Managing Peak MW Levels in Power Grids

Little game for my kids, and everyone else. Comparing two electricity markets, one with a price peak in the evening, the other without. Which color, or technology is missing in the power mix in Germany? Earlier this week we have seen an arrival of a kind of new phenomena in our energy system: peak power prices in the summer. Driven by increasing AC-loads and climate change related lower capacity factors of European nuclear and fossil power fleets, prices peak in the evening. You can find a nice write down of this by Matthias Janssen here: https://lnkd.in/ehymPvCK Before somebody tells us, that this is a clear sign that we need to build (let's pick a random number) 20 GW of gas peakers as soon as possible (to at least a Schnellboot) ... or bring back coal plants from the grid reserve ... let's ask ourselves, how are other markets dealing with this. A peak price period of 4 hours, plannable, and right after a day with plenty of renewables at negative power prices. Wouldn't it be great, if we could 😵 STORE electricity? The comparison with CAISO (California) painfully shows what is missing in the German (and European) power mix. In CAISO, on June 30th at a load of 31 GW, BESS delivered 10GW of peaking power, keeping gas peakers offline. The power prices in the evening never went above $100/MWh. In Germany without BESS to dispatch at scale gas peakers ramped up, burning expensive fuel and putting their start up cost for only limited hours of operation into the market bid, resulting in a power price of 476€/MWh. In fairness, around 5.4 GW of pumped hydro plants were dispatched at the same time, mirroring what a large BESS fleet would do during this time. Clear and simple message: If you want to prevent high peak power prices, allow flexibility into your power mix. BESS is the only technology class that is being build in Germany without subsidies or state-supported revenue guarantees. But policy makers, regulators and grid operators on many fronts try to slow down the single success story of the German power market (from an investment site). Some just want to build gas plants, others are overwhelmed with lacking digitalization of their grids and the regulator just wants to squeeze as much money as possible from the business case of batteries. But bottom line, the CAISO example shows us, what our future power mix should look like, and that BESS are the key tool to provide peaking power and reduce scarcity pricing in power markets.

⚡ Technical Engineering Insight | Utility Scale BESS (20 MW / 40 MWh) Developed a detailed technical study and engineering overview for a 20 MW / 40 MWh Battery Energy Storage System (BESS) covering complete SLD, CAPEX & OPEX architecture aligned with modern grid integration requirements. 🔹 System Configuration: • 20 MW / 40 MWh (0.5C Configuration) • Grid Connected at 33 kV Level • Utility Scale Lithium-Ion BESS Architecture • Integrated EMS / SCADA Monitoring & Control 🔹 Major Technical Components: ✅ Battery Racks & Battery Management System (BMS) ✅ Power Conversion System (PCS) – Bidirectional Inverter ✅ 0.69/33 kV Step-Up Transformer (ONAN/ONAF) ✅ 33 kV Switchgear, CT/PT & Protection Relay ✅ Fire Detection & Suppression System ✅ HVAC Based Thermal Management ✅ Grid Synchronization & Dynamic Response Control 🔹 Engineering Scope Covered: ⚡ Single Line Diagram (SLD) Development ⚡ AC/DC System Integration Philosophy ⚡ Protection Coordination & Interlocking ⚡ Auxiliary Power Requirement Analysis ⚡ EMS-PCS-BMS Communication Logic ⚡ CAPEX Distribution & Lifecycle OPEX Estimation ⚡ Battery Safety & Thermal Runaway Mitigation ⚡ Grid Code Compliance & Ancillary Service Readiness 🔹 Estimated Financial Overview: • CAPEX: ~₹80–120 Cr • OPEX: ~₹2.5–4 Cr/year • OPEX ≈ 2–4% of Total CAPEX 🔹 Grid Support Applications: ✔ Peak Shaving ✔ Frequency Regulation ✔ Voltage Support ✔ Renewable Smoothing ✔ Black Start Capability ✔ Reactive Power Compensation ✔ Ancillary Services Participation The future of modern power systems will strongly depend on intelligent integration of BESS with Renewable Energy and Smart Grid infrastructure for ensuring stability, flexibility and decarbonization of the grid. Prepared By: Kushlesh Pandey Engineer – BESS & Renewable Energy #BESS #BatteryEnergyStorageSystem #EnergyStorage #UtilityScaleBESS #RenewableEnergy #SmartGrid #GridStability #AncillaryServices #SCADA #EMS #BMS #PCS #PowerSystem #ElectricalEngineering #Substation #HVEngineering #EHV #GridModernization #Transformer #Switchgear #ProtectionSystem #BatteryTechnology #LithiumIon #RenewableIntegration #CleanEnergy #PowerGrid #SolarEnergy #WindEnergy #EnergyTransition #GridCode #ElectricalInfrastructure

This is a typical 24-hour load curve of LESCO, and it perfectly captures how high solar penetration is reshaping our grid dynamics. During the midday hours, the system demand dips sharply, not because consumption has fallen, but because rooftop and distributed solar generation are offsetting grid demand. Many consumers meet a significant share of their own needs during daylight hours. However, as the sun sets, this demand rebounds rapidly, creating a steep evening ramp of nearly 1,000 MW that the grid must supply within just a few hours. This phenomenon, known globally as the “Duck Curve,” is now clearly visible in Pakistan’s power sector and is becoming more pronounced in LESCO. While solar is helping reduce daytime demand and emissions, it is also creating new operational challenges such as steep evening ramps, lower system inertia, and potential overvoltage in solar-dense feeders. Moving forward, utilities like LESCO will need to adopt flexible solutions such as battery energy storage systems, demand response programs, time-of-use tariffs, and smarter grid management tools to maintain stability as Pakistan transitions toward higher levels of clean energy integration.

Our grid is built to manage the marginal day when demand is highest. That means there is plenty of room on the non-marginal days for additional generation to be made use of (Data Centers?). In the graphic below, explore this a little with 2024 data - generation data from https://lnkd.in/e9DcWKk and capacity data from https://lnkd.in/gaxU2JAd. Top graphic - the line is the hourly demand (load) in GW for the contiguous US 48. The shading from bottom to top is the coal fleet (black), combined cycle fleet (light brown), nuclear fleet (green), gas boiler - steam turbine fleet (darker brown), and gas turbine fleet (orangish) nameplate capacity. Roughly about 850 GW. That ignores our big solar, wind, and hydro fleet, and also ignores that while the nameplate on those dispatchable coal, gas, and nuclear assets may be 850 GW, with reliability and maintenance windows, the actual available capacity is not likely 850 GW across the year. So the middle graphic takes the load (line) from 2024, and subtracts the hydro, solar, and wind generation on each hour to give a net load as the new blue line. I.e. the load less what we got from supply from water, solar, and wind in 2024 in the contemporaneous hour. It also pro-rates the gas, coal, and nuclear capacity, assuming we get 90% from the CCGT, coal, gas boiler steam turbines, and nuclear in summer and winter, and 70% from the gas turbines in summer. In the off season for demand it drops these to 70% and 50% to allow for annual maintenance intervals. The bottom graphic then takes a look at how much room we have between the net load and the available capacity for every hour of the year, and lines it up from most room to least. The tightest pinch-point is over 50GW, if the data centers we built provided demand response or had backup power for the 150 tightest hours of the year this window opens up to 120GW. That is more than the peak demand for electricity in all of Germany (about 150% the peak). Which is to say if data centers can offer a little flex, we have a lot of room to scale. DCFlex is a large EPRI Initiative exploring the technical details of how this can be done at a grand scale. https://dcflex.epri.com/ The initiative has its first big annual meeting this October in Florida! https://lnkd.in/eivCnD3P

The power grid is strained, but not around the clock. The real pressure shows up during peak days, especially during summer heat waves. In Northern Virginia, most data center facilities are colocation data centers with annual PUE values around 1.2 to 1.5 and even higher real-time PUE during the summer days. Because they host many independent tenants who operate their own servers, these colocation facilities have limited ability to use computational load shifting for load shedding. As a result, their main option for load shedding today remains onsite diesel generators, which emit large amounts of NOx and pose health risks to nearby communities. This raises a natural question: what if we revisit some old technologies such as thermal energy storage and shifting? For data centers that rely on chilled-water cooling, operators can pre-chill water and store it in insulated tanks, then draw from that reserve during peak hours without running chillers. For facilities that do not use chilled water, server heat can be captured to warm cold water stored in a tank, reducing the need for energy-intensive cooling to reject heat outdoors during peak periods. What is the potential impact? I estimate that this type of thermal energy management, using something on the order of 25 Olympic-sized swimming pools of water (and icepacks would be even better lol), could provide gigawatt-scale grid relief on peak days. When combined with other load-shedding strategies, the benefits can be even greater! No computational flexibility? No problem! Thermal energy storage and shifting is the way to go!

🔋BATTERIES BATTERIES BATTERIES!!! ⚡️ California is rightfully busy building a massive amount of energy storage, and the results are nothing short of a grid revolution. If you’ve been watching the California grid lately, you’ve seen the shift. Batteries are no longer just a science experiment—they are the new backbone of our evening peak. Why the hype? Here are the facts: 📉 Zero-Lag Response: Batteries replace expensive natural gas peaker plants because they react instantly. 📉 Frequency Control: They have dominated the Ancillary Services Market. They are so effective at maintaining grid frequency (60 Hz) that Regulation Up/Down prices have fallen off a cliff lately, decoupling entirely from gas prices. 📈 Massive Growth: We’ve gone from ~500 MW in 2020 to over 13,000 MW of installed capacity today. That is a 25x increase in under 5 years! 📈 Record Breaking: During the 2024 heatwaves, batteries routinely discharged over 8 GW back into the grid—meeting nearly 25% of the state's total load during critical windows. 👀 The Visual Proof (See Attached Maps): I have been working hard to update the database behind the California Power Grid Map to reflect this breakneck build-out. My latest update takes us from a mere 13 facilities to 115 active battery plants (~15GW total capacity tracked!). Check out the difference in the two screenshots below: 1️⃣ 1:00 PM (Solar Peak): The battery symbols are Turquoise. They are soaking up that abundant, cheap solar energy. 2️⃣ 7:00 PM (Net Peak): The sun is down, solar is offline, but look at the batteries now. The symbols turn Bright Yellow as they discharge massive power to keep the lights on. It won't be long before we see them move deeper into long-duration territory. We are building at breakneck speed, and so we should. Play with the load shape yourself to see how the grid reacts: https://lnkd.in/gBsN3jva #EnergyTransition #BatteryStorage #California #Renewables #GridModernization #DataViz

🔋One of the most interesting results of the DC Industrie-2 project was demonstrating a reduction in the peak power rating of the grid-connecting active front end (AFE). The figure shows the demonstration “production line” with two robots, two machines, AFE, and a battery energy system (BESS).   🍃The comparison of the power fed from the AC grid with and without the BESS system on the DC bus shows the peak power reduction. Reducing the peak power rating of the grid connection eases 1️⃣ the strain on the AC grid, reduces 2️⃣ the initial investment (lowering grid charges), and reduces 3️⃣ the operational costs as it increases self-consumption.   💡The droop control ensures load sharing, and the figure also includes the droop curves. In the first zone, both AFE and BESS supply power to the load. However, in the second zone (marked with grey borderline), AFE supplies only the base load, and BESS supplies the peak power. The direct connection of the battery to the same bus as robots and machines also increases recuperation in Zone 5. The last zone is reserved for the option when the BESS system is fully charged and energy needs to be returned to the grid.   🎯DC distribution systems are protected by solid-state circuit breakers, which reduce the transferred fault energy and peak fault current several times compared to using fuses. Since operating efficiency is crucial, Infinitus delivers a 0.9 V drop at 1000 A.  https://lnkd.in/dZjf3dcW #directcurrent #solidstatecircuitbreaker #renewable #cleanenergy #bess #energystorage #gridmodernization

Demand Response: A Key to Reducing Peak Load Pressure on Utilities Demand response is not a new topic and keeps getting rebranded as something else every few years. The basic idea behind Demand Response is that the utility, when needed, sheds load that volunteered into the program when it makes financial sense or for reliability reasons. Usually, it is more the former than the latter. Shedding load is usually the last-ditch thing that a utility wants to do, but if someone is part of a Demand Response program, they are volunteering to be a front-line solution. What are some reasons for utilities to want Demand Response load? The first is just that as generation is dispatched, it generally is dispatched from the cheapest megawatt to the most expensive megawatt. This can be related to the efficiency of a unit. For example, an old coal plant might not be as efficient as newer plants at turning water into steam, and that might be a reason it is dispatched only during peaks. Often, peaker generation is not just old units but gas turbine generators. These units are in the 60-80 MW range. These units are half as efficient as combined cycle gas generation plants, which are in the 1000 MW range but are built explicitly to serve peaking loads with having a lot of unused capacity. The cost for peaking generation is very expensive, not only due to units being dispatched from cheapest to most expensive megawatt but mainly due to low utilization rates. There are fixed costs associated with just having and maintaining generation. On base units, units that run 90%+ of the time, these fixed costs are spread out over a large number of hours so fuel cost tends to be more of a driver for cost per MWh. With peaker units, their utilization might be seasonal or just a few weeks a year. This drives up run costs because fixed expenses are spread over only a few hours a year. To give you an idea of the difference, inside ERCOT, it is normal for electrical prices to be around $20-30 per MWh. During system peaks, electrical prices can spike above $1000 per MWh. These high prices are necessary to pay for the peaker units and incentivize people to build peaker generation. Inside of ERCOT, there has been a problem with getting generator companies to build additional peaker generation and the market cap kept getting raised to $9,000 per MWh to try to incentivize people to build more generation. It was lowered after the 2021 Texas Blackout to $5,000 per MWh. The cap is in place to prevent market manipulation as was done by Enron. The second reason why utilities like demand response load is for reliability reasons. Occasionally, the grid starts running short on capacity, usually on a hot summer day, and shedding load would help keep the grid from potentially having brownouts. Market prices for unregulated regions would spike, too, during these shortages, so it often isn't just the only reason. #electricalengineering #utilities #renewables #energystorage #substations

The Netherlands just unlocked 9GW of grid capacity without even building new lines. They’re using it to connect record levels of battery storage. 🇳🇱 Around a third of Dutch homes have rooftop solar, offshore wind will be the biggest source of energy by 2030, and the country has the highest penetration of EV chargers in Europe. It also has one of the most congested grids. While it is clear that our grids need to be modernised and expanded to integrate renewables, this can take years. Years that we don’t have, as new renewables, batteries, heat pumps are struggling to get connected. To better manage the grid, the Dutch TSO TenneT introduced “off-peak” flexible connection contracts. A user, such as a solar farm, would only have full access the grid 85% of the time. During peak periods for the grid, the TSO can partially or fully limit use. The TSO calculates that 9GW of capacity is available during off-peak hours, and is awarding 6GW to battery storage projects, which themselves can help better manage congestion further. The grid is not just about build, build, build, and the Netherlands shows it. We need countries to enact Rapid Capacity Plans, a toolbox of measures to unlock capacity today, while buying the time needed to expand the grid.