-
Testing Topological Data Analysis for Condition Monitoring of Wind Turbines
Authors:
Simone Casolo,
Alexander Stasik,
Zhenyou Zhang,
Signe Riemer-Sorensen
Abstract:
We present an investigation of how topological data analysis (TDA) can be applied to condition-based monitoring (CBM) of wind turbines for energy generation. TDA is a branch of data analysis focusing on extracting meaningful information from complex datasets by analyzing their structure in state space and computing their underlying topological features. By representing data in a high-dimensional s…
▽ More
We present an investigation of how topological data analysis (TDA) can be applied to condition-based monitoring (CBM) of wind turbines for energy generation. TDA is a branch of data analysis focusing on extracting meaningful information from complex datasets by analyzing their structure in state space and computing their underlying topological features. By representing data in a high-dimensional state space, TDA enables the identification of patterns, anomalies, and trends in the data that may not be apparent through traditional signal processing methods. For this study, wind turbine data was acquired from a wind park in Norway via standard vibration sensors at different locations of the turbine's gearbox. Both the vibration acceleration data and its frequency spectra were recorded at infrequent intervals for a few seconds at high frequency and failure events were labelled as either gear-tooth or ball-bearing failures. The data processing and analysis are based on a pipeline where the time series data is first split into intervals and then transformed into multi-dimensional point clouds via a time-delay embedding. The shape of the point cloud is analyzed with topological methods such as persistent homology to generate topology-based key health indicators based on Betti numbers, information entropy and signal persistence. Such indicators are tested for CBM and diagnosis (fault detection) to identify faults in wind turbines and classify them accordingly. Topological indicators are shown to be an interesting alternative for failure identification and diagnosis of operational failures in wind turbines.
△ Less
Submitted 24 June, 2024;
originally announced June 2024.
-
Balancing the Norwegian regulated power market anno 2016 to 2022
Authors:
Pål Forr Austnes,
Signe Riemer-Sørensen,
David Andreas Bordvik,
Christian Andre Andresen
Abstract:
The balancing market for power is designed to account for the difference between predicted supply/demand of electricity and the realised supply/demand. However, increased electrification of society changes the consumption patterns, and increased production from renewable sources leads to larger un-predicted fluctuations in production, both effects potentially leading to increased balancing. We ana…
▽ More
The balancing market for power is designed to account for the difference between predicted supply/demand of electricity and the realised supply/demand. However, increased electrification of society changes the consumption patterns, and increased production from renewable sources leads to larger un-predicted fluctuations in production, both effects potentially leading to increased balancing. We analyse public market data for the balancing market (manual Frequency Restoration Reserve) for Norway from 2016 to 2022 to investigate and document these effects. The data is newer than for similar analyses and the eight years of data is more than double the time span previously covered. The main findings are: a) The balancing volumes are dominated by hours of zero regulation but for non-zero hours, the balancing volumes are increasing during the eight-year period. b) The balancing prices are primarily correlated with day-ahead prices and secondary with balancing volumes. The latter correlation is found to be increasingly non-linear with time. c) The balancing volumes and the price difference between balancing price and day-ahead price are strongly correlated with the previous hour. d) The increasing share of wind power has not impacted the frequency of balancing, which has remained stable during the 8 years studied. However, the volumes and share of balancing power compared to overall production have increased, suggesting that the hours which are inherently difficult to predict remain the same. e) Market data alone cannot predict balancing volumes. If attempting, the auto-correlation becomes the main source of information.
△ Less
Submitted 14 February, 2024;
originally announced February 2024.
-
Pseudo-Hamiltonian system identification
Authors:
Sigurd Holmsen,
Sølve Eidnes,
Signe Riemer-Sørensen
Abstract:
Identifying the underlying dynamics of physical systems can be challenging when only provided with observational data. In this work, we consider systems that can be modelled as first-order ordinary differential equations. By assuming a certain pseudo-Hamiltonian formulation, we are able to learn the analytic terms of internal dynamics even if the model is trained on data where the system is affect…
▽ More
Identifying the underlying dynamics of physical systems can be challenging when only provided with observational data. In this work, we consider systems that can be modelled as first-order ordinary differential equations. By assuming a certain pseudo-Hamiltonian formulation, we are able to learn the analytic terms of internal dynamics even if the model is trained on data where the system is affected by unknown dam** and external disturbances. In cases where it is difficult to find analytic terms for the disturbances, a hybrid model that uses a neural network to learn these can still accurately identify the dynamics of the system as if under ideal conditions. This makes the models applicable in some situations where other system identification models fail. Furthermore, we propose to use a fourth-order symmetric integration scheme in the loss function and avoid actual integration in the training, and demonstrate on varied examples how this leads to increased performance on noisy data.
△ Less
Submitted 2 January, 2024; v1 submitted 9 May, 2023;
originally announced May 2023.
-
A Levenberg-Marquardt algorithm for sparse identification of dynamical systems
Authors:
Mark Haring,
Esten Ingar Grøtli,
Signe Riemer-Sørensen,
Katrine Seel,
Kristian Gaustad Hanssen
Abstract:
Low complexity of a system model is essential for its use in real-time applications. However, sparse identification methods commonly have stringent requirements that exclude them from being applied in an industrial setting. In this paper, we introduce a flexible method for the sparse identification of dynamical systems described by ordinary differential equations. Our method relieves many of the r…
▽ More
Low complexity of a system model is essential for its use in real-time applications. However, sparse identification methods commonly have stringent requirements that exclude them from being applied in an industrial setting. In this paper, we introduce a flexible method for the sparse identification of dynamical systems described by ordinary differential equations. Our method relieves many of the requirements imposed by other methods that relate to the structure of the model and the data set, such as fixed sampling rates, full state measurements, and linearity of the model. The Levenberg-Marquardt algorithm is used to solve the identification problem. We show that the Levenberg-Marquardt algorithm can be written in a form that enables parallel computing, which greatly diminishes the time required to solve the identification problem. An efficient backward elimination strategy is presented to construct a lean system model.
△ Less
Submitted 23 March, 2022;
originally announced March 2022.
