Data Analytics

Tata Power-DDL has achieved envious feats in loss reduction, reliability and other key parameters through technology adoption and innovation. In this journey, the organization has relied on building internal capacity and capability to learn and implement new technologies. Data Analytics was identified by the organization six years ago, as a key enabler to bring data driven approach to complement the conventional wisdom of power sector professionals. Accordingly infrastructure and skills were gradually built over the years and Big Data Practices were initiated from 2018 onwards. Subsequently business critical analytics practices have been developed which have contributed immensely to improve key parameters of the organization.

Analytics landscape

One of the most complex IT-OT integrated architecture comprising of state of art systems like GIS, ADMS, Big Data and MDM etc.

Network Management & Optimization

Automatic Consumer Mapping Correction

Transformer to consumer mapping anomaly detection and predict correct mapping using integrated analytics on GIS and Smart Meter data. The benefit of this is an expected saving of INR 4,658,000 which is recurring in nature. Accurate network topology which is integral to correct outage management and loss calculations.

Asset Sweating

Optimize transformer utilization and ensuring load balancing resulting CAPEX saving and enhanced asset life through data captured from smart meter and buildings logics on top of it.

Following action taken by Operations team:

• MF corrected for 171 distribution transformers.
• Total cost saving INR 3.94 Cr of 146 Distribution Transformers against Distribution Transformer swapping activity in FY 23-24.

To further increase asset utilization at lower cost, outage and infrastructure a flexible loading dashboard with geo spatial analytics has been developed as shown below which can also prescribe mitigation of load by extension of LT network only without even swapping the transformers.

Vegetation Management

Decision support system to analyse high tree tripping prone feeders, to predict best suitable time for tree trimming and generate notification for tree trimming. This helps utility in minimizing loss on account of outages and most effective maintenance output.

Best Time for Tree Trimming

Risk Based Matrix for Distribution Transformer

Distribution Transformer health index calculation based upon different parameters as Age, Loading, Power Quality factors (harmonics, % load unbalance and voltage profile etc. from Smart Meter) and distribution transformer result (BDV, IR Value) to improve equipment reliability, reduce distribution transformer failure rate and reduce operational expenditure.

Power Management

Day ahead and intraday power management in the most accurate way to include sudden weather impact and with minimal human intervention. Tata Power-DDL's in-house automated power management application is benchmarked against the external vendor and found to be at par in terms of prediction accuracy.

Revenue Protection

Indian power sector factor loses around $16.2 Billion on account of electricity theft every year and this makes revenue protection as highest priority for any power distribution company. Tata Power-DDL has managed this problem remarkably well and minimized AT&C losses to 5.9% by optimizing process with the help of AI. To accurately predict energy theft at consumer level, various attributes like consumer master data, meter data, reading data, billing data, smart meter interval data and event data, etc. are used to predict the percentage probability of theft. The teams can prioritize their activities accordingly.

Data visualization has been also used to identify historical pockets of theft which acts as another input for the teams to target traditionally pilfer prone areas and decide their course of action. A snapshot is given below for reference.

A dashboard to get a quick overview of consumer’s consumption pattern, historical theft history, and meter replacement and act accordingly during site visits. A snapshot is added below:

Payment Default Prediction

Indian Discoms are plagued by payment defaults by consumers. The amount is approx. INR 1,37,000 Crores pan India. To recover such dues further expenses are incurred. To avoid defaults and ensure timely and complete collection, a predictive model has been built so that pre emptive actions of follow up are initiated for prospective defaulters, The previous behaviours of consumers are used to segment them and customized recovery measures like soft calling, knocking etc. are planned and employed.

Cash Flow Prediction

Power sector being a regulatory industry and the model in Delhi being post paid for connections, the Finance teams needs a visibility of the cash flow so that they can plan the usage of the same and maintain a balanced scenario. The cash flow prediction model has been created to predict the incoming liquidity in advance blocks of 15 days.

Analytics on Cloud

“Analytics on cloud” is to provide highly matured in-house developed products embedded with unparalled domain expertise as a ready-to-use service with minimal customization in least possible timeframe. Leveraging organizational data analytics capabilities for generating additional revenue stream.

Smart Meter in Tata Power-DDL

Smart Meter installation has been scaled up year over the years. The communication medium predominantly through RF mesh, NB-IoT technology has also been introduced. Apart from the inherent advantages of smart meter in terms of remote reading, disconnection, event logs etc., the data available from smart meters have been leveraged to improve network and revenue parameters through analysis and analytics.

AMI Network Performance

Smart Meter Use Cases

Smart Meter Communication Analytics

Decision support system to find reason for smart meter non-communication and provide prescriptive solution to improve it further.

Consumer Sentiment Analysis

360° analysis of consumer voice at multiple digital platforms like Website, Facebook and Twitter etc. to provide actionable insight for further customer satisfaction improvement.