Dhafar Al-Ani

A new Sliding Mode Controller for Electro-Hydraulic Actuator (EHA) applications

In any type of system, unexpected fault conditions may occur. A minor fault may cause the system efficiency and performance to deviate from norm and/or deteriorate; a major fault condition may cause a system failure. Fault detection and diagnosis (FDD) and fault tolerance are two important concepts that need to be considered in industrial applications and more specifically in safety critical systems. This paper considers the development of a fault tolerant sliding mode control (SMC) strategy for a form of hydraulic actuation system known as the Electro-Hydrostatic Actuator (EHA). EHA systems are being increasingly used in flight control applications. An EHA prototype is used for demonstrating the effectiveness of the proposed fault tolerant control strategy.

A New Particle Swarm Optimization and Differential Evolution Technique for Constrained Optimization Problems

Real-world problems are often complex and may need to deal with constrained optimization problems (COPs). This has led to a growing interest in optimization techniques that involve more than one objective function to be simultaneously optimized. Accordingly, at the end of the multi-objective optimization process, there will be more than one solution to be considered. This enables a trade-off of high-quality solutions and provides options to the decision-maker to choose a solution based on qualitative preferences. Particle Swarm Optimization (PSO) algorithms are increasingly being used to solve NP-hard and constrained optimization problems that involve multi-objective mathematical representations by finding accurate and robust solutions. PSOs are currently used in many real-world applications, including (but not limited to) medical diagnosis, image processing, speech recognition, chemical reactor, weather forecasting, system identification, reactive power control, stock exchange market, and economic power generation. In this paper, a new version of Multi-objective PSO and Differential Evolution (MOPSO-DE) is proposed to solve constrained optimization problems (COPs). As presented in this paper, the proposed MOPSO-DE scheme incorporates a new leader(s) updating mechanism that is invoked when the system is under the risk of converging to premature solutions, parallel islands mechanism, adaptive mutation, and then integrated to the DE in order to update the particles’ best position in the search-space. A series of experiments are conducted using 12 well-known benchmark test problems collected from the 2006 IEEE Congress on Evolutionary Computation (CEC2006) to verify the feasibility, performance, and effectiveness of the proposed MOPSO-DE algorithm. The simulation results show the proposed MOPSO-DE is highly competitive and is able to obtain the optimal solutions for the all test problems.Copyright

Optimal pump operation for water distribution systems using a new multi-agent Particle Swarm Optimization technique with EPANET

The optimal pump scheduling allows for computing the most economical energy costs and provides more efficient operations for complex water distribution systems (WDS) with multiple pumping stations. The proposed technique employs the latest advances in multi-agent Particle Swarm Optimization (MOPSO) to automatically determine the most cost-effective solutions for scheduling/operation multiple pumps in multiple pumping stations, while satisfying both loading conditions and hydraulic performance requirements. The present work considers a bi-objective pump-scheduling problem, where the objectives are: minimize the electrical energy cost (

State estimation of a faulty actuator using the second-order smooth variable structure filter (The 2 ND -order SVSF)

/KW.h) and minimize the maintenance costs in terms of the total number of pump switches. In additional to the bi-objective pump-operational problem, pressure and tank levels (i.e., initial, minimum, and maximum) are considered as constraints in this paper for computing the most cost-effective solutions. The constraint-handling method, the Modified MOPSO (M-MOPSO) algorithm, and the modified EPANET Toolkit 2.0 are used to solve the constrained multi-objective problem. The results showed that the new MOPSO algorithm produced the most economical pump scheduling solutions.

Robustness comparison of some state estimation methods with an explicit consideration of modeling uncertainties

This paper presents the application of the new developed second-order Smooth Variable Structure Filter, 2nd-order SVSF, for fault detection under uncertain conditions. The 2nd-order SVSF is a novel modelbased state estimation method formulated in a predictor-corrector form. It produces robust state estimation under uncertain conditions, while decreasing the measurement error and its difference, at the same time. The stability of the 2nd-order SVSF is proven using the Lyapunovs stability criterion with the given bounded noise and modeling uncertainties. The corrective gain of the filter is pushing the measurement error and its difference toward zero. This results in providing higher degrees of accuracy, robustness and smoothness in state estimation under uncertain situations. Due to the robust performance of the 2nd-order SVSF for state estimation, it applies to an experimental setup of an Electro-Hydrostatic Actuator (EHA) for fault detection. Experimentation are performed for the EHA setup under the normal and faulty conditions. For the faulty EHA setup, there exists a major friction condition in the EHAs cylinders. The robustness of the new developed 2nd-order SVSF is then verified by comparing its performance with the state-of-the-art filters including the Kalman Filter (KF) and the 1st-order SVSF.

Cutting Tools Offset in Lathe Machines Using a Modern Heuristic Technique

This paper presents a comparative analysis of well-known state estimation methods that are commonly used in real systems. The aim of this research is to measure and then evaluate the robustness (i.e., a measure of performance when a small and deliberate changes are made to the method conditions) of these methods against modeling uncertainties. The state estimation methods include the Kalman Filter, the 1st-Order Smooth Variable Structure Filter (1st-order SVSF), and the new developed Dynamic 2nd-Order SVSF. A relatively new performance robustness criterion (so-called robustness index) is adopted in this work to first measure the robustness of the estimation methods, and then, evaluate their performance. The robustness index is calculated for each method when modeling uncertainties are explicitly considered. Simulation analysis is performed using the linear model of an Electro-Hydrostatic Actuator (EHA) setup under the normal and uncertain conditions. Simulation results showed the superior performance of the Dynamic 2nd-Order SVSF over other methods in terms of robustness against modeling uncertainties.

Fault Prognosis of Roller Bearings Using the Adaptive Auto-Step Reinforcement Learning Technique

In this paper, an optimal control strategy (i.e., offset settings system) based on the new Multi-objective Particle Swarm Optimization with Differential Evolution (MOPSO-DE) technique is developed and presented. The MOPSO-DE algorithm is used for calculating the optimal positions (i.e., offset settings) for the cutting tools in lathe machines. This optimal control strategy yields interesting results without a need to go through the complex mathematical modeling of the lathe system. The proposed technique is validated considering a real-world industrial system. This strategy is designed to take an action every 20 pieces, and it takes only 2.5 sec to run the code and optimally calculate the new settings. The control strategy is implemented using two high precision linear stepper motors. By implementing the new optimal control strategy, the estimated number of the defective pieces per day can be reduced by 85%.Copyright

A Robust Controller of Multi DOF-Cooperating Planar Robotic Manipulators Using a Tuned PID Approach

This paper presents the implementation of a new adaptation algorithm to model the crack propagation of roller bearings and to predict their Remaining Useful Life (RUL). The developed algorithm is designed based on the adaptive auto-step reinforcement-learning method combined with a crack propagation model. The advantage of this algorithm is that it is now able to not only estimate the defect growth rate online, but also to predict the RUL of a roller bearing element. The presented defect propagation model incorporated in this work is an extension to the Paris’s formula that is well known in the fracture mechanics community. Further, a new adaptive filtering technique, referred to as the auto-step, is presented in this paper and is used to estimate the parameters of the crack propagation model in real-time. The prognosis structure is first compares values of both the predicted and the measured defect sizes, and then, tunes the parameters of the crack propagation model. Simulation results obtained by the auto-step method are then compared with results obtained by the Recursive Least Square (RLS) adaptive filter. The proposed prognosis strategy is distinct itself from other approaches in terms of obtaining higher accuracy as well as faster convergence rate.Copyright

Operational Costs Optimization in Water Distribution Systems

Usually, a dynamic system with impact conditions is an interesting problem with practical applications in the fields of dynamics, vibrations, and control. One difficulty in controlling robotics (i.e., a multi DOF two-cooperating or two-link planar) is the subject to impact between the end-effectors of manipulators is that the dynamics (i.e., equations of motion) are different when the system status changes suddenly from a non-contact state to a contact state. In this paper, a Tuned PID controller with different design scenarios is developed to regulate the states of two dynamic systems that collide. Further, in this work, three types of errors are used to compare among different cases that are; (1) the steady state error, (2) the root mean square error, and (3) the final value error. The results of the Tuned PID controller are compared to those obtained by a classical PID controller. The PID controller is tuned using the Ziegler–Nicholas approach. The simulation results of the robotic manipulators confirmed the theoretical effectiveness of the proposed controller, based on MATLAB/Simulink. Unlike the classical PID results (i.e., the impact-induced force is found to be 2.0 N), the Tuned PID controller successfully determined the impact-induced force as same as the desired force (i.e., 0.6 N). Moreover, the Tuned PID satisfied all other desired design values.Copyright

A new adaptive control scheme based on the interacting multiple model (IMM) estimation

Pump management and reservoir management have many similarities, and therefore, should ideally be analyzed in an integrated way to plan effectively the daily operation of water distribution systems. Historically, these two management activities have been evolved as separate tasks in energy-efficiency (i.e., energy optimization) studies and are often carried out in an isolated way. The latter being most often associated directly with the concepts of multimodal and multi-objective optimization problems, whereas the former is usually considered as a single optimization problem to be solved. When some single optimization problems appear at part of the solution tied to a local (i.e., regional) search-space (i.e., objective space), this artificial integration (i.e., multi-modal and multi-objective optimization) can always obtain optimal solutions. Similarly when system constraints and load conditions are considered, a set of feasible and innovative optimal solutions can be obtained in order to continue the enhancement of energy consumption that turns into a significant reduction in the overall operational cost (i.e., a potential of 6.24% cost savings) without affecting the level of services provided to the clients in a safe and protected manner.Copyright