FAQs

Terminologies

Terminologies	Explanations
cbc	Choice-based Conjoint
`cbcTools`	An R package developed by Professor John Helveston for cbc survey designs
logit	The natural logarithm of odds
`logitr`	An R package developed by Professor John Helveston estimating MNL and MXL models with preference space and WTP
MNL	Multinomial Logit
MXL	Mixed Logit
WTP	Willingness to Pay
DCM	Discrete Choice Modeling
MLE	Maximum Likelihood Estimation
PEV	Plug-in Electric Vehicles, includes both BEV and PHEV
BEV	Battery Electric Vehicles
PHEV	Plug-in Hybrid Electric Vehicles
CV	Conventional Vehicles
GHG	Greenhouse Gas
V2G	Vehicle to Grid
SOC	State of Charge (appears as % of battery)
HiGRID	The Holistic Grid Resource Integration and Deployment model, developed using Matlab at UC Irvine by Eichman et al. (2013)
GOOD	The Grid Optimized Operation Dispatch model, developed at UC Davis to capture effects of PEV adoption an integration onto the electric grid

Research Thrust 1

Design the conjoint questionnaire using the cbcTools package and use DOE (Design of Experiment) process to design the survey. The survey questions should at least contain compensation and the 3 controlled charging attributes (time window, proportion, type), and then some additional questions about customer preferences.
Implement the survey using formr.org.
The survey should be in 3 stages:
- The “think aloud” pilot survey - among a small amount of BEV owners;
- Pilot survey - using Amazon Mechanical Turk;
- Final survey - using platforms like Dynata or Qualtrics; via partner coordination; and in online BEV forums.
I found 2 versions for “preliminary survey designs”:
- p.36 - 2500 completed respondents in batches of 500, out of which we need 2000 final samples.
- p.12 - 4000 completed respondents in batches of 500, out of which we need 3000 final samples.
With the results of survey, we use MLE (Maximum Likelihood Estimation) to estimate the parameters of DCM (Discrete Choice Modeling) and generate utility functions with both deterministic and random components, out of which we construct the logit model. This requires the use of the logitr package.
The marginal utility will be converted to dollars, which is a direct measure of values from each feature.
Use the model to conduct simulations to predict the best smart charging programs.
Integrate with Research Thrust 2…
In summary, our R packages, platforms, and math are:
- R packages: cbcTools, logitr,
- Platforms: formr, survey platforms, BEV forums,
- Math: MLE, DCM, logit.

Q & A

Question 1: I know the meaning of “utility” in the context of DCM (i.e. utility maximization, or marginal utility), but in the proposal, “utility” can also refer to the “provider” or the “grid”, right?

Answer 1: Yes “utility” refers to the grid electricity provider. It is also used in the context of DCM.

Question 2: Does “compensation” refer to the supplement to the smart charging implementation in a lager scale, or the bonus awards to the participants (the BEV owners), or both?

Answer 2: “Compensation” refers to some sort of incentive to participate in smart charging.

Question 3: What is the key difference between “high BEV adoption” and “high renewable energy adoption”, in the 2 case studies in California? I know BEV is on the demand side, and renewable energy is clean energy supply (since we also consider low carbon emission), but what roles do they play differently in our study?

Answer 3: “High BEV adoption” means a lot of people buying BEVs and “high renewable energy adoption” means grids that have a lot of renewable power, like solar and wind. CA has a lot of both. If we have a lot of BEVs enrolled in smart charging, they can be used to help integrate the renewable energy more smoothly.

Question 4: There will be simulations required for our prediction process (pp.38-39 Building predictive models). How do we implement the simulations? Are there specific tools and processes? Or do we follow the other team in Thrust 2?

Answer 4: We’ll get to simulations later. The point of them is to predict how many BEV owners would enroll in the smart charging program and compute the benefits on the grid side in terms of energy savings and renewable energy added to the grid.

Question 5: I read through our Gantt Chart. I believe the green colored tasks are mine. How about the orange color? Does it mean utility providers or collaboration with our Co-PI and the other student?

Answer 5: We’re focused on the conjoint part for now (year 1). My colleague at UC Irvine will be leading the energy modeling, and he’ll be working this year to put together some basic energy models that do not include smart charging preferences. Then later after our survey is done we’ll add that part in.

References

Eichman, Joshua D., Fabian Mueller, Brian Tarroja, Lori Smith Schell, and Scott Samuelsen. 2013. “Exploration of the Integration of Renewable Resources into California’s Electric System Using the Holistic Grid Resource Integration and Deployment (HiGRID) Tool.” Energy 50 (February): 353–63. https://doi.org/10.1016/j.energy.2012.11.024.

--- title: "FAQs" bibliography: references.bib --- ## Terminologies | Terminologies | Explanations | |:-----------------:|-----------------------------------------------------| | cbc | Choice-based Conjoint | | [`cbcTools`](https://jhelvy.github.io/cbcTools/) | An R package developed by Professor John Helveston for cbc survey designs | | logit | The natural logarithm of odds | | [`logitr`](https://jhelvy.github.io/logitr/) | An R package developed by Professor John Helveston estimating MNL and MXL models with preference space and WTP | | MNL | Multinomial Logit | | MXL | Mixed Logit | | WTP | Willingness to Pay | | [DCM](https://eml.berkeley.edu/books/choice2.html) | Discrete Choice Modeling | | MLE | Maximum Likelihood Estimation | | [PEV](https://www.transportation.gov/rural/ev/toolkit/ev-basics/vehicle-types) | Plug-in Electric Vehicles, includes both BEV and PHEV | | BEV | Battery Electric Vehicles | | PHEV | Plug-in Hybrid Electric Vehicles | | CV | Conventional Vehicles | | GHG | Greenhouse Gas | | V2G | Vehicle to Grid | | SOC | State of Charge (appears as % of battery) | | HiGRID | The *Holistic Grid Resource Integration and Deployment* model, developed using Matlab at UC Irvine by @eichman2013 | | GOOD | The *Grid Optimized Operation Dispatch* model, developed at UC Davis to capture effects of PEV adoption an integration onto the electric grid | ## Research Thrust 1 1. Design the conjoint questionnaire using the `cbcTools` package and use **DOE** (Design of Experiment) process to design the survey. The survey questions should at least contain compensation and the 3 controlled charging attributes (time window, proportion, type), and then some additional questions about customer preferences. 2. Implement the survey using [formr.org](formr.org). 3. The survey should be in 3 stages: - The "think aloud" pilot survey - among a small amount of BEV owners; - Pilot survey - using Amazon *Mechanical Turk*; - Final survey - using platforms like *Dynata* or *Qualtrics*; via partner coordination; and in online BEV forums. 4. I found 2 versions for "preliminary survey designs": - p.36 - 2500 completed respondents in batches of 500, out of which we need 2000 final samples. - p.12 - 4000 completed respondents in batches of 500, out of which we need 3000 final samples. 5. With the results of survey, we use **MLE** (Maximum Likelihood Estimation) to estimate the parameters of **DCM** (Discrete Choice Modeling) and generate utility functions with both deterministic and random components, out of which we construct the *logit model*. This requires the use of the `logitr` package. 6. The marginal utility will be **converted** to dollars, which is a direct measure of values from each feature. 7. Use the model to conduct simulations to **predict** the best smart charging programs. 8. Integrate with Research Thrust 2... 9. In summary, our R packages, platforms, and math are: - R packages: `cbcTools`, `logitr`, - Platforms: *formr*, *survey platforms*, *BEV forums*, - Math: **MLE**, **DCM**, **logit**. ## Q & A **Question 1**: I know the meaning of "utility" in the context of DCM (i.e. utility maximization, or marginal utility), but in the proposal, "utility" can also refer to the "provider" or the "grid", right? **Answer 1**: Yes "utility" refers to the grid electricity provider. It is also used in the context of DCM. **Question 2**: Does "compensation" refer to the supplement to the smart charging implementation in a lager scale, or the bonus awards to the participants (the BEV owners), or both? **Answer 2**: "Compensation" refers to some sort of incentive to participate in smart charging. **Question 3**: What is the key difference between "high BEV adoption" and "high renewable energy adoption", in the 2 case studies in California? I know BEV is on the *demand* side, and renewable energy is clean energy *supply* (since we also consider low carbon emission), but what roles do they play differently in our study? **Answer 3**: "High BEV adoption" means a lot of people buying BEVs and "high renewable energy adoption" means grids that have a lot of renewable power, like solar and wind. CA has a lot of both. If we have a lot of BEVs enrolled in smart charging, they can be used to help integrate the renewable energy more smoothly. **Question 4**: There will be simulations required for our prediction process (*pp.38-39* Building predictive models). How do we implement the simulations? Are there specific tools and processes? Or do we follow the other team in Thrust 2? **Answer 4**: We'll get to simulations later. The point of them is to predict how many BEV owners would enroll in the smart charging program and compute the benefits on the grid side in terms of energy savings and renewable energy added to the grid. **Question 5**: I read through our Gantt Chart. I believe the green colored tasks are mine. How about the orange color? Does it mean utility providers or collaboration with our Co-PI and the other student? **Answer 5**: We're focused on the conjoint part for now (year 1). My colleague at UC Irvine will be leading the energy modeling, and he'll be working this year to put together some basic energy models that do not include smart charging preferences. Then later after our survey is done we'll add that part in.