A novel self-tuning control method based on regulated bi-objective emotional learning controller's structure with TLBO algorithm to control DVR compensator

doi:10.1016/j.asoc.2014.08.051

Applied Soft Computing

Volume 24, November 2014, Pages 912-922

https://doi.org/10.1016/j.asoc.2014.08.051 Get rights and content

Highlights

•
Presenting self-tuning PI controller based on human brain emotional learning to control DVR.
•
Suggesting a bi-objective structure for emotional controller to satisfy voltage sag and THD.
•
Considering sensitivity of the controller's coefficients, they are regulated by an algorithm.
•
TLBO algorithm is used to optimize emotional controller's parameters.
•
Convergence speed and final answer are great better in TLBO algorithm than PSO algorithm.

Abstract

DVR is one of the custom power devices for compensating power quality indices. A self-tuning controller with a bi-objective structure is presented for controlling the DVR compensator in order to improve the THD and voltage sag indices of a sensitive load in the network. In this paper, the emotional controller which is based on emotional learning of human brain is proposed for controlling the DVR compensator. This controller has such a structure that makes it capable of considering a second objective in the control process of the system. So far, this capability of the emotional controller has not been used in any researches. The results of the paper demonstrate that compensating and controlling the voltage THD signal in the control process has caused more improvement in the voltage sag of the sensitive load. It was reported that the performance of the emotional controller depends on the selection of the values of its coefficients. Therefore, in order to better improve the proposed controller, these coefficients are tuned by an optimization algorithm. Teaching–learning-based optimization algorithm is considered as optimization algorithm to regulate these coefficients. According to simulation results, it works significantly better than classic PI controller and some intelligent controllers that have introduced in other researches already.

Introduction

Nowadays by increasing the number of sensitive loads, demand for access to stable and high quality electrical power has increased significantly. In industrial competitive environment, with the development of commercial production of power electronic devices, computer processors and nonlinear loads, any interruption or diversion from the standard range causes economic losses. The realization of this economic loss can be studied in such frameworks as production competition opportunity loss, efficiency reduction and production cost increase, low-quality products, reduced equipment lifetime and increased repair cost, production interruption and energy losses. Thus, access to high power quality, applies a great influence on the asset savings and economic advantages for a firm [1].

Disturbance in power distribution system causes harmful defects in distribution system such as interruption, voltage sag, voltage swell, and flicker. Among the above disturbances the most important is voltage sag. According to the IEEE standard, it is defined as a sudden voltage decrease in the range of 10–90% for 0.5 cycles to 1 min [2]. That is the result of natural phenomena such as system asymmetric errors and electromagnetic phenomena such as start and inrush current.

“Custom power” device have been introduced by experts in order to compensate the harmful effects of disturbances on sensitive loads. Among these devices, DVR is capable to compensate voltage sag and swell effects for sensitive loads devices. The structure of DVR in simple terms consists of: electrical storage source, voltage source inverter and coupling transformer. Recognizing voltage sag in feeder connected to the sensitive load, DVR generates proper voltage using coupling transformer which is in series with sensitive load and injects proper voltage to the network and decrease voltage sag effect.

The control system of a DVR plays an important role. It should have a fast response in presence of voltage sags and variations in the connected load. The main purpose of the control system is to preparation of a constant voltage for sensitive load, under system disturbances [3]. The performance analysis and control of the DVR, with different control strategies, have been studied and examined by researchers. Most of the published works on the DVR have used an ordinary proportional-integral (PI) controller in a synchronous frame [4]. The classical PID method has poor flexibility since its parameters cannot be changed. Furthermore, when it is applied to such complicated systems as power system in fault conditions specially, proper results cannot be obtained in most cases [5]. Therefore, control strategies such as predictive control [6], sliding mode [7] and robust control [8] are used in order to control injected voltage. Also in [9] and [10] H∞ controller and a controller based on iteration are used respectively for having better operation in steady and transient states. A multi-level inverter with optimal predictive control structure is used in [11]. These controllers are based on classical and nonlinearity control theory. The problems of control plants that arise in systems such as power systems and its components such as DVR can be classified under three categories. The first problem is complex computation of DVR control. Therefore these theories are rarely used in practical systems. The second is the presence of nonlinearities in these systems that make the control problem complicated. The last is uncertainty in these systems. Necessary information required in the mathematical model of these systems such as intensive dynamic behaviors in normal and fault mode. Most classic and nonlinear control methods are model base. Hence it is possible that they have complexity in control. Therefore, based on the nonlinear control theory as fine as the human ability to comprehend, reason and learn, intelligent methods may be used to overcome the these problems [12]. Thus some researchers investigated on intelligent techniques in order to control DVR compensator. Emotional controller is implemented as an adaptive controller in [13] in DVR control. In [14], [15], [16], improving voltage THD index has also been considered as an objective and a control criterion. In [14], [15] is introduced a multi-objective PSO and multi-objective modified PSO (CAPSO) algorithms to optimize voltage sag and THD. An adaptive neural network controller based on Hebb learning theory discussed in [16]. This controller was made bi-objective with fuzzification of goals. However, in all of the aforementioned references, algorithms are complex. Although applied control strategies are capable to reduce impulses caused by voltage sag in sensitive loads, but most of these approaches do not consider reducing voltage THD. In many sensitive loads such as medical equipment and adjustable speed motor drives, this level of sensitivity can be very important.

In most of the aforementioned researches, it is tried to utilize a stable controller to have a robust system in the presence of fault conditions. Refs. [14], [15] are one the rare researches which has tried to improve sensitive load voltage THD and voltage sag together. In this research evolutionary algorithms have been used to optimize two-objective mentioned together in DVR. This approach has two shortcomings. Firstly, power systems and compensators have a completely nonlinear nature. This may cause that these algorithms operate based on random search to encounter the problems and converge to a local optimum. Also, it is necessary to say that in real power systems this search process takes long time. Secondly, power systems have different dynamic behaviors especially during fault conditions. Therefore, PID controllers which are optimized by off-line search algorithms may not have a good performance under these conditions. In this paper it is tried to solve these problems. Consequently, a controller based on intelligent systems with two-objective structure should be better. We try to create a controller that has more accuracy in DVR control. This controller is based on emotions called emotional controller and it is adaptive based on intelligent control. In the following, we introduce this controller.

The adaptive method which is inspired by learning nature of human brain is used as a self-tuning PI controller. In order to put capability of having an appropriate performance during voltage sag and voltage THD for sensitive load, a two-objective structure is proposed. Proper regulating of emotional controller's parameters is enormous essential to have better performance [17]. Hence we decide to regulate these parameters with an optimization algorithm. Considering some drawbacks of most optimization algorithms, in this paper we use TLBO algorithm for regulating these parameters. Most optimization methods require algorithm parameters that affect the performance of the algorithm. For instance, GA requires the crossover probability, mutation rate, and selection method; PSO requires learning factors, the variation of weight, and the maximum value of velocity. The proper tuning of the algorithm specific parameters is very crucial factor, which affect the performance of the above mentioned algorithms. The improper tuning of algorithm-specific parameters either increases the computational effort or yields the local optimal solution [18], [19], [20]. Considering this fact, Rao et al. [21] introduced the teaching–learning-based optimization (TLBO) algorithm which does not require any algorithm parameters to be tuned, thus making the implementation of TLBO simpler [21]. Therefore in this paper, TLBO algorithm is used for regulating bi-objective emotional controller's coefficients to have appropriate performance. Also we use PSO algorithm for regulate parameters of this controller to compare this PSO algorithm with TLBO algorithm. To investigate efficiency of the proposed algorithm, performance of DVR compensator is tested during various faults in a typical network and compared with those of emotional and classic PI controller and other controllers which are introduced in previous researches. Also, in this paper in order to have better comparison, some other controllers based on intelligent systems have been used. These controllers are such as PI controller regulated with multi-objective PSO and CAPSO algorithm [14], [15], Hebb learning controller and its bi-objective version [16].

DVR operation is introduced in Section 2; PI emotional controller and its multi-objective structure (proposed controller) are introduced in Sections 3 Structure of human brain emotional learning, 4 DVR intelligent control based on emotional learning and its bi-objective control structure respectively. The final section contains the simulation and results.

Section snippets

DVR's structure and functionality

DVR is one of the “custom power” devices in distribution network which is series with transmission line. Load voltage becomes balance by injecting three controlled voltages during disturbance in the power system. Thus DVR is based on injection of appropriate voltage during voltage sag in order to compensate it. DVR functionality can be categorized as two modes: standby mode and injection mode [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]. In the standby mode a low voltage inject

Structure of human brain emotional learning

In 2000, this controller was introduced for the first time by Moren and Balkenious. These researchers started to develop computational models of those parts of the human brain which carry out emotional processing. In [23], a new model for performance of the brain emotion processing parts consisting of Amygdala, orbitofrontal cortex, thalamus, and finally sensory cortex has been presented in Fig. 2, Fig. 3.

Considering the aforementioned model and according to new theories, the

DVR intelligent control based on emotional learning and its bi-objective control structure

By combining Eqs. (1), (2), (3): $MO = (G_{a} - G_{oc}) \cdot SI \equiv G (SI, EC, \dots) \cdot SI$

In the other words, the Amygdala–orbitofrontal emotional learning system output is equal to a varying gain G (which is dependent to various factors such as emotional signal EC, input stimulant SI, etc.,) multiplied by input stimulant SI [24]. Due to the value and high performance of self-tuning PID controller in DVR performance control domain, a simple and appropriate suggestion to formulate stimulant input signal is a template similar

Teaching–learning based optimization algorithm

TLBO algorithm, originally developed by Rao et al. [21], is a population-based optimization algorithm. In TLBO algorithm a population of solutions is utilized to proceed to the global solution. To this end, in TLBO algorithm, a group of learners is chosen as population and different design variables are considered as different subjects offered to the learners and learners’ result is similar to the ‘fitness’ value of the optimization problem [25], [26], [27]. In the whole population, the best

Proposed method

A special characteristic of emotional controller that makes it effective is its flexibility. This controller has several gains that give good freedom for choosing favorite response [28]. Here we use TLBO algorithm to choose the best set of these gains for our controller to achieve the desired response. Emotional controller is used to improve the step response of DVR compensator. The TLBO algorithm was utilized to determine six optimal controller parameters, such that the controlled system could

Simulation and results

The power distribution system case study consists of two voltage buses one of which includes sensitive load. The simple schematic of electrical network is shown in Fig. 7 and its parameters are introduced in Table 1. This network is applied in [14], [15], [16].

The distractive effect of fault increases by decreasing distance between event locations to sensitive load. In order to analyze more critical conditions, two faults are simulated. The first fault is occurred just after series injection

Conclusion

According to the results obtained from applying proposed algorithm to the test system, it can be said that this method is a appropriate approach for improving power quality for customers. As it is clear, due to dynamic behavior of power system under normal and fault conditions it is difficult to have a good model. This method is applied as bi-objective controller to modify both of indices of power qualities such as sag and THD voltage. We should have accurate and efficient controller to

References (29)

R.V. Rao et al.
Teaching–learning-based optimization: a novel method for constrained mechanical design optimization problems
Comput. Aided Des.
(2011)
M. Abirami et al.
Source and transmission line maintenance outage scheduling in a power system using teaching learning based optimization algorithm
Appl. Soft Comput.
(2014)
B. Mandal et al.
Multi-objective optimal power flow using quasi-oppositional teaching learning based optimization
Appl. Soft Comput.
(2014)
J.V. Milanovic et al.
Global minimization of financial losses due to voltage sags with FACTS based devices
IEEE Trans. Power Deliv.
(2010)
IEEE std. 1159–1995, IEEE Recommended Practice for Monitoring Electric Power...
S.A. Mohammed et al.
Conventional dynamic voltage restorer (DVR) for mitigation of voltage sag in power distribution systems
Int. J. Adv. Eng. Technol.
(2013)
M. Karimian et al.
Proportional repetitive control of a dynamic voltage restorer (DVR) for power quality improvement
Int. J. Tech. Phys. Probl. Eng.
(2012)
X. Liu
Single Neuron Self-tuning PID Control for Welding Molten Pool Depth
H. Zhou et al.
Study on Dynamic matrix predictive control of single-phase DVR, experimental thermal and fluid science
G.V. Nagesh Kumar et al.
DVR with sliding mode control to alleviate voltage sags on a distribution system for three phase short circuit fault

Y.W. Li et al.

A robust control scheme for medium-voltage-level DVR implementation

IEEE Trans. Ind. Electron.

(2007)

Y. Wei Li et al.

Design and comparison of high performance stationary-frame controllers for DVR implementation

IEEE Trans. Power Electron.

(2007)

P. Roncero-Sánche et al.

A versatile control scheme for a dynamic voltage restorer for power-quality improvement

IEEE Trans. Power Deliv.

(2009)

J. Dionísio Barros et al.

Multilevel optimal predictive dynamic voltage restorer

IEEE Trans. Ind. Electron.

(2010)

Cited by (41)

Performance evaluation of dynamic voltage restorer using intelligent control algorithm
2023, e-Prime - Advances in Electrical Engineering, Electronics and Energy
Power quality has become an important issue due to increasing applications of modern industrial and commercial sensitive equipment's. Due to high penetration of modern power electronics-based devices leads to voltage disturbances. To address the voltage disturbances the commonly employed series device called Dynamic Voltage Restorer (DVR) is used which requires an effective control technique. A control scheme namely Volterra adaptive expanded filter is developed which incorporate the higher ordered cost function. The conventional filters have like least mean square (LMS) algorithm is restricted to power error of second order. This restricts the computation of fundamental quantity from the disrupted signals and results in degradation under dynamical scenarios. The recommended approach surmount this aforesaid issue with a solution of a mixed learning of step size. This is derived from the error of second and fourth-order. The Volterra based LMS/Fourth (LMS/F) control has the merit of estimating the fundamental quantity under any power quality (PQ) scenario with less delay and low computational burden. The conventional controller like Proportional Integral (PI) has limited gain coefficients to achieve the optimized the voltage regulation. Furthermore, this is overcome by a Fractional Order PID (FOPID) with an additional coefficient of λ and γ obtain more stabilized DC and AC bus voltage by integrating with Kepler optimizer. For achieving optimal gain, an online tuning of FOPID is done with Kepler Optimizer. The Kepler is inspired by planet motion to achieve best gain values. The evaluation of FOPID is made by considering as a Fitness Function (FF) an Integral Time Square Error (ITSE). Owing to the superior performance of Kepler optimizer based FOPID under transitory state and an enhanced DVR performance is observed. The optimized FOPID has improved response in terms of less overshoot (2.9 %), undershoot (5.6 %), and fast settle time (0.104 s). To authenticate the simulating and experimental results dSPACE board is employed to validate the efficacy and accuracy of the proposed Volterra LMS/F technique.
Emotional neural networks with universal approximation property for stable direct adaptive nonlinear control systems
2020, Engineering Applications of Artificial Intelligence
Universal approximation, continuity, and differentiability are desirable properties of any computational framework, including those that rise from human cognition and/or are inspired by nature. Emotional machines constitute one such framework, but few studies have addressed their mathematical properties. Here, we propose a Continuous Radial Basis Emotional Neural Network (CRBENN) that benefits from the universal approximation property, continuous output, and simple structure of RBF; while keeping the fast response properties of emotion-based approaches. As such, CRBENN is amenable to a wide array of challenging problems in systems engineering and artificial intelligence. Here, we propose a CRBENN-based direct adaptive robust emotional neuro-control approach (DARENC) for a class of uncertain nonlinear systems. Stability is theoretically established using Lyapunov analysis of the closed-loop system. DARENC is then applied to control two nonlinear systems, and the performance of the controller is numerically compared with several competing fuzzy, neural, and emotional controllers. The simulation results indicate improved tracking performance, better disturbance rejection, and less control effort. Finally, DARENC is implemented on a real-world 3-PSP (spherical–prismatic–spherical) parallel robot in our laboratory. The experimental results show the satisfactory performance of the robot in tracking the desired trajectory with low control effort.
A hybridized ELM-Jaya forecasting model for currency exchange prediction
2020, Journal of King Saud University - Computer and Information Sciences
Citation Excerpt :
The algorithm needs only common controlling parameters and does not require any algorithm specific control parameters. Keeping in view of Teaching Learning Based Optimization (TLBO) algorithm, which also does not require (Khalghani and Khooban, 2014) any algorithm specific parameters to be tuned, Jaya is introduced, but again Jaya differs from TLBO, as unlike two phases that is teaching phase and learner phase of TLBO, Jaya is having only one phase and it is simpler to apply. In this study an exchange rate prediction based on ELM, Functional Link Artificial Neural Network (FLANN) and (Neural Network) NN with a set of technical indicators and statistical measures have been experimented.
This paper establishes a hybridized intelligent machine learning based currency exchange forecasting model using Extreme Learning Machines (ELMs) and the Jaya optimization technique. This model can very well forecast the exchange price of USD (US Dollar) to INR (Indian Rupee) and USD to EURO based on statistical measures, technical indicators and combination of both measures over a time frame varying from 1 day to 1 month ahead. The proposed ELM-Jaya model has been compared with existing optimized Neural Network and Functional Link Artificial Neural Network based predictive models. Finally, the model has been validated using various performance measures such as; MAPE, Theil's U, ARV and MAE. The comparison of different features demonstrates that the technical indicators outperform both the statistical measures and a combination of statistical measures and technical indicators in ELM-Jaya forecasting model.
A hybrid elicit teaching learning based optimization with fuzzy c-means (ETLBO-FCM) algorithm for data clustering
2018, Ain Shams Engineering Journal
Citation Excerpt :
Unlike other optimization techniques, it does not require any controlling parameter setting which makes the algorithm simple and effective to use in real life domains. Since its inception, it has been a key interest among a wide range of researchers and has been successfully implemented in various application areas including Mechanical Design Optimization [28], Optimal Power flow [29], Multi-objective optimization [30], Hydro Thermal Scheduling [31], Power System [32,33], Machining processes [34], Scheduling of Thermal Power Systems [35], Power Quality [36] and global optimization [37,38]. Due to its easy implementation and less design issues, it attracts researchers to use it in various application areas.
Since its inception, Fuzzy c-means (FCM) technique has been widely used in data clustering. The advantages of FCM such as balancing of individual number of cluster points, drifting of small cluster centers to large neighboring cluster centers, and presence of fuzzy factor, make it more popular. However, early trapping at local minima and high sensitivity to the cluster center initialization are the major limitations of FCM. In this paper, a novel Elicit Teaching learning based optimization (ETLBO) approach has been incorporated with the Fuzzy c-means clustering algorithm to obtain the improved fitness values of the cluster centers. The simulation results of the proposed method have been compared with some other existing methods such as GA, PSO and IPSO. Experimental results show that the proposed approach is superior to the other methods in terms of their fitness value calculations.
Chaos synchronization using higher-order adaptive PID controller
2018, AEU - International Journal of Electronics and Communications
Citation Excerpt :
The process of TLBO can be divided into two parts: namely, Teacher Phase and Learner Phase [26]. The pseudo-code of TLBO algorithm has been given in [30]. Now, to discuss the robust performance of the proposed controller, the effect of the external disturbance is investigated.
This paper is devoted to designing higher-order adaptive PID controllers as a new generation of PID controllers for chaos synchronization, in which second order integration and second-order derivative terms to the PID controller (PII²DD²) are employed. The five PII²DD² control gains are updated online with a stable adaptation law driven by Lyapunov’s stability theory. This is the unique advantage of the proposed approach. Furthermore, it is equipped with a novel robust control term to improve controller robustness against system uncertainties and unknown disturbances. An important feature of the proposed approach is that it is a model-free controller. In addition, to determine the control design parameters and avoid trial and error, the Teaching–learning-based optimization algorithm (TLBO) is employed to regulate these parameters and enhance the performance of the proposed controller. Based on the Lyapunov stability theory, it is proven that the proposed control scheme can guarantee the synchronization and the stability of closed-loop control system. The case study is the Duffing–Holmes oscillator. Comparative simulation results are presented which confirm the superiority of the proposed approach.
A TLBO based gradient descent learning-functional link higher order ANN: An efficient model for learning from non-linear data
2018, Journal of King Saud University - Computer and Information Sciences
All the higher order ANNs (HONNs) including functional link ANN (FLANN) are sensitive to random initialization of weight and rely on the learning algorithms adopted. Although a selection of efficient learning algorithms for HONNs helps to improve the performance, on the other hand, initialization of weights with optimized weights rather than random weights also play important roles on its efficiency. In this paper, the problem solving approach of the teaching learning based optimization (TLBO) along with learning ability of the gradient descent learning (GDL) is used to obtain the optimal set of weight of FLANN learning model. TLBO does not require any specific parameters rather it requires only some of the common independent parameters like number of populations, number of iterations and stopping criteria, thereby eliminating the intricacy in selection of algorithmic parameters for adjusting the set of weights of FLANN model. The proposed TLBO-FLANN is implemented in MATLAB and compared with GA-FLANN, PSO-FLANN and HS-FLANN. The TLBO-FLANN is tested on various 5-fold cross validated benchmark data sets from UCI machine learning repository and analyzed under the null-hypothesis by using Friedman test, Holm’s procedure and post hoc ANOVA statistical analysis (Tukey test & Dunnett test).

View all citing articles on Scopus

View full text

A novel self-tuning control method based on regulated bi-objective emotional learning controller's structure with TLBO algorithm to control DVR compensator

Highlights

Abstract

Introduction

Section snippets

DVR's structure and functionality

Structure of human brain emotional learning

DVR intelligent control based on emotional learning and its bi-objective control structure

Teaching–learning based optimization algorithm

Proposed method

Simulation and results

Conclusion

Comput. Aided Des.

Appl. Soft Comput.

Appl. Soft Comput.

Global minimization of financial losses due to voltage sags with FACTS based devices

IEEE Trans. Power Deliv.

Conventional dynamic voltage restorer (DVR) for mitigation of voltage sag in power distribution systems

Int. J. Adv. Eng. Technol.

Proportional repetitive control of a dynamic voltage restorer (DVR) for power quality improvement

Int. J. Tech. Phys. Probl. Eng.

Single Neuron Self-tuning PID Control for Welding Molten Pool Depth

Study on Dynamic matrix predictive control of single-phase DVR, experimental thermal and fluid science

DVR with sliding mode control to alleviate voltage sags on a distribution system for three phase short circuit fault

A robust control scheme for medium-voltage-level DVR implementation

IEEE Trans. Ind. Electron.

Design and comparison of high performance stationary-frame controllers for DVR implementation

IEEE Trans. Power Electron.

A versatile control scheme for a dynamic voltage restorer for power-quality improvement

IEEE Trans. Power Deliv.

Multilevel optimal predictive dynamic voltage restorer

IEEE Trans. Ind. Electron.