Automatic detection of solar array faults reduces maintenance costs and increases efficiency. In this paper, we address the problem of fault detection, localization, and classification in utility-scale photovoltaic (PV) arrays using machine learning methods. More specifically, we develop a series of customized neural networks for detection and classification of solar array faults. We evaluate fault detection and classification using metrics such as accuracy, confusion matrices, and the Risk Priority Number (RPN). We examine and assess the use of customized neural networks with dropout regularizers. We develop and evaluate neural network pruning strategies and illustrate the trade-off between fault classification model accuracy and algorithm complexity. Our approach promises to elevate the performance and robustness of PV arrays and compares favorably against existing methods.

A novel distributed algorithm for estimating the maximum of the node initial state values in a network, in the presence of additive communication noise is proposed. Conventionally, the maximum is estimated locally at each node by updating the node state value with the largest received measurements in every iteration. However, due to the additive channel noise, the estimate of the maximum at each node drifts at each iteration and this results in nodes diverging from the true max value. Max-plus algebra is used as a tool to study this ergodic process. The subadditive ergodic theorem is invoked to establish a constant growth rate for the state values due to noise, which is studied by analyzing the max-plus Lyapunov exponent of the product of noise matrices in a max-plus semiring. The growth rate of the state values is upper bounded by a constant which depends on the spectral radius of the network and the noise variance. Upper and lower bounds are derived for both fixed and random graphs. Finally, a two-run algorithm robust to additive noise in the network is proposed and its variance is analyzed using concentration inequalities. Simulation results supporting the theory are also presented.

A distributed spectral clustering algorithm to group sensors based on their location in a wireless sensor network (WSN) is proposed. For machine learning and data mining applications in WSN's, gathering data at a fusion center is vulnerable to attacks and creates data congestion. To avoid this, we propose a robust distributed clustering method without a fusion center. The algorithm combines distributed eigenvector computation and distributed K-means clustering. A distributed power iteration method is used to compute the eigenvector of the graph Laplacian. At steady state, all nodes converge to a value in the eigenvector of the algebraic connectivity of the graph Laplacian. Clustering is carried out on the eigenvector using a distributed K-means algorithm. Location information of the sensor is only used to establish the network topology and this information is not exchanged in the network. This algorithm works for any connected graph structure. Simulation results supporting the theory are also provided.

Sampling one or more effective solutions from large search spaces is a recurring idea in machine learning (ML), and sequential optimization has become a popular solution. Typical examples include data summarization, sample mining for predictive modeling, and hyperparameter optimization. Existing solutions attempt to adaptively trade off between global exploration and local exploitation, in which the initial exploratory sample is critical to their success. While discrepancy-based samples have become the de facto approach for exploration, results from computer graphics suggest that coverage-based designs, e.g., Poisson disk sampling, can be a superior alternative. In order to successfully adopt coverage-based sample designs to ML applications, which were originally developed for 2-D image analysis, we propose fundamental advances by constructing a parameterized family of designs with provably improved coverage characteristics and developing algorithms for effective sample synthesis. Using experiments in sample mining and hyperparameter optimization for supervised learning, we show that our approach consistently outperforms the existing exploratory sampling methods in both blind exploration and sequential search with Bayesian optimization.

A consensus based distributed algorithm to compute the spectral radius of a network is proposed. The spectral radius of the graph is the largest eigenvalue of the adjacency matrix, and is a useful characterization of the network graph. Conventionally, centralized methods are used to compute the spectral radius, which involves eigenvalue decomposition of the adjacency matrix of the underlying graph. Our distributed algorithm uses a simple update rule to reach consensus on the spectral radius, using only local communications. We consider time-varying graphs to model packet loss and imperfect transmissions, and provide the convergence characteristics of our algorithm, for both static and time-varying graphs. We prove that the convergence error is a function of principal eigenvector of adjacency matrix of the graph and reduces as $\mathcal {O}(1/t)$ , where $t$ is the number of iterations. The algorithm works for any connected graph structure. Simulation results supporting the theory are also presented.