Dynamic Heave Compensation

We treat the generalized Kalman Filter (GKF) approach to analyze heave dynamics data more effectively. An optimum solution that simultaneously accounts for the accuracy of the estimates and stability (lateral continuity) is achieved. Conventional Kalman Filters (KF) have been used in source heave compensation. One of the roblems continuity, since each trace is analyzed separately (single channel o eration). The approach establishes a trade off cost related to the lateral discontinuity of the estimates. By assigning pro er weights for accuracy and stability (WA and U’S) in the oRjective function the desired balance between accuracy and stability is achieved. Computational results are offered to illustrate the trade-offs involved. associatedwithKFis the difficulties inmaintaining t Tl elateral between t K e cost associated with estimation error and the