gschivley/battery_model

Sample project modeling battery storage and dispatch

battery_model's Issues

Fix SOC in last hour for full year optimization

I'm starting SOC at half of maximum. Finish as same in last hour to avoid model discharging all power? The max discharge in 24-hour constraint might prevent this though - check results.

Explore lmp data

Nothing too fancy. Maybe a heat map of mean/median price by hour/week?

Explore results

Similar to #7 but with charge/discharge and profit

Multiple-period optimization

Once #2 is done, wrap the code in a function. Allow for parameter inputs (length of look-ahead, length of schedule to keep). Output charge state, charge/discharge data. Use outputs to ensure charge continuity and build record of activity of the full year.

To keep the 24-hour discharge constraint, each optimization will probably need to know the proceeding 23-hour discharge schedule. I suppose this could be done by passing in the entire discharge schedule up to that point and using the values as a constraint.

Potentially incorrect operator in the Positive_charge Function

Background
The positive_charge function in the read battery_model.py file appears to constrain Eout when the S < Rmax.

Current Behavior
Currently the battery is exporting more power than it can hold. for example; When S = 100, RMax = 250 and eta = 0.90, Eout currently equals 105

Expected Behavior
the desired output should be as follows; When S = 100, RMax = 250 and eta = 0.90, Eout should equal 95

Possible Solution
In the positive_charge function switch the operator from divide to multiply

Aggregate results

Need to know most profitable month and plot dispatch/LBMP for the most profitable week. Aggregate and do calculations using pandas and the timestamp from original data.

Use integer index rather than datetime

I forgot about having to deal with daylight savings and don't think it's worth the bother to convert to UTC for this exercise. Just use 8760 instead.

Solve optimization for whole year at once

Start with the easiest method by solving for the entire year.

Write out parameters

All parameters/constraints/variables and the objective. Might as well get in equations to keep as a nice reference when writing code.