PQ Data, whether its voltage disturbances, current unbalance or any such other critical parameter, is characterized as voluminous. While it’s a challenge to deal with this large volume of data, it also carries many valuable insights.
To acquire, organize and analyze PQ related data is often intimidating, mostly because of its volume. In fact, it still remains to be a critical hurdle in understanding the very nature of PQ. The data related to PQ is also heterogeneous, distributed with low value density when considering the PQ monitoring system of the grid. Data Analytics along with data cleaning, cluster analysis and co-relation can be leveraged to unfold a new set of insights to improve PQ. Further, decision models, based on data analytics, with specific goals, such as predictive maintenance or energy efficiency, can be built to ensure sustenance of good PQ.
This blog evaluates the possibilities of using data analytics at the consumer end and outlines a set of priorities for the users to leverage analytics to improve PQ.
DATA ANALYTICS – WHAT CAN IT DELIVER FOR PQ?
Given the volume of data, analytics as a science has a vast number of potential applications when it comes to improving PQ. However, challenges in acquisition of data – the most important raw material for analytics, imposes severe limitations on the possibilities. The cost of developing analytics, awareness among the professionals in energy domain, subject matter expertise and availability of historical data (for companies developing analytics solutions) are a few other prominent limitations.
Several studies, tools and initiatives at the consumer end are leveraging the power of data analytics in varying levels of strengths. The schematic below attempts to summarise a few prominent areas where end users are leveraging the power of analytics to improve PQ in their electrical network.
Considering the electrical network wide presence of PQ, the current set of issues which can be effectively addressed through analytics are of limited nature. However, with falling data acquisition costs and rising computational powers of PQ measurement and monitoring devices, the scope for using data analytics to improve PQ will only grow in near future. A recent study that focused on using data analytics to predict transformer failures met with mixed success. The study revealed that use of predictive data analytics models using PQ parameters to assess the possibility of equipment failure yields reasonable results for individual devices with the data acquired downstream from those devices. However, to perfect the data analytics models at the electrical network level, Utilities will need to install infrastructure to process higher volumes of data.
DATA ANALYTICS APPLICATIONS IN IMPROVING PQ
The applications below are represented from various research studies that indicate use of Data Analytics for long-term improvement of PQ. Typically, these applications call for historical data analysis and need strategic initiatives to improve PQ.
Fault anticipation in Power Distribution
Patterns in unbalanced Voltage or Current waveforms can be spotted very effectively using data analytics. Expertise in disturbance analysis, coupled with data analytics, helps to accurately predict the disturbance source and synchronise the data for validation and advanced analysis.
Fault locations and Root Cause
Investigating the root causes of PQ problems can be challenging. With Data analytics providing the capability to observe deviations at a granular level and correlate the same to reveal causal and effect correlations helps to understand fault locations for PQ with its root causes.
Creating harmonic signatures for critical assets
With widespread use of acquisition devices such as the smart meters and PQ huge volume of asset level performance data are available. Typically, the data acquisition devices come with limited data storage and computational configuration to perform long-range analysis. Often, such systems rely on the cloud based data analytics platforms to store and analyse data. The ability to draw meaningful insights from data analytics are driven by the reliability of communication networks, which includes the cloud based data storage and the supporting IT infrastructure to transfer high-volumes of data. The long term performance of such data can be mined to create harmonic signatures of the equipment. These signatures are highly valuable when it comes to predicting the impact of the asset on the electrical network and also understand the results of changes in electrical network on the equipment.
Parallel to this, the concerns about data security and responsible use of insights from data analytics must also be rightly addressed. The sensitive data and insights on critical assets stands the risk of being misused in many ways, starting from a centralised and monopolistic control (data colonisation) to the threat of system level cyber attacks. The large-scale data related to Power industry is vital to the growth and sustenance of economy. Therefore, legal as well as ethical thrust on ensuring availability of such data for larger benefit of the society will be required with greater use of data analytics.
Load patterns to estimate PQ instances
PQ concerns emerge with presence of non-linear nature of the loads. With availability of historical data on demand and consumption patterns for the equipment or a load center, data analytics can be employed to estimate the occurrence of PQ instances with changes in load patterns. Such analysis can also be leveraged to enforce initiatives such as the proposed change of law to separate wires and supply business (Carriage and Content) in electricity distribution. The insights can help the Regulators to ensure quality power is made available by utilities or network owners, especially in view of growth of Distributed Generating. At a macro level, it also allows more accurate load forecasts and better long-term planning for the Regulators.
Electrical power disturbances offer information that can provide deep insights into the electrical network health as a whole. The present nature of PQ monitoring platforms is of general nature. Data analytics enables significant knowledge extraction and synthesis and will therefore emerge as a reliable platform for power system monitoring in near future. However, setting clear goals, with a calculated imagination of the benefits, before investing in the data analytics initiatives, will be key to success for the end users.
- Power Quality Data Analytics – http://canrev.ieee.ca/cr73/Power_Quality_Data_Analytics.pdf
- PQ Monitoring using unsupervised Data Mining https://ro.uow.edu.au/cgi/viewcontent.cgi?referer=https://www.google.com/&httpsredir=1&article=2230&context=infopapers
- Predicting Distribution Transformer Failures – https://www.tdworld.com/asset-managementservice/predicting-distribution-transformer-failures