(216e) Energiapy - an Open Source Python Package for Multiscale Modeling & Optimization of Energy Systems
AIChE Annual Meeting
2023
2023 AIChE Annual Meeting
Computing and Systems Technology Division
Software Tools and Implementations for Process Systems Engineering
Tuesday, November 7, 2023 - 1:42pm to 2:00pm
The energy transition poses formidable challenges [1], such as a plethora of technology and resource
options, renewable intermittency, uncertainty of future scenarios, a dearth of materials to establish
technology pathways, resiliency to disruptions, the accounting of carbon emissions over the life cycle
of technology use, interaction and competition with vital chains such as food and water [2]. Tools
that support data-driven decision making are essential to achieve net-carbon neutrality over the
coming decades. To this end, numerous energy systems modeling frameworks have been developed
utilizing disparate architectures and programming languages such as AMPL, Python, Julia, etc.
[3, 4].
options, renewable intermittency, uncertainty of future scenarios, a dearth of materials to establish
technology pathways, resiliency to disruptions, the accounting of carbon emissions over the life cycle
of technology use, interaction and competition with vital chains such as food and water [2]. Tools
that support data-driven decision making are essential to achieve net-carbon neutrality over the
coming decades. To this end, numerous energy systems modeling frameworks have been developed
utilizing disparate architectures and programming languages such as AMPL, Python, Julia, etc.
[3, 4].
Here, we present energiapy, a decisions making and risk analysis tool for the design and optimiza-
tion of energy systems. The novelty of energiapy lies in the highly configurable system agnostic
component-based approach. Notably, formulations can span from the level of individual processes
to networks spanning multiple regions. The object-oriented approach allows deterministic factors
to be assigned at the appropriate level, viz. localization factors at the location level, and pro-
cess parameters at the process level. The components involved follow a suggested hierarchy which
allows the formulation of models ranging from multi-scale mixed integer programs (MIPs) to multi-
parametric, and robust stochastic programs [5]. This lends significant flexibility to the user in both
developing network models, as well as the scenarios over which these models are tested. Moreover,
scenarios can also be aggregated and solved over a set of representative temporal periods, used to
generate graphs for network analysis, and both the data input and solutions can be visualized.
These developments are demonstrated over a detailed hydrogen economy case study [6] which is
solved through different approaches. First as a determinsitic mixed integer linear program (MILP),
then as a multi-parametric linear program (mpLP), and then as its robust equivalent. The key
applications illustrated in the study include: 1) the design of future energy systems (network
design), 2) scheduling under uncertainty, 3) life-cycle and environmental impact assessment, 4)
techno-economic analysis, 4) system resiliency and reliability characterization.
solved through different approaches. First as a determinsitic mixed integer linear program (MILP),
then as a multi-parametric linear program (mpLP), and then as its robust equivalent. The key
applications illustrated in the study include: 1) the design of future energy systems (network
design), 2) scheduling under uncertainty, 3) life-cycle and environmental impact assessment, 4)
techno-economic analysis, 4) system resiliency and reliability characterization.