Energiewende in Deutschland: Investition in die Zukunft oder teurer Umweg? YouTube video by Forum Energiedialog

(3/3)
Das Video wurde hier veröffentlicht: www.youtube.com/watch?v=v2PJ...

Towards Efficient Multi-Objective Optimisation for Real-World Power Grid Topology Control Power grid operators face increasing difficulties in the control room as the increase in energy demand and the shift to renewable energy introduce new complexities in managing congestion and maintaini...

Great paper on #reinforcementlearning in power grid optimization by #enliteAI. This is AI thats actually generating a benefit for society. The advantages of #RL scaled to to real-world use cases.

arxiv.org/abs/2502.00034

China's electrification is pushing fossil fuels towards peak demand across key sectors ⚡
🏬Buildings: Electricity displaced coal with electric heaters
🏭Industry: Electricity overtook coal as largest energy source
🚗Transport: Early signs of oil decline as EVs accelerate

https://loom.ly/psjqvhE

Classic tit-for-tat between the US and Denmark with the added benefit of propping up struggling fossil firms

Is indonesias grid operator ready to connect all those turbines to the mainland grid? Who is going to finance these assets?

A Solid-State Battery Mercedes EQS Went 749 Miles On A Charge. It Still Had Range Left It’s all thanks to solid-state batteries that will gradually make their way into production EVs sooner rather than later

Mercedes is showcasing the potential of solid-state #batteries by delivering impressive range in an prototype EQS. But will these numbers transition onto real conditions and at what cost?

#emobility #solidstate

insideevs.com/news/771682/...

A pegada de carbono e as margens térmicas dependem da disponibilidade de GNL e gás, mas são menos dominantes mês a mês do que em secas anteriores.

#brasil #energia #ember

Espere spreads de pico mais apertados quando o vento tiver baixo desempenho e durante restrições de transmissão; suavização fora do pico quando a energia solar/eólica estiverem alinhadas.

O que isso significa para os preços e o despacho:

A energia hidrelétrica de reservatório está se tornando cada vez mais um “amortecedor” flexível, moldando os picos e preservando a água em períodos de seca.

Wind and solar generate over a third of Brazil’s electricity for the first month on record | Ember

Isso evidencia que o crescimento da VRE está amenizando a escassez hidrelétrica, em vez de aumentar o despacho térmico.

(ember-energy.org/latest-insig...)

Em agosto, a energia eólica e solar forneceram 34% (19 TWh) da energia elétrica consumida no Brasil, enquanto a hidrelétrica caiu para 48% (a segunda vez abaixo da metade) e os combustíveis fósseis ficaram em 14%, contra 26% em agosto de 2021, em condições de seca.

🗣️ “Brazil has established itself as a global clean power leader,” says Ember’s Dr Raul Miranda

China’s solar and wind generation grew 27% in the first half of 2025, contributing to a 2% decline in coal generation.

Combined with rapid electrification, China’s clean energy boom is rewiring the economy for deep decarbonisation

https://loom.ly/psjqvhE

Bankers urge Latin America to ease hurdles to attract renewable energy funds Latin American countries need to address regulatory barriers, improve electrical interconnection, and enhance renewable energy project structuring to attract the $200 billion annual investment required for the energy transition, multilateral bank executives said on Tuesday.

Bankers urge Latin America to ease hurdles to attract renewable energy funds reut.rs/4lHgx9N

Liebreich: The Pragmatic Climate Reset – Part II: A Provocation | BloombergNEF As the debate on transition deepens, Michael Liebreich details how a pragmatic climate and energy reset might unfold.

It's here folks. Part II of the Pragmatic Climate Reset. Prepare to be provoked! about.bnef.com/insights/cle...

ASSUME - Agent-based Simulation for Studying and Understanding Market Evolution Maurer et al., (2025). ASSUME - Agent-based Simulation for Studying and Understanding Market Evolution. Journal of Open Source Software, 10(112), 8515, https://doi.org/10.21105/joss.08515

Great toolbox for anyone who wants to model energy markets using #reinforcementlearning; by Mauerer et al (2025).
The use of agentic models is becoming more influential in energy systems. ASSUME enables the introduction of #RL and #AI into this ecosystem. #opensource.
joss.theoj.org/papers/10.21...

Kim K. Miskiw: Reinforcement Learning for Agent-based Electricity Market Simulations YouTube video by DTU: Lectures on Power & Energy Systems

Great paper on rl for #power market simulation. Goesto show that new #ai methods can revolutionize our understanding of markets.

www.youtube.com/watch?v=NBzO...

However the sparse digital infrastructure in the grid is often the greatest bottleneck! Will the #AI revolution in energy systems cut short by aging grids?

Check out their paper for the L2RPN challenge!

arxiv.org/pdf/2302.07654 (2)

Powergrid Congestion Management with Model-based Reinforcement Learning YouTube video by AAIC - Applied AI Conference

#enliteAI is a startup for #smartgrid redispatch and real-time control. Using RL at various steps of the process of managing a power network, they show a way to incorporate AI into actual control rooms of TSOs.

youtube.com/watch?v=0zqSAsj_86I (1)

a chart

a chart

a chart

a chart

The new @ember-energy.org report out today shows that both wind and solar are still growing - but that wind's growth has been decelerating for four straight years now.

IMO this is pretty worrying. At the system level and globally, wind and solar complement each other. A solar-only approach = worse

The authors admit more research must be done on inter-level dependencies and larger networks. Will we see #AI like this making #decisions on DSO level one day? (4)

The presented approach introduces a clever setup of hierarchial multi-agent RL. Agents with different policies each control sections of the grid. Rule based policies safeguard the grid, while low level agents are free to explore the smartest strategies (3)

The Problem:
Power grids are complex & dynamic, especially with a rise in res and storage. Optimizing their structure — like switching lines or buses — is a huge challenge due to the enormous action space involved. Traditional methods hit limits. (2)

Multi-Agent Reinforcement Learning for Power Grid Topology Optimization Recent challenges in operating power networks arise from increasing energy demands and unpredictable renewable sources like wind and solar. While reinforcement learning (RL) shows promise in managing ...

Cutting-edge Grid #AI: Multi-Agent RL for Power Topology Optimization
Van der Sar, Zocca and Bhulai explore how multi-agent reinforcement learning (MARL) can optimize power grid topology with high efficiency and scalability. (1)

arxiv.org/abs/2310.02605

#SmartGrid #ReinforcementLearning

Finanzieung der Energiewende: Wie viele Investitionen in Wind und Solar werden - Stand heute - über welchen Mechanismus finanziert? Prosuming aka Eigenverbrauch machen ein Drittel der Invest von rund 40 Mrd p.a. aus. Ungefähr gleichviel: Marktbasiert (offshore, PV-Freifläche).

Highlights!

> Although most papers present older algorithms efficiency across all problems improves by 10-20%
> There are open problems which could be solved by RL #sectorcoupling
> Promising research in the area of #multiagents and the intersection to local optimization (4)

The authors classify problems into
> Building Energy Management Systems
> Dispatch
> Vehicles and Devices
> Grid
> Markets

and analyse for complexity, #reproducability, computational needs (3)

They show that RL as a #datadriven method has the potential to deal with the insane complexity of large state and control variable spaces which appear all the time in energy systems. (2)

Great Paper about RL in Energy Systems!

Perera and Kamalarubans "Applications of #reinforcementlearning in energy systems" provides an broad overview as well as thorough analysis of the sota of RL in Energy Systems.

t.co/ae1rJzuEaq (1)