Parallel self-organising hierarchical neural network-based fast voltage estimation (original) (raw)
Related papers
Computers & Electrical Engineering, 2003
Increased interconnections and loading of power systems, sometimes, lead to insecure operation. Since insecure cases often represent the most severe threats to secure system operation, it is important that the user be provided with a measure for quantifying the severity of the cases both in planning and operational stages of a power system. The Euclidean distance to the closest secure operating point has been used as a measure of the degree of insecurity. Recently, artificial neural networks are proposed increasingly for complex and time-consuming problems of power system. This paper presents a parallel self-organised hierarchical neural network based approach for estimation of the degree of voltage insecurity. Angular distance based clustering is used to select the input features. The proposed method has been tested on IEEE 30-bus system and a practical 75-bus Indian system and found to be suitable for real time implementation in Energy management centre.
Two-phase neural network based estimation of degree of insecurity of power system
Computers & Electrical Engineering, 2003
Increased loading and contingencies often lead to situations where the optimal power flow solution no longer remains within the secure region. In such situations there is a need of determining control actions to be taken quickly, as otherwise the system may become unstable. Hence it is important to quantify the degree of insecurity of the power system both in planning as well as at operational stages. The distance in parameter space between an insecure operating point and the closest point on feasible (secure) hypersurface has been used as a measure of degree of insecurity. A method based on two-phase optimization neural network has been presented to compute the degree of insecurity and the voltages and angles at all the buses of the system corresponding to the closest secure point. Inclusion of security limits on power system variables assures a solution representing a secure system. When compared with conventional non-linear optimization techniques, the proposed neural network is superior, as it can be easily implemented using digital hardware and is highly suitable for real time implementation in energy management system. The proposed method has been tested on IEEE 30-bus test system and a practical 75-bus Indian system. The results achieved are compared with results from a conventional method. Insecurity arising due to increase in load and contingencies has been considered in this work.
A neural network-based method for voltage security monitoring
IEEE Transactions on Power Systems, 1996
In this paper, a neural network-based method is proposed for monitoring on-line voltage security of electric power systems. Using a dynamic model of the system, voltage stability is measured totally, considering a suitable stability index for the whole system, and locally, by defining appropriate voltagemargins for detecting the area of the system where the instability phenomenon arises. A three-layer feedforward neural network is trained to give, as outputs to a pre-defied set of input variables, the expected values of the above defined indices. The neural network is designed by using a fast learning strategy that allows the optimal number of hidden neurons to be easily determined. Moreover, it is shown that, in the operation mode, the system power-margin and the bus power-margins can be easily evaluated using the value of the voltage stability index given by the designed NN. The effectiveness of the proposed approach has been demostrated on the IEEE 118-bus test system.
Static Security Assessment of Power System Using Self Organizing Neural Network
2002
A Neural Network aided solution to the problem of static security assessment of a model power system is proposed. This paper utilizes the artificial neural net of Kohonen’s Self Organizing Feature Map (SOFM) technique that transforms input patterns into neurons on the two dimensional grid to classify the secure / insecure status of the power system. The SOFM uses the line flows under different component cases as inputs and perform self-organization to obtain the cluster of the components based on their loading limits. The output of the SOFM provides information about the violation of the constraints from which the operating state of the power system can be identified as the result of which the system can be classified as secure or insecure. The above method of security assessment was successfully tested for a model 3-generator, 6 bus system and latter extended to an IEEE 14 bus and 30 bus systems respectively.
Optimal Fuzzy Self-Organizing Structure for Voltage Security Margin Estimation
2005
In recent years, research efforts have been focused to estimating voltage security margins. Voltage security margin shows how close the current operating of a power system is to a voltage collapse point. One main disadvantage of these techniques is that they require a large amount of computations, therefore they are not efficient for on-line use in power control centers. Therefore the intelligent networks classification techniques for systems that are illustrated by series of data are offered. In this paper, two methods are used to estimate voltage security margin. The first method is general fuzzy minmax neural network (GFMM NNs.) with on-line adaptation. The second method which we named it as "Fuzzy Self-Organizing Network," combines two structures of Kohonen and GFMM neural network then by using Akaike criterion the optimal values of the proposed network parameters are determined. The data set needed for training this structure is obtained from the minimum singular value of the power flow Jacobian matrix. These methods are applied on the IEEE 30-bus system with 2000 simulated data randomly generated from different operating conditions. Finally, the results compared with the three-layer feed forward neural network as the most common used neural networks. The results clearly show the advantage and high efficiency of the proposed structure. Key words: kohonen neural network, general fuzzy Min-Max neural network, akaike information criteria, power systems, voltage security margin estimation 15th PSCC, Liege, 22-26 August 2005 Session 30, Paper 6,
Power system static security assessment using self-organizing neural network
2006
Artificial neural network approach to the problem of static security assessment of power system is presented. This paper utilizes the artificial neural net of Kohonen's self-organizing feature map (SOFM) technique that trans- forms input patterns into neurons on the two-dimensional grid to classify the secure/insecure status of the power system. SOFM uses the line flows under different component cases as inputs and self-organizes to obtain the clus- ter of the components based on their loading limits. The output of SOFM provides information about the violation of the constraints from which the operating state of the power system can be identified, which can be classified as secure or insecure. The proposed method of security assessment was initially demonstrated for a model 3 genera- tor 6-bus system and later extended to IEEE-14, -30 and -57 bus systems.
Contingency ranking for voltage collapse using parallel self-organizing hierarchical neural network
International Journal of Electrical Power & Energy Systems, 2001
On-line monitoring of the power system voltage security has become a vital factor for electric utilities. This paper proposes a voltage contingency ranking approach based on parallel self-organizing hierarchical neural network (PSHNN). Loadability margin to voltage collapse following a contingency has been used to rank the contingencies. PSHNN is a multi-stage neural network where the stages operate in parallel rather than in series during testing. The number of ANNs required is drastically reduced by adopting a clustering technique to group contingencies of similar severity into one cluster. Entropy based feature selection has been employed to reduce the dimensionality of the ANN. Once trained, the proposed ANN model is capable of ranking the voltage contingencies under varying load conditions, on line. The effectiveness of the proposed method has been demonstrated by applying it for contingency ranking of IEEE 30-bus system and a practical 75-bus Indian system. q
International Journal of Electrical Power Energy Systems, 2012
This paper presents the application of cascade neural network (CANN) based approach for integrated security (voltage and line flow security) assessment. The developed cascade neural network is a combination of one screening module and two ranking modules, which are Levenberg-Marquardt Algorithm based neural networks (LMANNs). All the single line outage contingency cases are applied to the screening module, which is 3-layered feed-forward ANN having two outputs. The screening module is trained to classify them either in critical contingency class or in non-critical contingency class from the viewpoint of voltage/line loading. The screened critical contingencies are passed to the corresponding ranking modules, which are developed simultaneously by using parallel computing. Parallel computing deals with the development of programs where multiple concurrent processes cooperate in the fulfillment of a common task. For contingency screening and ranking, two performance indices: one based on voltage security of power system (VPI) and other based on line flow (MWPI) are used. Effectiveness of the proposed cascade neural network based approach has been demonstrated by applying it for contingency selection and ranking at different loading conditions for IEEE 30-bus and a practical 75-bus Indian system. The results obtained clearly indicate the superiority of the proposed approach in terms of speedup in training time of neural networks as compared to the case when the two ranking neural networks were developed sequentially to estimate VPI and MWPI.
Classification of Voltage Security States Using Unsupervised ANNs
Energy Systems in Electrical Engineering, 2015
This chapter focuses on the application of self-organizing neural networks that are capable of extracting valuable data from their working surroundings. The basic role of self-organization lies in the invention of significant patterns without the intervention of a teaching input. An important aspect of the implementation of such a system is that all adaptations must be based on the data that are accessible locally to the neural connection from the pre-and postsynaptic neuron signals and activations. Self-organization must lead eventually to a state of knowledge that provides useful information concerning the environment from which patterns are drawn. As an alternative to the multilayer perceptron, Kohonen's self-organizing neural network offers some advantages, particularly in clustering-type applications. Faster learning rate and straightforward interpretation of the classification results make self-organizing map (SOM) an ideal choice for the classification of voltage security states in multi-bus power networks. This chapter describes an artificial neural network-based approach, in which Kohonen's self-organizing feature map technique has been applied to classify the power system operating states based on their degree of static voltage stability.
ANN based integrated security assessment of power system using parallel computing
International Journal of Electrical Power & Energy Systems, 2012
This paper presents the application of cascade neural network (CANN) based approach for integrated security (voltage and line flow security) assessment. The developed cascade neural network is a combination of one screening module and two ranking modules, which are Levenberg-Marquardt Algorithm based neural networks (LMANNs). All the single line outage contingency cases are applied to the screening module, which is 3-layered feed-forward ANN having two outputs. The screening module is trained to classify them either in critical contingency class or in non-critical contingency class from the viewpoint of voltage/line loading. The screened critical contingencies are passed to the corresponding ranking modules, which are developed simultaneously by using parallel computing. Parallel computing deals with the development of programs where multiple concurrent processes cooperate in the fulfillment of a common task. For contingency screening and ranking, two performance indices: one based on voltage security of power system (VPI) and other based on line flow (MWPI) are used. Effectiveness of the proposed cascade neural network based approach has been demonstrated by applying it for contingency selection and ranking at different loading conditions for IEEE 30-bus and a practical 75-bus Indian system. The results obtained clearly indicate the superiority of the proposed approach in terms of speedup in training time of neural networks as compared to the case when the two ranking neural networks were developed sequentially to estimate VPI and MWPI.