Skootik – Optimal construction planning of electrical projects
Skootik
Sector: Energy
Business Case
We have several electrical projects (substations, lines, repowering, etc.) to be built. There is not enough capacity to start building them all at once, so one has to plan when to build each project. Each project involves a construction cost, and the end of the project brings with it a financial return.
Objectives
Propose a plan for the construction of projects ensuring that the plan will comply with the capabilities of the construction company and the external limits imposed on the company (legal, bureaucratic limits, delays, etc.). In turn, each project entails a financial return to the construction company, which is why it seeks to maximise the immediate investment by prioritising the most profitable projects so that they can be built as soon as possible.
Use case
The planning problem has been addressed by proposing an integer-mixed programming model and solving it either by exact methods (a B&C solver) or by genetic algorithm. Both the projects to be built and their proposed optimal planning are displayed in a user-friendly interface where the user can plan by manually adjusting the parameters describing the exact conjuncture of the project portfolio.
Infrastructure
It is a Cloud service, so a server is procured to host the system along with a platform for user interaction with the data and an ingest system that connects to the client’s data lake.
Technology
Automatic or Deep Learning
Data
The construction company’s project portfolio along with information about each project and information about the company’s construction capacity. All data are private.
Resources
One back-end and one front-end developer was needed, along with a project coordinator. On the part of the construction company, a contact person with weekly availability for video calls and daily availability for emails was necessary, together with the involvement of several departments to enrich the project portfolio and obtain representative data.
Difficulties and learning
The processing, standardisation and unification of data from different sources was a complex challenge. In addition, PIM models called for considerable execution times due to the number of projects to be planned and the large number of constraints imposed on the planning.
KPIs (business impact and metrics of the model)
The main KPI was the total planned remuneration for the first 5 years of planning. The impact on the business, in addition to the financial one, has been to make bottlenecks visible at the time of construction.
Funding
Private, by the client
Collaborators, Partners
–
