Statistical Methods for Degradation Estimation and Anomaly Detection in Photovoltaic Plants (original) (raw)
Related papers
Progress in Photovoltaics: Research and Applications
Fault detection and classification in photovoltaic (PV) systems through real-time monitoring is a fundamental task that ensures quality of operation and significantly improves the performance and reliability of operating systems. Different statistical and comparative approaches have already been proposed in the literature for fault detection; however, accurate classification of fault and loss incidents based on PV performance time series remains a key challenge. Failure diagnosis and trend-based performance loss routines were developed in this work for detecting PV underperformance and accurately identifying the different fault types and loss mechanisms. The proposed routines focus mainly on the differentiation of failures (e.g., inverter faults) from irreversible (e.g., degradation) and reversible (e.g., snow and soiling) performance loss factors based on statistical analysis. The proposed routines were benchmarked using historical inverter data obtained from a 1.8 MWp PV power plant. The results demonstrated the effectiveness of the routines for detecting failures and loss mechanisms and the capability of the pipeline for distinguishing underperformance issues using anomaly detection and change-point (CP) models. Finally, a CP model was used to extract significant changes in time series data, to detect soiling and cleaning events and to estimate both the performance loss and degradation rates of fielded PV systems.
A Statistical Tool to Detect and Locate Abnormal Operating Conditions in Photovoltaic Systems
Sustainability
The paper is focused on the energy performance of the photovoltaic systems constituted by several arrays. The main idea is to compare the statistical distributions of the energy dataset of the arrays. For small-medium-size photovoltaic plant, the environmental conditions affect equally all the arrays, so the comparative procedure is independent from the solar radiation and the cell temperature; therefore, it can also be applied to a photovoltaic plant not equipped by a weather station. If the procedure is iterated and new energy data are added at each new run, the analysis becomes cumulative and allows following the trend of some benchmarks. The methodology is based on an algorithm, which suggests the user, step by step, the suitable statistical tool to use. The first one is the Hartigan's dip test that is able to discriminate the unimodal distribution from the multimodal one. This stage is very important to decide whether a parametric test can be used or not, because the parametric tests-based on known distributions-are usually more performing than the nonparametric ones. The procedure is effective in detecting and locating abnormal operating conditions, before they become failures. A case study is proposed, based on a real operating photovoltaic plant. Three periods are separately analyzed: one month, six months, and one year.
Pertanika Journal of Science and Technology
This paper presents the characterization and performance of six-year field data for two different systems of PV module technologies from the rooftop grid-connected system installed at Universiti Teknologi MARA (UiTM) Shah Alam. Two different PV module technologies are used as case studies to establish a method of anomaly detection on the system performance. The selected parameters such as string voltage, string current and AC power output are used in the analysis, while solar irradiance and module temperature are used as a reference basis. Based on the results obtained, both systems having degradation rates differently. System A had shown stable performance before it degraded by 16.09% after the 4th-year of operation, whereas system B continuously decreased by 39.35% during the monitored period. However, the string current of system A degraded up to 4.4% and, interestingly, no degradation for the string voltage. In contrast, system B has experienced a degradation of the string curre...
Statistics to Detect Low-Intensity Anomalies in PV Systems
Energies
The aim of this paper is the monitoring of the energy performance of Photovoltaic (PV) plants in order to detect the presence of low-intensity anomalies, before they become failures or faults. The approach is based on several statistical tools, which are applied iteratively as the data are acquired. At every loop, new data are added to the previous ones, and a proposed procedure is applied to the new dataset, therefore the analysis is carried out on cumulative data. In this way, it is possible to track some specific parameters and to monitor that identical arrays in the same operating conditions produce the same energy. The procedure is based on parametric (ANOVA) and non-parametric tests, and results effective in locating anomalies. Three cumulative case studies, based on a real operating PV plant, are analyzed.
Energies
This paper deals with the monitoring of the performance of a photovoltaic plant, without using the environmental parameters such as the solar radiation and the temperature. The main idea is to statistically compare the energy performances of the arrays constituting the PV plant. In fact, the environmental conditions affect equally all the arrays of a small-medium-size PV plant, because the extension of the plant is limited, so any comparison between the energy distributions of identical arrays is independent of the solar radiation and the cell temperature, making the proposed methodology very effective for PV plants not equipped with a weather station, as it often happens for the PV plants located in urban contexts and having a nominal peak power in the 3÷50 kWp range, typically installed on the roof of a residential or industrial building. In this case, the costs of an advanced monitoring system based on the environmental data are not justified, consequently, the weather station is often also omitted. The proposed procedure guides the user through several inferential statistical tools that allow verifying whether the arrays have produced the same amount of energy or, alternatively, which is the worst array. The procedure is effective in detecting and locating abnormal operating conditions, before they become failures.
Multivariate statistical monitoring of photovoltaic plant operation
Energy Conversion and Management, 2020
Detecting anomalies in a photovoltaic system play a core role in keeping the desired performance and meeting requirements and specication. For this propose, a simple and ecient monitoring methodology using principal component analysis model and multivariate monitoring schemes is designed to monitor PV systems. The principal component analysis model is used to generate residuals for anomaly detection. Then, the residuals are examined by computing the monitoring schemes (T 2 and square predicted error) for the purpose of fault detection. However, these conventional schemes are usually derived under the hypothesis of Gaussian distribution. Thus, the major aim of this paper is to bridge this gap by designing assumption-free principal component analysis-based schemes. Specically, a nonparametric approach using kernel density estimation is proposed to set thresholds for decision statistics and compared with the parametric counterparts. Real measurements from an actual 9.54 kWp grid-connected PV system are used to illustrate the performance of the studied methods. To evaluate the fault detection capabilities of the proposed approach, six case studies are investigated, one concerning a string fault, one involving a partial shading, and one concerning the loss of energy caused by inverter disconnections. Results testify the ecient performance of the proposed method in monitoring a PV system and its greater exibility when using nonparametric detection thresholds.
Tailored Algorithms for Anomaly Detection in Photovoltaic Systems
Energies, 2020
The fastest-growing renewable source of energy is solar photovoltaic (PV) energy, which is likely to become the largest electricity source in the world by 2050. In order to be a viable alternative energy source, PV systems should maximise their efficiency and operate flawlessly. However, in practice, many PV systems do not operate at their full capacity due to several types of anomalies. We propose tailored algorithms for the detection of different PV system anomalies, including suboptimal orientation, daytime and sunrise/sunset shading, brief and sustained daytime zero-production, and low maximum production. Furthermore, we establish simple metrics to assess the severity of suboptimal orientation and daytime shading. The proposed detection algorithms were applied to a set of time-series of electricity production in Portugal, which are based on two periods with distinct weather conditions. Under favourable weather conditions, the algorithms successfully detected most of the time-series labelled with either daytime or sunrise/sunset shading, and with either sustained or brief daytime zero-production. There was a relatively low percentage of false positives, such that most of the anomaly detections were correct. As expected, the algorithms tend to be more robust under favourable rather than under adverse weather conditions. The proposed algorithms may prove to be useful not only to research specialists, but also to energy utilities and owners of small-and medium-sized PV systems, who may thereby effortlessly monitor their operation and performance.
Renewable Energy, 2019
Over the last decade, research into photovoltaic (PV) technology has shifted from a race for the highest efficiency to the increase of the performance reliability in the field. A major part of current research activities focuses on the reliability of the installations and the guaranteed lifetime output through constant, solid and traceable PV plant monitoring. In this domain, several PV monitoring strategies for the early diagnosis of failures in grid-connected PV systems have been proposed in the literature and this study seeks to provide an overview of all the data analytic methods used by the research community and industry for the detection and classification of failures from acquired performance data of grid-connected PV systems. Insight into the performance monitoring requirements (parameters and resolution) for the detection of failures in monitored PV systems, as well as the various techniques used for their classification is also provided. Finally, this overview covers the data analytic methods based on statistical, numerical and electrical signature analysis and are summarised according to the type of failure, input requirements and validation procedure.
A method for detecting malfunctions in PV solar panels based on electricity production monitoring
Solar Energy, 2017
In this paper a new method is developed for automatically detecting outliers or faults in the solar energy production of identical sets (sister arrays) of photovoltaic (PV) solar panels. The method involves a two-stage unsupervised approach. In the first stage, "in control" energy production data are created by using outlier detection methods and functional principal component analysis in order to remove global and local outliers from the data set. In the second stage, control charts for the "in control" data are constructed using both a parametric method and three non-parametric methods. The control charts can be used to detect outliers or faults in the production data in real-time or at the end of the day. As an illustration, the method is applied to analysis of the real energy production data of six sets of "identical" PV solar panels over a period of three years. Tests indicate that the proposed method is able to successfully detect a reduction in efficiency in one of the solar panel sets by up to 5%. Control charts based on parametric and non-parametric methods both show good performance results.