Due to the increasing penetration of distributed energy resources, congestion problems are already emerging in Dutch distribution grids. The available flexibility of assets in the built environment could have the potential to reduce congestion if prosumers are properly incentivized by distribution grid operators (DSOs). However, it is not yet clear what (combinations of) flexibility a ctivation mechanisms will be effective for congestion management in Dutch Distribution grids. To shed light on this issue, the GO-e consortium aims at performing large-scale agent-based simulations of up to 120 low-voltage networks and a large variety of possible instruments and scenarios. For this reason, we developed a novel scalable time-discrete simulation framework for distributed agent-based simulations of energy systems. We demonstrate the framework on a case-study in which we assess the effectiveness of a dynamic bandwidth tariff instrument on overloading problems in a low-voltage network containing solar panels, batteries, and heat pumps. It was shown that a dynamic bandwidth tariff can successfully resolve forecasted congestion if the associated costs are high enough compared to the day-ahead prices. However, the resulting load shifting can cause new congestion intra-day aswell.