Artificial Neural Network and Support Vector Machine Approach for Locating Faults in Radial Distribution Systems

The Industry 4.0 revolution is insisting strongly for use of machine learning-based processes and condition monitoring. In this paper, emphasis is given on machine learning-based approach for condition monitoring of shaft misalignment. This work highlights combined approach of artificial neural network and support vector machine for identification and measure of shaft misalignment. The measure of misalignment requires more features to be extracted under variable load conditions. Hence, primary objective is to measure misalignment with a minimum number of extracted features. This is achieved through normalization of vibration signal. An experimental setup is prepared to collect the required vibration signals. The normalized time domain nonstationary signals are given to discrete wavelet transform for features extraction. The extracted features such as detailed coefficient is considered for feature selection viz. Skewness, Kurtosis, Max, Min, Root mean square, and Entropy. The ReliefF algorithm is used to decide best feature on rank basis. The ratio of maximum energy to Shannon entropy is used in wavelet selection. The best feature is used to train machine learning algorithm. The rank-based feature selection has improved classification accuracy of support vector machine. The result obtained with the combined approach are discussed for different misalignment conditions.

A rule-based expert system designed to locate the faults in a distribution system is presented. Distribution system component data and network topology are stored in the database. A set of heuristic rules is compiled from the dispatcher's experience and imbedded in the rule base. To locate distribution system faults, an inference engine is developed to perform deductive reasoning on the rules in the knowledge base. The inference engine comprises three major parts: the dynamic searching method, the backtracking approach, and the set intersection operation. The expert system is implemented on a personal computer using the artificial intelligence language Prolog. To demonstrate the effectiveness of the proposed approach, the expert system has been used to locate faults in a real underground distribution system within the service area of the Taipei City District Office of the Taiwan Power Company. It is found that the expert system can identify fault location very efficiently. Therefore, it can serve as a valuable tool to help distribution system dispatchers determine fault locations.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

The rising demand for stable power supply in distribution systems has increased the importance of reliable supply. Thus, a networked distribution system (NDS) linked with individual lines is being adopted, gradually replacing the radial distribution system (RDS) currently applied to most distribution systems. Implementing the NDS can lead to various improvements in factors such as line utilization rate, acceptance rates of distributed power, and terminal voltages, while mitigating line losses. However, compared with the RDS, the NDS can experience bidirectional fault currents owing to its interconnected lines, thereby hindering protection coordination, which must be addressed before the NDS can be implemented in real-world power systems. Due to the characteristics of NDS, the reverse fault current is relatively small. However, this phenomenon becomes more severe when the high impedance fault (HIF) occurs. In this paper, the malfunction of protective devices during the HIF is directly verified and analyzed in the NDS. As a result, when the HIF occurs, the issue of the reverse protective device malfunctioning worsens because of a reduction in fault current and a failure in direction detection. To solve this issue, this work proposes a communication-based protection algorithm. Through the comparative verification of the proposed algorithm and the conventional protection method, protection coordination can be secured in the case of an HIF without new devices. It must be highlighted that the proposed method does not affect the settings of the protective device and provides a cost-effective and efficient solution since this algorithm is added independently to the existing relay.

A novel data-based method is proposed to solve a multi-step fault classification and identification problem in distribution systems. Several machine learning techniques such as artificial neural network (ANN), support vector machine (SVM), bagged tree (BT), and adaptive boosting (AdaBoost) are utilized by employing the sensor and smart meter data containing the voltage and current measurements in the presence of distributed generations (DG) at certain buses on the feeder. Considering the strength of each technique, a new method is applied to combine the predictions of different classifiers at three different steps: (a) classification of faulty phase, (b) detection of impedance level, and (c) identification of faulty line segment to improve the overall accuracy. The proposed method is validated on the modified IEEE 13 bus test system for phase to ground (L-G) type faults. The performance of the proposed method is also evaluated considering system uncertainties like missing sensor data and varying DG penetration level. K-means clustering method is used to predict the missing data to improve overall accuracy.

Electrical distribution system, which is the largest portion of networks, acts as a final power delivery path to the end users. Hence, ensuring its efficient and reliable operation is an extremely important and challenging task. Utilities, all over the world are competing for improvement in service to consumers, and the area of fault detection and diagnosis in distribution systems have attracted the significant attention of researchers in last few decades. With availability of inadequate system information and presence of high impedance faults, locating faults in a distribution system pose a major challenge to the utility operators. Most of the previous work on fault identification and location concentrated on estimating the status of circuit breakers and relays with aid of some algorithmic approaches. In this paper, a fault identification and location technique using wavelet multi-resolution approach for radial distribution systems are proposed. The current measurements at the substation, available in the distribution network have been utilized and the effectiveness of the proposed approach is demonstrated on 7-node and 19-node three-phase test systems.

The Differential Protection Scheme for Microgrid Using S-Transform

High‐impedance faults (HIFs) in electrical power distribution systems produce a very random, non‐linear and low‐magnitude fault current. The conventional overcurrent (OC) relaying‐based distribution system protection schemes find difficulty in detecting such low‐current HIFs. In this study, a simple two criteria‐based protection scheme is proposed for detection and isolation of HIFs in multi‐feeder radial distribution systems. It utilises one‐cycle sum of superimposed components of residual voltage for HIF detection and the maximum value of one‐cycle sum of superimposed components of negative‐sequence current for faulted feeder identification. The performance of the proposed scheme is evaluated for a wide variety of possible test cases by generating data through power systems computer‐aided design/electro‐magnetic transient design and control software. Results clearly show that the proposed scheme can assist conventional OC relay for detection and isolation of HIFs in distribution systems with any grounding connections in a more reliable and faster way.

Estimation of static formation temperatures in geothermal wells by using an artificial neural network approach

An exhaustive literature survey shows, that a very little effort has been done towards Artificial Neural Network (ANN) approach in the area of concrete technology [1, 2, 3]. In the present investigation, development of ANN approach for prognostication of concrete mix proportion in lieu of conventional laboratory approach. The traditional lab approach attracts some drawbacks such as lot of manual involvement, time consuming, chances of creeping of human errors, uncertain prediction and always invasive in nature. Hence to reduce above said drawbacks, this study is undertaken to develop a ANN between concrete mix ingredient properties namely maximum size of aggregate, degree of quality control, degree of workability, type of exposure, characteristic compressive strength required in the field at 28 days and concrete mix proportion. Prognostication of concrete mix proportion is essential for all structural works. The present work deals with collection of huge input data base from literatures, ANN’s training and its testing are adopted to fix the appropriate weighted matrix (Illustrated in Fig [1]) which in turn Prognosticates the appropriate concrete mix proportion. Indian standard code (IS 10262:1982) procedure is also adopted to compare the concrete mix proportions of same samples. The Prognosticated concrete mix proportion from ANN approach yielded very high accuracy results (As shown in fig [2]) compared with IS code method. To account for larger sample data the results of this work will contribute for the prognostication of concrete mix proportions up to a certain degree of level, which will assist a structural engineer in estimation of concrete mix proportion, with minimum effort and non- invasive technique.

Artificial neural network approach for prediction of ammonia emission from field-applied manure and relative significance assessment of ammonia emission factors

The common fault in distribution systems due to line outages consists of single-line-to-ground (SLG) faults, with low or high fault impedance. The presence of arcing is commonplace in high impedance SLG faults. Recently, artificial intelligence (AI) based techniques have been introduced for low/high impedance fault diagnosis in ungrounded distribution systems and high-impedance fault diagnosis in grounded distribution systems. So far no tool has been developed to identify, locate and classify faults on grounded and ungrounded systems. This paper describes an integrated package for fault diagnosis in either grounded or ungrounded distribution systems. It utilizes rule-based schemes as well as artificial neural networks (ANN) to detect, classify and locate faults. Its application on sample test data as well as field test data are reported in the paper.

Silk fibroin/xanthan composite was investigated as a suitable biomedical material for controlled drug delivery, and blending ratios of silk fibroin and xanthan were optimized by response surface methodology (RSM) and artificial neural network (ANN) approach. A non-linear ANN model was developed to predict the effect of blending ratios, percentage swelling and porosity of composite material on cumulative percentage release. The efficiency of RSM was assessed against ANN and it was found that ANN is better in optimizing and modeling studies for the fabrication of the composite material. In-vitro release studies of the loaded drug chloramphenicol showed that the optimum composite scaffold was able to minimize burst release of drug and was followed by controlled release for 5days. Mechanistic study of release revealed that the drug release process is diffusion controlled. Moreover, during tissue engineering application, investigation of release pattern of incorporated bioactive agent is beneficial to predict, control and monitor cellular response of growing tissues. This work also presented a novel insight into usage of various drug release model to predict material properties. Based on the goodness of fit of the model, Korsmeyer-Peppas was found to agree well with experimental drug release profile, which indicated that the fabricated material has swellable nature. The chloramphenicol (CHL) loaded scaffold showed better efficacy against gram positive and gram negative bacteria. CHL loaded SFX55 (50:50) scaffold shows promising biocomposite for drug delivery and tissue engineering applications.

In order to run power plant operations smoothly, power plant faults need to be detected, located, and classified quickly. For this, artificial neural network approaches are considered significant tool in related application of power system. This research paper focuses on power plant faults detection and classification using intelligent approach of artificial neural network. There is frequent nature of power plant faults, in particular transmission line faults. The key intention of this research is to use ANN approach to detect and classify power plant faults, and in the end to compare their performance and accuracy. In this proposal, the three-phase current and voltage are use as input. Backpropagation (BPNN) technique is used in developing the learning algorithm for multi-layer neural network that can accurately and reliably classify faults. The multi-layer neural network such as Levenberg-Marquardt and Bayesian regularization are used for offline training of the data. Findings from this research show that both Levenberg-Marquardt and Bayesian regularization show good accuracy to detect and classify faults in transmission line. However, LM algorithm trained data much faster than BR. LM shows fastest detection of faults with overall mean square error value of 6.13785e-3 at the 91 epoch. Whereas, BR trained network at the 717 epochs with overall mse of 4.67123e-3. It is to be noted that the accuracy of trained network to classify all the fault types is found to be 81.1%, which is higher than previous studies using Backpropagation neural network technique. Training and test results show that these neural network-based methods can effectively detect and classify faults and faults types and have adequate performance. Matlab is used in this research because it has toolbox, which has effective features to train required data and achieve effective result.

Software project development has high failure rate. Software project risk management may gain a high rate of return in investment. Establishing an intelligent risk evaluation model for project will be valuable in the analysis and control of project risks. In this paper, we employed neural network (NN) and support vector machine (SVM) approaches to establish a model for risk evaluation in project development. In the model, the input is a vector of software risk factors that were obtained through interview with 30 experts, and the output is the final outcome of the project. The data for modeling were collected from 120 real software projects through questionnaires. The experiment shows the model is valid. Interestingly, SVM is a powerful supervised learning method, and some believe that it is a more promising classification method that may someday supercede NN. In our study, the standard neural network model had lower prediction accuracy compared to SVM due to its tendency in finding local optima. However, after attempt in optimizing the neural network model with genetic algorithm, the experimental results showed that our enhanced model surpassed SVM in performance.

An improved fault-location method for distribution system using wavelets and support vector regression