Home

Welcome to the SolOptXMR wiki!

TODO

Minimum

• all jsons under the local config directory
• weather: predict 3 days ahead
• proper solar readout
• UIs: console + curses + Qt
• temperature safety - use the script at startup
• present multiple solutions in a HTML file (Partly done: using console)
• document rectangle recognition
• Docker images
• Conda environment & virtualenv
• Windows port
• OSX fixes
• Rpi CI
• minimize Python dependencies, based on User choices

Major

• measurements
• interface for custom energy sources like wind
• automated download of mining software
• scheduling for further days requires setting these days for the at command as well (+ tests of course)
• translate measured voltage into battery charge in Ah
• model charge voltage to State of Charge
• allow for a custom energy input plugin
• model voltage loss from attaching consumers
• multiply the voltage of the voltage functions by the multiple of (system.json voltage / 12V)
• 2 phased optimization: first see which setups treat the battery well, then from these select those which maximize profits
• Support more PCs by improving the combination algorithm (currently up to 2 PCs are supported)
• automatically switching on & off the PCs. Needs documentation & scripts
• hashrate predictions into model
• OSX support. Issue link

Testing

• prep. testing scheme for the end results without having to do heavy calculations
• simulate some found solutions to test the production path in a fast way
• the search algo gets confused if the initial conditions show battery overcharge

Known bugs

• setting wind turbine's count to 0 leads to crashes after calling the opti program from Python
• opti app writes to /tmp, rather than to the ConfSol.m_outDir. It might be related to the OSX bug.

Minor

• type hints
• doc: how to enable MSR module
• doc: cable sizes vs max current
• doc: fuses on battery, cables (industrial), mppt's radiator, that can overheat and inverter's grounding. Mention synthetic load
• doc: measure and document battery voltage vs. temperature, esp. across the night.
• doc: grounding https://www.solaris-shop.com/blog/grounding-techniques-to-prevent-inverter-damage/
• 3 line plots by stacking 2 blocks on top of one another
• arbitrary number of line plots by stacking N blocks on top of one another
• automatically scramble the geo location

Lifepo4 bats

• https://footprinthero.com/lifepo4-battery-voltage-charts
• https://pushevs.com/2018/01/29/2018-nissan-leaf-battery-real-specs/
• https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries
• https://www.rebel-cell.com/knowledge-base/the-discharge-and-capacity-of-batteries/

Future work

• compare historical results against each other across time, to measure the discrepancy
• scheduled energy production with crank, similarly to habits
• detailed cloud forecast 1 2 3
• cloud forecast abstraction & fallbacks
• leverage Boost Units to avoid potential conversion errors.
• synthetically increase the load during days of rapidly alternating clouded/sunny sky to reduce the shock on the inverter
• predict and evaluate future network difficulty fluctuations. Right now we're just displaying the current situation.
• more work on optimization algorithm: fuzzification between the bands
• algo: use decision tree, or genetic algo, rather than a fully random choosing
• flywheels & hydro storage
• choose habits via console and their text ID
• habits need to be able to be declared as optional, and ran only when there's enough power
• use the installation scripts to setup i2p network automatically