Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Distribution systems are continuously exposed to fault occurrences due to various reasons, such as lightning strike, failure of power system components due to aging of equipment and human errors. These phenomena affect the system reliability and results in expensive repairs, lost of productivity and power loss to customers. Since fault is unpredictable, a fast fault location and isolation is required to minimize the impact of fault in distribution systems. Therefore, many methods have been developed since the past to locate and detect faults in distribution systems with distributed generation. The methods can be divided into two categories, conventional and artificial intelligence techniques. Conventional techniques include travelling wave method and impedance based method while artificial intelligence techniques include Artificial Neural Network (ANN), Support Vector Machine (SVM), Fuzzy Logic, Genetic Algorithm (GA) and matching approach. However, fault location using intelligent methods are challenging since they require training data for processing and are time consuming. In this paper, most of the techniques that have been developed since the past and commonly used to locate and detect faults in distribution systems with distributed generation are reviewed. Research works in fault location area, the working principles, advantages and disadvantages of past works related to each fault location technique are highlighted in this paper. Hence, from this review, the opportunities in fault location research area in power distribution system can be explored further.

Fault location and detection techniques in power distribution systems with distributed generation: A review

Accurate and timely diagnosis of transmission line faults is key for reliable operations of power systems. Existing fault-diagnosis methods rely on expert knowledge or extensive feature extraction, which is also highly dependent on expert knowledge. Additionally, most methods for fault diagnosis of transmission lines require multiple separate subalgorithms for fault classification and location performing each function independently and sequentially. In this research, an integrated framework combining fault classification and location is proposed by applying an innovative machine-learning algorithm: the summation-wavelet extreme learning machine (SW-ELM) that integrates feature extraction in the learning process. As a further contribution, an extension of the SW-ELM, i.e., the summation-Gaussian extreme learning machine (SG-ELM), is proposed and successfully applied to transmission line fault diagnosis. SG-ELM is fully self-learning and does not require ad-hoc feature extraction, making it deployable with minimum expert subjectivity. The developed framework is applied to three transmission-line topologies without any prior parameter tuning or ad-hoc feature extraction. Evaluations on a simulated dataset show that the proposed method can diagnose faults within a single cycle, remain immune to fault resistance and inception angle variation, and deliver high accuracy for both tasks of fault diagnosis: fault type classification and fault location estimation.

Combined Fault Location and Classification for Power Transmission Lines Fault Diagnosis With Integrated Feature Extraction

/pdf/support-vector-machine-based-fault-classification-and-38mohwlpza.pdf

Support vector machine based fault classification and location of a long transmission line

In the last decade, artificial intelligence (AI) techniques have been extensively used for maximum power point tracking (MPPT) in the solar power system. This is because conventional MPPT techniques are incapable of tracking the global maximum power point (GMPP) under partial shading condition (PSC). The output curve of the power versus voltage for a solar panel has only one GMPP and multiple local maximum power points (MPPs). The integration of AI in MPPT is crucial to guarantee the tracking of GMPP while increasing the overall efficiency and performance of MPPT. The selection of AI-based MPPT techniques is complicated because each technique has its own merits and demerits. In general, all of the AI-based MPPT techniques exhibit fast convergence speed, less steady-state oscillation and high efficiency, compared with the conventional MPPT techniques. However, the AI-based MPPT techniques are computationally intensive and costly to realize. Overall, the hybrid MPPT is favorable in terms of the balance between performance and complexity, and it combines the advantages of conventional and AI-based MPPT techniques. In this paper, a detailed comparison of classification and performance between 6 major AI-based MPPT techniques have been made based on the review and MATLAB/Simulink simulation results. The merits, open issues and technical implementations of AI-based MPPT techniques are evaluated. We intend to provide new insights into the choice of optimal AI-based MPPT techniques.

/pdf/artificial-intelligence-based-mppt-techniques-for-solar-122xph2f0n.pdf

Artificial Intelligence Based MPPT Techniques for Solar Power System: A review

Transient stability is very important in power system. Large disturbances like fault in a transmission line are a concern which needs to be disconnected as quickly as possible in order to restore the transient stability. Faulty current and voltage signals are used for location, detection and classification of faults in a transmission network. Relay detects an abnormal signal, and then the circuit breaker disconnects the unhealthy transmission line from the rest of the health system. This paper discusses various signal processing techniques, impedance-based measurement method, travelling wave phenomenon-based method, artificial intelligence-based method and some special technique for the detection, location and classification of various faults in a transmission network. In this survey, paper signifies all method and techniques till August 2017. This compact and effective survey helps the researcher to understand different techniques and methods.

Fault detection, location and classification of a transmission line

Electrical power load forecasting has at all times been an major issue in energy trade. Load forecasting is generally made through developing models on relative knowledge, reminiscent of local weather and previous load demand knowledge. Such forecast is almost always aimed towards brief-time period prediction like one-day forward prediction, on the ground that longer interval prediction (mid-term or long term) will not be reliant as a result of error propagation. The accurateness of load predicting needs a massive effect on an electricity services process and making cost. Exact load predicting is hence significant, particularly with the fluctuation shappening within the utility industry because of deregulation and competition. Several outmoded approaches such as regression model, time series model and expert system have been proposed for short term load forecasting by different degree of achievement. Artificial Neural Network established short term load forecasting model has its own importance due to its transparent model, easy implementation, and superior performance. In this paper ANN trained through back propagation in combination with Genetic Algorithm model is used aimed at short term load forecasting. In back propagation, the weights of neuron changes according to the gradient descent which may tend to local minima, so Genetic Algorithm is implemented which gives better forecasting result as compared to back propagation.

Short term load forecasting by artificial neural network

This paper gives an analyzing study on fault detection and classification in a long transmission line which is series compensated using artificial neural network and wavelet transform. The proposed scheme makes use of one cycle pre fault and one cycle post fault samples of the three phase current signals to find the ground current signal. Daubechies is used as the mother wavelet while using the discrete wavelet transform technique.The differential energy, based on discrete wavelet transform is applied to feed a system, designed for the classification of all eleven fault types. Finally, the optimal features which are the energies obtained from discrete wavelet transform of the current signals selected are fed to neural networks for the purpose of fault classification. The reliability of the suggested technique is experienced in a 735-kV, 50 Hz power systems under altered operating settings using MATLAB . The fault was detected and classified using discrete wavelet transform as well as artificial neural network. The results indicate that the proposed scheme can correctly classify every possible fault with large variations in system conditions.

Fault Detection and Classification of a Transmission Line Using Discrete Wavelet Transform & Artificial Neural Network

A hybrid technique for fault distance estimation in a distribution line with wind farm is presented in this paper. Here, one cycle of post fault current samples are taken for fault location from the distributed generation end. The collected samples are then decomposed by wavelet transform and thereafter six statistical features are extracted from the reconstructed detail coefficients of the current signal. Further best features are selected from the total feature set by forward feature selection method. These selected features are then fed as input to the artificial neural network for fault location. In the proposed method, the simulation conditions for the test pattern are completely different from the train one in order to make it robust. Simulation result shows that the proposed hybrid fault location method gives high accuracy for the distribution system.

Debani Prasad Mishra

Papers

Support vector machine based fault classification and location of a long transmission line

Fault detection, location and classification of a transmission line

Short term load forecasting by artificial neural network

Fault Detection and Classification of a Transmission Line Using Discrete Wavelet Transform & Artificial Neural Network

Artificial Intelligence Based Fault Location in a Distribution System