Juan Manuel Escaño

Development and Experimental Validation of a Dynamic Model for a Fresnel Solar Collector

Abstract this paper presents a lumped parameter dynamic model of a Fresnel collector field of a solar refrigeration plant. The plant is located in the Escuela Superior de Ingenieros of the University of Seville. The dynamic model parameter model developed can be used as a control model or as a simulation tool to test controllers. The lumped parameters have been determined by using real data from the plant in different operating conditions. The model has been validated against a data validation set obtained from the plant. The model has shown to reproduce the system behavior with a good compromise in accuracy and model complexity.

Ambulatory wireless sensor network power management using constrained explicit generalised predictive control

This work considers the problem of controlling transmit power within a wireless sensor network (WSN), where the practical constraints typically posed by an ambulatory healthcare setting are explicitly taken into account, as a constrained received signal strength indicator (RSSI) tracking control problem. The problem is formulated using an explicit generalised predictive control (GPC) strategy for dynamic transmission power control that ensures a balance between energy consumption and quality of service (QoS) through the creation of a stable floor on information throughput. Optimal power assignment is achieved by an explicit solution of the constrained GPC problem that is computed off-line using a multi-parametric quadratic program (mpQP). The solution is shown to be a piecewise-affine function. The new design is demonstrated to be practically feasible via a resource-constrained, fully IEEE 802.15.4 compliant, Moteivs Tmote Sky sensor node platform. Design utility is benchmarked experimentally using a representative selection of scaled ambulatory scenarios.

Explicitly constrained generalised predictive control strategies for power management in ambulatory wireless sensor network systems

This paper develops an explicit generalised predictive control (GPC) strategy for a wireless sensor network (WSN) power control problem that addresses practical constraints typically posed by an ambulatory healthcare problem scenario. An explicit solution of the particular GPC problem that arises facilitates the implementation of traditional on-line optimisation control strategies on a commercial sensor node platform with limited computational performance. The problem is shown to reduce to the evaluation of a piecewise linear function that can offer good performance with existing Mote type devices. The control law is validated experimentally against a number of existing strategies using a scaled, fully IEEE 802.15.4 compliant, testbed that emulates a selection of realistic wireless healthcare scenarios.

Fault tolerant MPC of a solar trough field based on classification and regression trees

Direct solar radiation is important in a solar trough plant because it is the perturbation that affects most the operation. Generally, it is measured locally. Production losses or dangerous situations may occur when controlling a plant with wrong measurements. A possible solution is to have several pyrheliometers though two problems arise: cost and sensor fusion (complexity). The aim of this paper is to prove that a solar estimation based on Classification and Regression Trees can be used to design a Fault Tolerant Model Predictive Control strategy capable to work with erroneous values of radiation or even none, avoiding dangerous situations or production losses.

Complexity Reduction in Fuzzy Systems Using Functional Principal Component Analysis

The ability to build fuzzy logic applications for control problems has been hindered by well-known problem of combinatorial rules explosion, causing complexity in modeling

Min-max model predictive control with robust zonotope-based observer

This paper considers the problem of robust estimation and constrained model predictive control (MPC). The paper deals with a discrete linear time-invariant system affected by additive bounded disturbances, whose states are measurable, but not directly accessible. In order to improve the control performance, a state estimator is desirable. The design problem of an observer based on zonotopes to estimate the system states of the uncertain system is addressed. Then, the min-max MPC optimisation problem formulation based on the designed robust observer as a quadratic program (QP) is described. An efficient implementation of the proposed robust observer-based control algorithm that can be solved by a standard QP is validated by simulation through the regulator problem of a cart pendulum system.

Game theory-based distributed predictive control: Application to shell oil fractionator

In this paper we propose to study the underlying properties of our formerly proposed distributed multi-rate predictive control strategy based on Nash game [1]. In this proposed method a set of multi-rate constrained agents communicate to accomplish their goals. The problems of how to decide the multirate communication strategy, share the inputs, estimated states and observers gains are solved using tools from game theory. The proposed scheme is demonstrated through a simulation example. Eventually different multi-rate scenarios have been simulated to present the performance of the method under all possible scenarios and also comparison of these scenarios together.

Solar radiation estimator and fault tolerant model predictive control of a parabolic-trough field

In this paper we propose an estimator of solar radiation and a heuristic radiation selector to design a solar radiation Fault Tolerant Model Predictive Control strategy for solar trough plants. The complete design of the control system proposed has been simulated using the model of ACUREX solar trough plant at the Plataforma Solar de Almería (PSA) using a Gain Scheduling Generalized Predictive Control(GS-GPC) control strategy. Simulation results (estimation/heuristic selector/GS-GPC) of the proposed scheme are compared with simulations of the same GS-GPC controller strategy when using only a pyrheliometer radiation measurement.

PCA based pressure control of a gas mixing chamber

PCA is a popular technique used in model reduction and fault diagnosis and isolation. In this work PCA is used to reduce the dimensionality of a MISO system. The coupling among the variables and the process output is taken into account through the projection into the PCA axis. The technique is applied to a gas mixing chamber in a Copper smelter factory, whose nonlinear behavior and large number of variables involved justify this approach. The control strategy is defined therefore in a straight and simple way making use of this new virtual and reduced system. The controller is simulated using a neurofuzzy model of the process that has been obtained using real data form the plant.

Dynamical & non-dynamical neurofuzzy models of a mixing chamber pressure

Two Neurofuzzy model of a mixing chamber pressure have been proposed. The process is a part of a copper smelter plant located in Huelva (Spain). The principal component analysis method has been used to reduce the inputs space, In addition, a non-recurrent model has been obtained, using hierarchical neurofuzzy models. The models have been validated with real data from the factory.