Adaptive Neuro-fuzzy System Research Articles

Prediction of L10 and Leq Noise Levels Due to Vehicular Traffic in Urban Area Using ANN and Adaptive Neuro-Fuzzy Interface System (ANFIS) Approach

Recently in urban areas, road traffic noise is one of the primary sources of noise pollution. Variation in noise level is impacted by the synthesis of traffic and the percentage of heavy vehicles. Presentation to high noise levels may cause serious impact on the health of an individual or community residing near the roadside. Thus, predicting the vehicular traffic noise level is important. The present study aims at the formulation of regression, an artificial neural network (ANN) and an adaptive neuro-fuzzy interface system (ANFIS) model using the data of observed noise levels, traffic volume, and average speed of vehicles for the prediction of L10 and Leq. Measured noise levels are compared to the noise levels predicted by the experimental model. It is observed that the ANFIS approach is more superior when compared to output given by regression and an ANN model. Also, there exists a positive correlation between measured and predicted noise levels. The proposed ANFIS model can be utilized as a tool for traffic direction and planning of new roads in zones of similar land use pattern.

Reduced Time Compression in Big Data Using MapReduce Approach and Hadoop.

An exponential rise has been observed in the data volume over the time when considering a real time environment. A phenomenal feature termed as 'Predictability' helps in predicting and portraying related data to the user according to their needs. Moreover, classification of Big Data is usually a tedious and lengthy task. The technique of MapReduce Framework performs the data processing that being paralleled by data distribution in small chunks through the clusters. This Map Reduce technique is being proposed which is employed to process heterogeneous data items. Few issues that are being targeted in the existing paper include associating climatologically and meteorological information with large variety of farming decisions. Using the well-known MapReduce framework the above issues and challenges can be resolved. The existing paper proposes empirical techniques of climate classification and prediction by adopting Co-EANFS (Co-Effective and Adaptive Neuro-Fuzzy System) approach for data handling. Furthermore, the paper examines association rule mining too, which is being implemented for examining the best crop production by relying upon the soil and weather condition. Lastly, a technique is proposed for managing various levels such as preprocessing, clustering, classification and prediction. First, the weather dataset is being collected which undergoes processing; thereafter the proposed model is implemented which results in formation of cluster data sets linked to each season. For evaluating the performance, accuracy predictions generated by Co-EANFS is used which being formulated with varying no: of inputs and variables. The proposed framework acquires least execution time.

Speed Control of a Separately Excited DC Motor Using New Proposed Fuzzy Neural Algorithm Based on FOPID Controller

The main goal of this paper is to control the speed of a separately excited DC motor (SEDM) with a new proposed fuzzy neural (FN) controller. This proposed method is used to adjust the fractional order proportional integral derivative (FOPID) parameters of the controller. Also the proposed control diagram solves the problem of parameter setting of the FN controller more effectively with use of particle swarm optimization (PSO) algorithm. In simulation with MATLAB 2017b, 250 series of data were used: 175 series of data, equivalent to 70% for training the designed neural network, and about 75 series, equivalent to 30% used to test and validate the neural network. The results show that the proposed method has a lower rise time and settling time for controlling the speed of SEDM in comparison with other methods such as Ziegler–Nichols, Cohen–Coon, PSO, genetic algorithm, artificial bee colony, artificial neural network, fuzzy logic controller and adaptive neuro-fuzzy interference system for PID and FOPID controllers.

Estimation of maximum scour depth downstream of an apron under submerged wall jets

AbstractAn analysis of laboratory experimental data pertaining to local scour downstream of a rigid apron developed under wall jets is presented. The existing equations for the prediction of the maximum scour depth under wall jets are applied to the available data to evaluate their performance and bring forth their limitations. A comparison of measured scour depth with that computed by the existing equations shows that most of the existing empirical equations perform poorly. Artificial neural network (ANN)- and adaptive neuro-fuzzy interference system (ANFIS)-based models are developed using the available data, which provide simple and accurate tools for the estimation of the maximum scour depth. The key parameters that affect the maximum scour depth are densimetric Froude number, apron length, tailwater level, and median sediment size. Results obtained from ANN and ANFIS models are compared with those of empirical and regression equations by means of statistical parameters. The performance of ANN (RMSE = 0.052) and ANFIS (RMSE = 0.066) models is more satisfactory than that of empirical and regression equations.

Compressive strength Prediction recycled aggregate incorporated concrete using Adaptive Neuro-Fuzzy System and Multiple Linear Regression

Compressive strength of concrete, renowned as one of the most substantial mechanical properties of concrete and key factors for the quality assurance of concrete. In the present study, two different data-driven models, i.e., Adaptive Neuro-Fuzzy Inference System (ANFIS), and Multiple Linear Regression (MLR) were used to predict the 28 days compressive strength of recycled aggregate concrete (RAC). 16 different input parameters, including both dimensional and non-dimensional parameters, were used for predicting the 28 days compressive strength of concrete. The present study established that estimation of 28 days compressive strength of recycled aggregate concrete was performed better by ANFIS in comparison to MLR. Besides, the performance of data-driven models with and without the non-dimensional parameters is explored. It was observed that the data-driven models show better accuracy when the non-dimensional parameters were used as additional input parameters. Furthermore, the effect of each non-dimensional parameter on the performance of each data-driven model is investigated and 28 days compressive strength of concrete is examined.

Distinguishing mental attention states of humans via an EEG-based passive BCI using machine learning methods

Recent advances in technology bring about novel operating environments where the role of human participants is reduced to passive observation. While opening new frontiers in productivity and lifestyle, such environments also create hazards related to the inability of human individuals to maintain focus and concentration during passive control tasks. A passive brain-computer interface for monitoring mental attention states of human individuals (focused, unfocused, and drowsy) by using electroencephalographic (EEG) brain activity imaging and machine learning data analysis methods is developed in this work. An EEG data processing pipeline and a machine learning mental state detection algorithm using the Support Vector Machine (SVM) method were designed and compared with k-Nearest Neighbor and Adaptive Neuro-Fuzzy System methods. To collect 25 h of EEG data from 5 participants, a classic EEG headset was modified. We found that the changes in EEG activity in frontal and parietal lobes occurring at 1–5 Hz and 10–15 Hz frequency bands were associated with the changes in individuals’ attention state. We demonstrated the ability to use such changes to identify individuals’ attention state with 96.70% (best) and 91.72% (avg.) accuracy in experimental settings using a version of continuous performance task with SVM-based mental state detector. The findings help guide the design of future systems for monitoring the state of human individuals by means of EEG brain activity data.

Performance evaluation of multilevel inverter based hybrid active filter using soft computing techniques

In this paper, a multilevel inverter based shunt hybrid active filter is used, which is composed of passive filter and shunt active filter, used for enhancing the quality of power, by reducing the total harmonic distortions below 5% as per IEEE-519 standard and managing the reactive power as well as correcting the power factor. The objective of this paper is to advance the power quality by reducing the harmonic distortions in the distribution line, which are affected because of utilising non-linear loads across the load end. For reference current generation and for controlling the switching signal, conventionally PI controller was used with different reference signal detection techniques. But with the evolution of different intelligence approaches, made more advantageous than the PI controller. In this paper, two artificial intelligence methods, i.e, artificial neural network and adaptive neuro-fuzzy interference system (ANFIS) are applied. Comparison of the results of the proposed methods is analysed using MATLAB/ SIMULINK tool. For reference current generation, synchronous reference frame theory is used and hysteresis current control technique is employed for generating the gating signal.

Asphaltene adsorption using green nanocomposites: Experimental study and adaptive neuro-fuzzy interference system modeling

Asphaltene deposition is a known problem that causes significant cost increases in the oil industry. Two bio-templated adsorbents, namely the NiO/ZSM-5 and NiO/AlPO-5 nanocomposites, were used as new green adsorbents to remove asphaltene from a model oil solution. Composite adsorbents were characterized by FTIR, BET, TEM and XRD analysis. Batch adsorption experiments were carried out as a function of D/C0 [(g)adsorbent/(mg/l)initial], pH, and temperature (K). Results showed that maximum adsorption is obtained at D/C0 = 0.072[g/(mg/l)] with a pH of 4.8 and a temperature of 298 K for NiO/ZSM-5 and D/C0 = 0.084[g/(mg/l)] with a pH of 3.4 and a temperature of 298 K for NiO/AlPO-5. In the experimental data, equilibrium adsorption models were introduced and their constants were calculated. The equilibrium adsorption data on NiO/ZSM-5 were well matched to the Freundlich model at 298 K and 325 K, and Temkin model at 342 K and 353 K. For the NiO/AlPO-5 adsorption data, the Temkin model was the best model showing strong adsorption interactions of asphaltene and adsorbent. The adaptive neuro-fuzzy interference system (ANFIS) was also used to model and predict the amount of asphaltene adsorbed by the proposed nanocomposites. ANFIS designed by triangular-shaped membership functions with three nodes and first-order polynomial Sugeno type FIS was the optimal structure and gave R2 = 0.9999 and R2 = 0.9996 for train and test data, respectively. Finally, Monte Carlo algorithm was used for sensitivity analysis on the input variables which is necessary for process optimization. Results demonstrated that D/C0, pH, and temperature have the highest effect on asphaltene removal by nanoparticles.

Rolling bearing fault detection of electric motor using time domain and frequency domain features extraction and ANFIS

Demands for various products, higher qualities, reduction of costs and competitiveness, have resulted in the use of intelligent fault detection systems. Bearing fault diagnosis as a major component of the electric motors has had an essential role in the operation of production units’ reliability. In addition, vibration analysis is one of the most powerful tools in diagnostics. Advances in signal processing technology and electrical equipment have developed a machinery condition monitoring for defect detection. This study has used the extracted features of vibration signals and the adaptive neuro-fuzzy interface system (ANFIS) network proposing a structure for fault detection and diagnosis of rolling bearings. Time-domain and frequency-domain statistical characteristics have been extracted fault information from vibration signals. Besides, the test data sets are presented to the ANFIS network. Simulation results indicated that the performance of the ANFIS network is acceptable. The results reveal that this method has more accuracy and better classification performance in comparison with other methods proposed in the literature.

Confined Aquifer’s Hydraulic Parameters Estimation by a Generalized Regression Neural Network

Precise estimation of aquifer hydraulic parameters can be an asset to more sustainable management of groundwater resources. The nonlinear correlation of aquifer parameters along with equation-level complicacy of aquifers’ charging functions can be efficiently handled using artificial intelligence models due to their flexibility in mapping between the observed data and output functions. This study focuses on performance comparisons for generalized regression neural network, artificial neural network (ANN) and adaptive neuro-fuzzy interference system (ANFIS) for estimation of hydraulic parameters, namely storage coefficient and transmissibility of confined aquifers. To acquire a greater precision of the estimated output functions, the principle component analysis of the pumping test data was initially undertaken to reduce the input dimensions by filtering out redundancies and insignificant variables’ correlations. Next, these data are passed into the training and validation process of the artificial intelligence models. Several error indices, mean absolute relative error (MARE), RMSE, MAE, RMRE, Bias and R2 were used during the validation of the prediction models. Finally, the estimated output functions were compared against the traditional and yet commonly used graphical Theis method. As an example, MARE in aquifer parameters estimation by ANN and graphical Theis method was 0.5564% and 1.1320%, respectively. More generally, among all the intelligence prediction models used to estimate the hydraulic parameters of a confined aquifer, ANFIS was more accurate and sensibly required much less computational time than the others and, hence, may be selected as the superior model in aquifer parameters estimation.

Application of adaptive neuro-fuzzy interference system on biosorption of malachite green using fir (Abies nordmanniana) cones biomass

Removal of malachite green (MG) onto fir (Abies nordmanniana) cones biomass (FCB) as a lingo-cellulosic-based structure material was investigated in the present study. Characterization of FCB was performed using Fourier transform infra red and scanning electron microscobe analyses. Several parameters (biomass dose and particle size, dye concentration, temperature, and pH) were investigated to determine optimal working conditions. Subsequently, FCB yelded a qe of 2.2 mg/g for 50 g/L FCB, in an MG solution of 110 mg/L, pH 3.3, at a temperature of 21 °C, on a 0.2-0.4 mm fraction powder after 146 h of contact. Adaptive neuro-fuzzy interference system modeling was applied to experimental data and results showed that predicted model fitted experimental data with R2 = 0.994. In a nutshell, it can be concluded that FCB shows good potential for treating MG contaminated waters.

Neuro-Adaptive Learning Approach for Predicting Production Performance and Pressure Dynamics of Gas Condensation Reservoir

In Reservoir Engineering, state-of-the-art data analysis enables engineers to characterize reservoirs and plan for developing fields. This allows production companies to save huge amounts that would otherwise be allocated to reservoir modelling and simulation and well testing. Numerical reservoir modelling and simulation is the standard use in industries today for comprehensive study of fields. However, the inflexible behaviour, development time and cost of numerical simulators are major challenges to production engineers, managers as well as modellers. On the other hand, Artificial Intelligence (AI) based reservoir models are characterised with low cost of development, short development time and fast tract analysis and flexibility to estimate the uncertainties normally found in numerical simulators. In this study, a strategy for controlling gas production and pressure drop in gas condensate reservoir is described. Numerical simulations of production rate and pressure drop were carried out first and their results were saved. An Adaptive Neuro-Fuzzy system was then developed and trained with some parts of the numerical simulation results. This AI-based system is checked and tested with part of the numerical simulation results that had not been used during the training. The developed system regenerates the numerical simulation results for both production rates and pressure drop at different Bottom Hole Pressures (BHPs) with very high accuracy (>98%). Results of this study showed that AI-based reservoir simulation can be considered a vital tool of help to production engineers, managers and modellers for a quick and more informed decision as regards field development plans that can meet operational targets.

Adaptive Neuro-Fuzzy System to Determine the Blood Glucose Level of Diabetic

Diabetes is a chronic disease that occurs when the pancreas does not produce enough insulin. The main aim of this research work was to determine the blood glucose level of diabetic patient using adaptive Neuro-fuzzy. Data of 80 diabetic patients were collected from Federal Medical Centre Jalingo. It was used for training and testing the system, Gaussian Membership function was used, hybrid training algorithm was used for training and testing, the error obtain is 0.0008333 at epoch 4 which shows that the training performance is exactly 99.99% and testing performance of the system are 99.99% at epoch 4. This shows that adaptive Neuro-fuzzy system can be applied to medical diagnosis because of the error obtained.

An Intelligent System for Real-Time Condition Monitoring of Tower Cranes

Reliability and safety are major issues in tower crane applications. A new adaptive neurofuzzy system is developed in this work for real-time health condition monitoring of tower cranes, especially for hoist gearboxes. Vibration signals are measured using a wireless smart sensor system. Fault detection is performed gear-by-gear in the gearbox. A new diagnostic classifier is proposed to integrate strengths of several signal processing techniques for fault detection. A hybrid machine learning method is proposed to facilitate implementation and improve training convergence. The effectiveness of the developed monitoring system is verified by experimental tests.

Intelligent Controller based Solar Photovoltaic with Battery Storage, Fuel Cell Integration for Power Conditioning

Providing clean power and meeting the power demand are the biggest issues in the current power scenario. These issues can be alleviated with the integration of renewable energy sources (RES) such as solar, wind etc. The solar photovoltaic source (SPV) became most prominent in the utility system. The day availability and stochastic nature of solar photovoltaic can be used efficiently with the integration of DC link system. In this work, the DC link integration is divided into two parts. In the first part the DC link is integrated with the solar photovoltaic, Fuel cell and battery storage system. The DC link voltages maintain constant with these hybrid sources. The stochastic nature of the solar photovoltaic source power is extracted using adaptive neuro fuzzy information system (ANFIS) based maximum power point tracking technique (MPPT). The battery storage system is used through the bi-directional converter and it’s operate based on the battery state of the charge. The battery state of the charge are monitored with the bi-directional controller. In the second part, the DC link integration is made with utility grid and load through the single phase voltage source Inverter. The grid synchronization can achieve with the grid synchronization unit even under dynamics in the load and sources. The power conditioning of proposed system through the DC link medium can be achieve by maintaining constant voltage at the DC link of voltage source inverter. The dynamic variation in the load and the utility grid connected hybrid system performance are simulated under different source conditions using MATLAB/Simulink software.

Proposal of an Adaptive Neurofuzzy System to Control Flow Power in Distributed Generation Systems

Systems of distributed generation have shown to be a remarkable alternative to a rational use of energy. Nevertheless, the proper functioning of them still manifests a range of challenges, including both the adequate energy dispatch depending on the variability of consumption and the interaction between generators. This paper describes the implementation of an adaptive neurofuzzy system for voltage control, regarding the changes observed in the consumption within the distribution system. The proposed design employs two neurofuzzy systems, one for the plant dynamics identification and the other for control purposes. This focus optimizes the controller using the model achieved through the identification of the plant, whose changes are produced by charge variation; consequently, this process is adaptively performed. The results show the performance of the adaptive neurofuzzy system via statistical analysis.

IAM: An Intuitive ANFIS-based method for stiction detection

Stiction in control valves is an industry-wide problem which results in degradation of control performance. A new approach to detect the presence of stiction by utilising only the PV-OP data from control loops is proposed using an Adaptive Neuro-fuzzy Inferencing System (ANFIS). Intuitively, the error between the output of an FIS model developed with stiction and a process with stiction would be minimal. When benchmarked against seventeen well-known industrial control loop case studies, the Intuitive ANFIS-based Method (IAM) accurately predicts the presence or absence of stiction in 65% of loops tested.

Artificial neural network (ANN) and adaptive neuro-fuzzy interference system (ANFIS) modelling for nickel adsorption onto agro-wastes and commercial activated carbon

Artificial neural network (ANN) and adaptive neuro–fuzzy interference system (ANFIS) were applied to model and analyze the adsorption of four different agro–wastes, namely sugarcane bagasse, passion fruit waste, orange peel and pineapple peel, and commercial activated carbon, for Ni2+ removal from aqueous solutions. The capacity of adsorption ranged from 14.75 to 63.50 mg g–1, and the results of the adsorption experiments revealed that sugarcane bagasse and orange peel presented the best adsorption performance for Ni2+ removal from aqueous solutions, even better than those of commercial activated carbon. The study also revealed that the adsorption capacity is affected by pHZPC and surface area. ANN and ANFIS were compared with the experimental data to determine the relationship of four input parameters on Ni2+ adsorption capacities: initial adsorbent concentration, adsorption time, pHZPC and surface area. The developed ANN and ANFIS could accurately predict the experimental data with correlation coefficient of 0.9926 and 0.9943, respectively. The Pearson's Chi–square measure was found to be 0.9508 for ANN and 0.5959 for ANFIS, indicating a small advantage of ANFIS over ANN.

Adaptive neuro-fuzzy and regression models for predicting microhardness and electrical conductivity of solid-state recycled EN AW 6082

In the last few years, there is a demand for developing new technologies in order to increase scrap reuse potential and CO2 emission savings. In this paper, aluminum was recycled from chips obtained by machining without any remelting in order to reduce environmental pollution and to increase material yield during the process. This process is called solid-state recycling (SSR) or direct recycling. SSR process consists of chips cleaning, cold pre-compaction, and hot direct extrusion followed by equal channel angular pressing (ECAP) at different temperatures. Influence of direct extrusion temperature, ECAP temperature, and number of ECAP passes on electrical conductivity and microhardness of the recycled EN AW 6082 aluminum chips was investigated. Microhardness and electrical conductivity of the recycled samples were comparable with commercially produced EN AW 6082. Experiments were planned utilizing design of experiments approach. Both adaptive neuro-fuzzy interference system (ANFIS) and regression models were developed and compared to describe the influence of input SSR process parameters on electrical conductivity and microhardness. Density and metallographic analysis of the recycled samples were also performed.

A New Structure for Direct Measurement of Temperature Based on Negative Temperature Coefficient Thermistor and Adaptive Neuro-fuzzy Inference System

Thermistors are very commonly used for narrow temperature-range high-resolution applications, such as in medicine, calorimetry, and near ambient temperature measurements. In particular, Negative Temperature Coefficient (NTC) thermistor is very inexpensive and highly sensitive, whose sensing temperature range and sensitivity are highly limited due to the intrinsic nonlinearity and self-heating properties of NTC thermistor at high operation currents. In this research, a new structure is proposed based on adaptive neuro-fuzzy system for the modeling of sensor nonlinearity. Apart from taking self-heating phenomenon of NTC thermistor sensor, the proposed structure also measures temperature directly, without any linearizing circuitry. Neuro-fuzzy network is trained and tested through data produced in the Laboratory environment. Examination of the proposed method on test data achieved a mean squared error of 0.0195, which is considered as a significant accomplishment.