1 Battery Model

A preliminary battery model is built by scaling the PSE battery model from a rated power and rated energy of 2000 kW/4400 kWh to the 1000 kW/4000 kWh rating of the Powin battery. We will update this as we collect data from EnergyNW.

The battery rate of SOC change as a function of SOC and Power is shown below.

2 Use Case 1 - Demand Charge Reduction

2.1 Data

We do not have this data yet. We require the formula for determining the cost of the monthly peak. This will update when we have the data. For preliminary cycle, we can generate assuming the price data is redundant.

2.2 Methodology

Discharge at rated power during predicted monthly peak. From Richland Load data, we know the peak four hours are hour 17 to hour 20. Hence we will begin discharge at hour 17 and provide rated power down to 20% SOC. Then we will charge back up after midnight.

2.3 Resulting Cycle

Below is the resulting cycle.

3 Use Case 2 - Load Shaping Reduction

3.1 Data

We do not have this data yet. We require hourly price data from the city of Richland. This will update when we have the data. For preliminary cycle, we can generate assuming the price data is redundant.

3.2 Methodology

Discharge at rated power during predicted maximum price. From Richland Load data, we know the peak four hours are hour 17 to hour 20. Hence we will begin discharge at hour 17 and provide rated power down to 20% SOC. Then we will charge back up after midnight.

3.3 Resulting Cycle

4 Use Case 3 - Transmission Charge Reduction

4.1 Data

Obtain transmission load data from BPA - data has resolution of 5 minutes. Arbitrarily select the range of 2021-04-22 06:00 to 2021-04-23 06:00 as the day as basis for producing a duty cycle.

4.2 Methodology

Discharge at rated power during predicted maximum transmission load. From BPA transmission Load data, we know the peak four hours are hour 7 to hour 10. Hence we will begin discharge at hour 7 and provide rated power down to 20% SOC. Then we will charge back up after midnight.

5 Use Case 4 - Volt Var

5.1 Data

For preliminary duty cycle, we use a reference for voltage vs time. We may later be able to collect this data from ENW. The voltage vs time is given below:

5.2 Methodology

We control the voltage by providing negative VAR when voltage is high, and positive VAR when voltage is low. This is done via a piecewise linear function that provides no VAR when voltage is between 0.99 to 1.01 of reference voltage. It maxes out at full rated negative VAR at voltage of 0.94 of reference voltage, and full rated positive VAR at 1.06 of reference voltage. This relationship is shown in the figure below:

5.3 Resulting Cycle

The resulting cycle is given in the figure below.

6 Use Case 5 - Outage Mitigation

6.1 Data

We already have load data from ENW for the city of Richland hourly load. Then scale it so the average load is (1.2 avg kW/house)*(113 houses) = 136 kW. The load is given in the graph below. Arbitrarily load in the date of 2018-07-23, which has the greatest standard deviation within the day. We’ll go from 6am on this date to 6am the next day.

6.2 Methodology

We will subject the battery to discharge during the 8 peak hours, for the worst case scenario. The power to discharge will be equal to the load. After midnight, charge back up at rated power.

6.3 Resulting Cycle

The resulting cycle is given below.

This cycle has an average discharge power of 220kW for 8 hours, and discharges to 42% SOC.

Below is the same cycle, but just the discharge part.