Systems and methods for managing diverse batteries
Systems and methods are disclosed for controlling packs of rechargeable batteries by: for each pack: determining electrical characteristics of the pack; determining energy correction for the pack; calculating a voltage correction for the pack; and determining a dispatch modifier for the pack. The system then normalizes the dispatch modifiers for all packs and recalculates the dispatch modifiers for all packs.
1. A method for controlling packs of rechargeable batteries, comprising:
sensing a voltage for each of the packs:
determining if the voltage is within a predetermined range of each of the packs;
determining a charge required to deliver a dispatch;
determining a slope of a discharge curve from the charge using an existing or a learned V-Ah model or discharge curve;
determining a voltage drop for the charge from the existing or learned V-Ah model or discharge curve;
normalizing the voltage drop to a dimensionless voltage drop by dividing it by maximum allowable voltage;
determining a modifier M based on required dispatch, the dimensionless voltage drop (ΔV i )* and dimensionless voltage V i * as M=D i ((ΔV i )*)(1−V i *);
determining a temperature deviation of each of the packs from a mean and normalizing the temperature deviation using a maximum allowable deviation;
utilizing a normalized temperature deviation in a gaussian distribution to calculate a temperature modifier (M T,i ); and
determining a battery dispatch signal {circumflex over (D)} i to discharge a battery using {circumflex over (D)} i =D i +M(V i ,Ah i )
where D i is a discharge power, V is the voltage, and Ah i is ampere-hour for pack i, wherein the modifier is used in battery discharge control.
2. The method of claim 1 , further comprising providing an energy correction with power draw or ampere hour correction.
3. The method of claim 1 , further comprising determining temperature correction for each of the packs.
4. The method of claim 1 , further comprising applying a neural network, curve fitting, or learning machine to determine a discharge slope for at least one pack.
5. The method of claim 4 , wherein data from time series is converted to a time invariant array comprising of temperature, dispatch and voltage and the neural network or learning machine is trained to generate the discharge slope as output.
6. The method of claim 1 , further comprising applying an auto regression technique to estimate slope.
7. The method of claim 6 , further comprising determining:
Ŝ ι =C+φ 1 S i-1 + . . . +φ P S i-p −θ 1 ε i-1 − . . . −θ q ε i-q
where S denotes the slope, while ∈, θ and φ denote the error and coefficients, respectively and where C is a constant, i is a time index, and p/q are event horizons for propagation of regression and its error.
8. The method of claim 1 , further comprising estimating dimensionless voltage by measuring a voltage and determining a maximum and minimum voltage.
9. A method for controlling packs of rechargeable batteries, comprising:
sensing a current or a voltage for each of the packs:
determining a voltage of each of the packs;
determining a discharge power and normalizing the voltage to a dimensionless voltage by dividing the voltage by a reference voltage;
determining a slope of a discharge curve from the voltage and the discharge curve and determining a drop in voltage from the slope and calculating a voltage correction therefrom; and
determining a modifier for a dispatch signal and generating a battery dispatch signal D i to control a battery discharge using {circumflex over (D)} i =D i +M(V i ,Ah i )
{circumflex over (D)} i =D i +M ( V i ,Ah i )
where D i is a discharge power, M is the modifier, V is the voltage, and Ah i is ampere-hour for pack i, and wherein the modifier is used in battery discharge control.
10. The method of claim 9 , wherein the sensing comprises sensing voltage or current trends.
11. The method of claim 9 , further comprising providing energy correction as power draw or ampere hour correction.
12. The method of claim 9 , further determining temperature correction for each of the packs considering a statistical distribution.
13. The method of claim 9 , further comprising determining training data.
14. The method of claim 13 , further comprising performing off-line training by obtaining historical battery data and training a model therewith.
15. The method of claim 13 , further comprising performing on-line training by measuring parameters from battery data, obtaining a base model, and training the base model.
16. The method of claim 1 , further comprising estimating voltage by measuring a voltage and determining a maximum and minimum voltage.
17. A system, comprising:
one or more packs of rechargeable batteries; and
a controller coupled to the one or more packs for controlling the packs of rechargeable batteries, wherein for each of the packs the controller executes code for:
sensing a current or voltage for each of the packs;
determining a voltage of the battery pack;
determining a discharge power and normalizing the voltage to a dimensionless voltage by dividing the voltage by the reference voltage;
determining a slope of a discharge curve from the voltage and the discharge curve and determining a drop in voltage from the slope and calculating a voltage correction therefrom; and
determining a modifier for a dispatch signal and generating a battery dispatch signal {circumflex over (D)} i to control a battery discharge using
{circumflex over (D)} i =D i +M ( V i ,Ah i )
where D i is a discharge power, M is the modifier, V is the voltage, and Ah i is ampere-hour for pack i, and wherein the modifier is used in battery discharge control.
18. The system of claim 17 , wherein the controller manages temperature correction for each of the packs considering a statistical distribution.
19. The system of claim 17 , further comprising a neural network, a learning machine, curve-fitting method, or an auto regression method to determine a differential change in discharge for each of the packs.