Enabling Intelligent Planning and Optimisation of Battery Energy Storage Systems to Maximise Grid Stability and Renewable Energy Integration 

What is the problem? 

As renewable energy penetration increases across India's grids, Battery Energy Storage Systems (BESS) are increasingly being deployed to manage intermittency and support grid stability. In Andhra Pradesh, utilities are planning to introduce BESS at multiple levels of the power system — at substations, at the distribution feeder level, and at renewable generation sites. However, most BESS installations operate on fixed charge-discharge schedules set at commissioning. These schedules do not adapt to real-time conditions — changes in renewable generation output, load fluctuations across feeders, or grid frequency signals. The result is that storage often charges or discharges at the wrong time, missing the moments where intervention would have the greatest impact on grid stability or cost. The full value of a BESS investment is therefore rarely realised in practice. 

Battery degradation compounds this problem. The long-term health of a BESS asset is shaped by how it is operated — cycle frequency, depth of discharge, temperature exposure, and rest periods all affect capacity retention over time. Without models that track these parameters together and recommend operational adjustments, batteries degrade faster than projected, replacement costs arrive sooner than planned, and utilities lose confidence in storage as a dependable grid asset. 

The need is for intelligent software that sits on top of existing and planned BESS infrastructure — optimising dispatch in real time, predicting degradation trajectories, and giving operators the visibility to manage storage assets across their full lifecycle.

Enabling Accurate Forecasting and Flexible Grid Operations to Manage Renewable Energy Variability Across Andhra Pradesh 

What is the problem? 

As solar and wind capacity grows across Andhra Pradesh, grid operators face increasing difficulty matching supply with demand in real time. Renewable generation is inherently variable — output changes with weather conditions, time of day, and season — and this variability creates scheduling and balancing challenges that require a new class of forecasting and flexibility tools working across the entire power system. 

On the generation side, short-term forecasting of solar and wind output remains limited. Without reliable predictions of when and how much renewable energy will be available, operators must make procurement and dispatch decisions with significant uncertainty. This leads to either over-reliance on expensive balancing power or the risk of supply shortfalls when renewable output drops unexpectedly. 

On the demand side, Andhra Pradesh's grid serves a diverse mix of consumer classes — including large agricultural, industrial, and residential loads — each with distinct and often poorly modelled demand patterns. This variation makes feeder-level demand difficult to predict reliably. Better demand characterisation at the feeder level would allow operators to plan dispatch more effectively and reduce the risk of asset overloads during peak demand periods. 

On the supply flexibility side, thermal generation units serve as the primary balancing resource when renewable output is low. However, the ability of thermal plants to ramp down and absorb renewable surpluses is constrained by technical minimum load requirements. Where these minimums are high, renewable energy may need to be curtailed even when it is available, reducing the overall efficiency of the grid's decarbonisation effort. Solutions that help thermal plants operate more flexibly — ramping down further and faster without equipment damage — would directly increase the grid's capacity to absorb renewable energy.

Enabling Intelligent Asset Management Across AP's Power System 

What is the problem? 

Across Andhra Pradesh's power system, unplanned equipment failures cause outages, strain maintenance crews, accelerate asset degradation, and reduce the overall reliability of electricity supply. The challenge exists across all three levels of the grid, but manifests differently at each. 

At the distribution level: The network operates a large number of assets — transformers, feeders, switchgear, and conductor infrastructure — across geographically dispersed service areas. Maintenance is largely reactive or scheduled at fixed intervals, regardless of actual asset condition. Field teams respond to failures after they occur, often without data on which assets are at highest risk. There is an opportunity to better utilise available data sources — including outage records, load history, and environmental conditions — to move towards condition-based maintenance. Physical inspection of distribution infrastructure across long stretches of line also remains time-consuming, resource-intensive, and difficult in remote terrain. 

At the transmission level: High-value assets such as power transformers, reactors, and switchgear are critical nodes whose failure causes large-scale supply disruptions across multiple parts of the network simultaneously. Transmission lines spanning long distances are exposed to conductor sag, vegetation encroachment, hotspot conditions, and physical damage that are difficult to detect through conventional patrolling. There is also an opportunity to move beyond conventional dispatch towards more dynamic network management — one that accounts for real-time conductor ratings, line capacity, and congestion patterns to improve utilisation across the transmission network. 

At the generation level: Thermal plants face growing maintenance complexity as units age and operational demands increase. Forced outages disrupt the grid's supply balance, particularly when renewable generation is low. Overhaul planning based on fixed schedules rather than actual equipment condition can lead to either premature interventions or missed warning signs. There is scope for AI and data-driven tools to support more intelligent maintenance scheduling, performance monitoring, and operational decision-making at the plant level. 

Together, the deployment of intelligent asset management tools across these three levels creates the conditions for a step change in how power system operations are managed — moving towards unified, real-time visibility across the asset base and operator decision support that is proactive rather than reactive. This is the foundation on which smarter, more integrated control room operations can eventually be built.