# Power Systems and Markets * reliability is king * frequency is a global signal * nuclear can be a cause of negative LMPs * hydro in theory is flexible, but not in practice * heavily regulated (sometime forced to turn on, other times forced to turn off) * hydro in PNW heavily impacts prices in caiso * hydro could have massive negative LMPs * wind can influence negative lmps by producing *only* due to the production tax credits * wind power curves (25+ m/s will lead to a shutoff) * turbine specific * this can be looked up * NOAA models are bad at predicting clouds * Always planning on palo verde going down * DA market for energy * Think of this as the "copper sheet" scenario. The only *constraints* that are considered here are generator specific. We do *not* consider transmission line constraints/grid constraints at this stage. This is only used for the price of energy. * run many production level models at once to get an idea of: "what is the probability that palo verde goes out, and if it does what does it do to the prices?" * e.g. plexos (but slow and a piece of shit black box) * aurora (russ philbrich) * open source is a better option * powersystems.jl [GitHub - NREL-Sienna/PowerSimulations.jl: Julia for optimization simulation and modeling of PowerSystems. Part of the Scalable Integrated Infrastructure Planning Initiative at the National Renewable Energy Lab.](https://github.com/NREL-Sienna/PowerSimulations.jl/tree/main) * [Site Unreachable](https://www.softxjournal.com/article/S2352-7110(21)00076-5/fulltext) * Mixed integer linear programming: harder to solve because we cannot used derivatives. The integer natures comes into play because generators can either be on (1) or off (0). This means we cannot use derivatives to solve for the max / min here. ### Questions * how much information can we share across gen types? * how much do these things tend to change over time? * How would you learn about this? * if a generator has incentive to not turn on/off and there are high prices in an area, could we determine what gen would have the highest impact on a given node that has high prices? E.g. if gen 1 could supply needed power to node A, but it won't (say it's nuclear), can we deterministically figure out what node would have the next largest impact *that may have incentive* to turn on? * does a high LMP mean that no generator has turned on? * how could we actually determine if a generator has turned on? * If an LMP remains high then maybe nothing *can* turn on that will relieve the pressure * Way of modeling: * Model the state of the world/reality * Load * weather * Generation * weather * Yesterdays prices / DA prices * Current constraints (grid topology) * Model what different generator types will do *given* the grid state * certain generators will care about certain parts of the whole reality / features (hydro will not care about solar) * Given what generator types will do, what will the impact to prices be * Say I have a generator at a node A. I see that the rtpd is $50 at hour 18. * Does that mean that that was the price for a generator near that node to turn on? * How can I "back out" the price required for it to turn on? * will a generators action be reflected in the price? ### Features * How quickly / when has a generator turned on in the past? E.g. does nuclear gen 1 never turn on/off (they are price insensitive)? This could be a "responsiveness" feature * how much can we share information across different generator types? * Can we *infer* an outage if a generator isn't acting as you'd expect? * even if we could infer this correct 70% of the time that would be powerful * How can we determine the *threshold* at which a generator will turn on? * If we could determine *when*/*what* would cause generators to turn on, and we could then determine what the grid is going to do, we could place trades. * [Four Ways to Enhance Virtual Trading](https://blog.yesenergy.com/yeblog/four-ways-to-enhance-virtual-trading?utm_campaign=2Q2023%20Live%20Power&utm_content=249270881&utm_medium=social&utm_source=linkedin&hss_channel=lcp-1822120) * If generation turns on can it get to where it needs to go (i.e. an outage could prevent this...) * Outage could impact a gens shift factor on a node. In the past it could be high, but could now be low. E.g. shift factors could be * compare a system forecast (i.e. ercot weather forecast) with your custom forecast. If there is a large difference, expect that there may be a large price difference there. --- Date: 20230711 Links to: Tags: References: * []()