Michele Brucoli

Discrete-time cellular neural networks for associative memories with learning and forgetting capabilities

A synthesis procedure for associative memories using Discrete-Time Cellular Neural Networks (DTCNNs) with learning and forgetting capabilities is presented. The proposed design technique generates networks with the capability of learning new patterns and forgetting old ones without recomputing the whole interconnection matrix and the input vector. >

A Gauss-Jacobi-Block-Newton method for parallel transient stability analysis (of power systems)

A parallel method for the transient stability simulation of power systems is presented. The trapezoidal rule is used to discretize the set of algebraic-differential equations which describes the transient stability problem. A parallel Block-Newton relaxation technique is used to solve the overall set of algebraic equations concurrently on all the time steps. The parallelism in space of the problem is also exploited. Furthermore, the parallel-in-time formulation is used to change the time steps between iterations by a nested iteration multigrid technique, in order to enhance the convergence of the algorithm. The method has the same reliability and model-handling characteristics of typical dishonest Newton-like procedures. Test results on realistic power systems are presented to show the capability and usefulness of the suggested technique. >

A generalized approach to the analysis of voltage stability in electric power systems

Abstract In this paper a new method for analysing the voltage stability problem in electric power systems is presented. The approach starts with a linearized model in state space form of a multi-machine power system, then rigorous voltage stability conditions are derived on the basis of a suitable aggregated model of the original system capable of retaining the dynamics of voltages at generator and load nodes. The approach allows a systematic individualization of all the dynamic factors which affect voltage instability phenomena and suggests the appropriate representation which has to be adopted for each dynamic component of the system. The validity and usefulness of the suggested method are illustrated by carrying out simulation studies on a sample power system.

Quadratic probabilistic load flow with linearly modelled dispatch

Abstract A second-order probabilistic load flow technique is presented, which takes into account the effects of nonlinearities in the system equations and of different dispatching strategies. This approach uses as input a normal probability distribution for the loads. The probability distribution for the generated powers is computed by linearly modelling the dispatching activity. The technique of moments is then applied to second-order approximations for the load flow and current equations, allowing the noniterative computation of the means and standard deviations for the various output quantities. Generation outages are separately treated, with the possibility of embedding different postoutage dispatching strategies. The cumulative results are finally obtained by combining the pertinent probabilistic load flows with the occurrence probabilities of the various conditions. Results on two sample systems are given.

A global approach to the design of discrete‐time cellular neural networks for associative memories

In this paper a global design method for associative memories using discrete-time cellular neural networks (DTCNNs) is presented. The proposed synthesis technique enables to realize associative memories with several advantageous features. First of all, grey-level as well as bipolar images can be stored. Moreover, the proposed approach generates networks with learning and forgetting capabilities. Finally, it is possible to design networks with any kind of predetermined interconnection structure. In particular, neighbourhoods without line crossings can be chosen, greatly simplifying the VLSI implementation of the designed DTCNNs. In the first part of this work a model of a multilevel threshold network is presented and a stability analysis is carried out using basic notions deriving from non-linear dynamical system theory. The synthesis procedure is then developed by means of a pseudoinversion technique, assuring learning and forgetting capabilities of the designed DTCNN. The use of a neighbourhood without line crossings is also discussed. Simulation results are reported to show the capability of the proposed approach.

State space representation of interconnected power systems for dynamic interaction studies

Abstract In this paper a linearized model of an interconnected power system in state space form is presented in order to analyse the dynamic interactions of its components across the interconnection network. The model, modularly structured, proves flexible and includes detailed descriptions of generators with their control devices, static and dynamic loads. The mathematical representation of the system is formulated such that the effect of interconnection on system characteristic polynomial and system dynamic stability can be systematically investigated. Additionally, a suitable index is defined to measure the degree of dynamic interaction which may occur between the machines of the interconnected power system. A numerical example is included using a four-generator ten-bus system to illustrate the capability of the developed model and the usefulness of the proposed technique.

Heteroassociative memories via cellular neural networks

In this paper a synthesis procedure for heteroassociative memories using Cellular Neural Networks (CNNs) is presented. The suggested method, by assuring the condition of symmetry of the interconnection matrix, guarantees the complete stability of the designed network, besides providing that all the stored patterns correspond to asymptotically stable equilibrium points. Numerical examples are carried out to show the behaviour of the designed memory with respect to input perturbations. Moreover, the storage capacity and the presence of spurious equilibria have been investigated.

An approach to the design of space-varying cellular neural networks for associative memories

In this work a design of a space-varying cellular neural network (CNN) in order to behave as an associative memory is presented. To this purpose, a new class of space-varying cellular neural networks with a nonsymmetric interconnection structure is considered. A stability analysis is firstly carried out. Then, a learning algorithm, based on the relaxation method, is used to compute the feedback parameters of the considered network. Simulation tests are reported to confirm the validity of the suggested approach.

Discrete-time cellular neural networks for associative memories: a new design method via iterative learning and forgetting algorithms

In this paper a synthesis procedure of discrete-time Cellular Neural Networks (DTCNNs) for associative memories with iterative learning and forgetting algorithms is developed, by which each pattern to be stored is learnt one at a time and each pattern to be forgotten is deleted one at a time. The proposed approach exploits the properties of pseudo inverse matrices and preserves the local connection feature of DTCNNs.

A probabilistic approach to the voltage stability analysis of interconnected power systems

Abstract In planning a power system it is always necessary to assess whether a voltage collapse occurs during a prefixed system operating condition. However, present approaches to the analysis of voltage instability phenomena in interconnected power systems are deterministic and, consequently, they cannot take into account the unavoidable uncertainties associated with the bus load forecast. This is indeed an important limitation. In this case the application of probabilistic techniques is the most feasible alternative. On the basis of this observation, in this paper a probabilistic approach to the voltage stability analysis of interconnected power systems is presented; it treats loads as random uncorrelated variables with normal distributions. The method proves suitable for determining systematically, for each expected system operating condition, the statistics of all the node voltages which are critical from the voltage stability viewpoint. The capability and usefulness of the suggested approach are illustrated by carrying out simulation studies on a sample power system.