1. Problem Statement

We are developing a system to predict problems with the power grid. We are analyzing large amounts of data from power grids in the midwest and building a machine learning model to understand fluctuations and potential problems and designing a system to predict such problems

2. Requirements & Constraints

List all requirements for your project . This includes functional requirements (specification), resource requirements, qualitative aesthetics requirements, economic/market requirements, environmental requirements, UI requirements, and any others relevant to your project. When a requirement is also a quantitative constraint, either separate it into a list of constraints, or annotate at the end of requirement as “(constraint)”. Other requirements can be a single list or can be broken out into multiple lists based on the category.

Frontend

• Use Google Cloud (Requirement)
• Map Based Visualization for all grid components (Requirement)
• Create User Interface for ML Grid Assistant and ML Rec. (Requirement)
• Admin Interface: User Interface data upload, file management, set up software in backend (Requirement)

Backend

• Use Google Cloud (Requirement)
• Microservice arch: docker-ize a tensorflow python ML container (Requirement)
• Add Time-series databases: optimal store/retrieve information (Requirement)
• Possibly use Influx DB (Constraint)

Machine Learning

• Use Google Cloud (Requirement)
• Able to generalize to different region’s power grids (Requirement)
• Must have a overall accuracy of greater than 90%
• ML Grid assistant: allow general queries/commands, search within the database(Requirement)
• ML Recommendation System: searching to wattage, can also see related fields (Requirement)

1. Engineering Standards

What Engineering standards are likely to apply to your project? Some standards might be built into your requirements (Use 802.11 ac wifi standard) and many others might fall out of design. For each standard listed, also provide a brief justification.

• IEEE/ISO/IEC 12207-2017: Software life cycle processes:

• The previous team used these requirements and our client recommended these as a base for our team.

• Google Cloud Standards:

• For cloud provider vendor specification

• Tensorflow/ Python ML:

• To meet Google cloud vendor specifications

2. Intended Users and Uses

Who benefits from the results of your project? Who cares that it exists? How will they use it? Enumerating as many “use cases” as possible also helps you make sure that your requirements are complete (each use case may give rise to its own set of requirements).

Those who will have the most significant benefit from this project will be power grid operators and their ability to service power grids. This system will allow operators to be made aware of potential problems and failures of the power grid system before they happen. Those who care that this exists are anyone who uses the power grid system or lives within proximity of a power grid node. Having a system that can predict failure points and help those who serve the power grid fix these systems faster will result in fewer power outages and a safer environment. This will help prevent minor power grid failure-related disasters such as forest fires and electrically charged water. System providers will use this system to service power grids to ensure that failures within proximity of the system will be less likely.