Overview of the Cairn tool

Cairn principles

CEA has been developing an optimization software called Cairn to conduct techno-economic and environmental studies that help decision making when looking for optimal energy mixes at several scales of the energy system, from industrial premises to territories or countries.

What is an energy system?

An energy system is a system designed to supply energy-services to end-users.
An energy system consists of inlet primary energies (fossil, renewables,…), conversion of primary energies to intermediate ones and outlet or final energies to end-users.
An energy system is constituted by all components related to the production, conversion, storage, delivery, and use of energy.
An energy system accounts for all technical, environmental, economic and social aspects.

The Cairn tool allows in a modular way to define and solve a techno-economic problem including environmental aspects by assembling several Mixed Integer Linear Programming (MILP) contributions of a dedicated C++ core library by modelling the main technological components of multi-energy systems.

Cairn is mainly used to solve the optimization problems of so-called unit-commitment problems with size optimization, trying to find the optimal energy mix for the energy system submitted to one or several time load profiles.

Specific constraints between components and their variables can be defined without any code development and through smart connectors.

Solving the problem is performed thanks to an Application Programming Interface (API) of various Open Source or commercial mixed integer linear solvers.

Caution

Cairn is not a solver.
Cairn does not supply solvers.
Solver has to be installed by users.

The user is able to build the energy system either through a C++ Application Programming Interface (API) or through a Graphical User Interface (GUI).

Modelling principles of Cairn 

A multi-energy system (energysystem) is modelled by an set of technological components exchanging flows of different energy carriers, and/or state data (internal temperature, operating status…). Each technological components is associated to a model that writes MILP contribution to a global optimization problem. An optimization problem consists of maximization or minimization of an objective function under constraints.

Each energy or mass term (electrical, thermal or material, H2) is associated with an energy carrier. These energy carriers give the nature of the flow associated to it and the different units to be used. Energy carriers are linked to smart connections (bus or nodes). These smart connections enable data exchange between models.

Technological components models are divided into 4 main classes

Converters models describing the behavior of conversion technologies (Heat Pumps, combined heat and power units, electrolyzers, methaners, biogas or biomethane units) with their efficiencies and constraints (start-up or ramp) to pass flow from a given energy carrier to another one.
Storage models modelling capacities of energy or mass.
Load models which impose source or load time profiles, with flexibility, load shedding or curtailment abilities.
Grid models which calculate injected or extracted flow, at a given cost.

A technological component model includes:

Input parameters : the techno-economic and environmental parameters (eg efficiencies, environmental impacts, costs), the boundary condition time series (eg the prices), the maps (CAPEX, performance…) and the setting options.

Decision variables :

About sizing : capacities (nominal production power or flow rate, maximum storage capacity).

About operation and control : state variables that are time-dependent, and instantaneous flows exchanged between components (UsedPower between Grid and Electrolyzer).

Constraints :

Linear constraints linking decision variables and parameters (e.g. the relationship between input and output, as a simple efficiency between electricity consumption and Hydrogen production in an Electrolysis system).

Advanced constraints like piecewise linear equations involving decision variables to reflect non linear operation constraints (e.g. an efficiency map for a Heat Pump).

Several modeling levels are possible to match the desired compromise between time and accuracy, as for the example of a Li-Ion battery :

it can be modeled either like a generic energy storage with constant charging and discharging powers assuming main use in the range 20% to 80% of full charge,
or it can be modeled using a performance map giving precisely charging and discharging powers as function of the state of charge of the battery.

Smart connections between ports of models use specific constraint modules; which can be of the following classes:

System constraints (Bus) :

Node Laws : the variables connected to the node are summed at each time step, a sum value can be imposed (eg 0 to ensure flow balances).

Node equality : the variables connected to the node must be equal at each time step.

Generic operational constraints to add to technological models (for instance ramp constraints).

Physical models to add physical generic equations to some technological components (temperature).

This modularity, based on object-oriented programming permits having a great modeling freedom as it is illustrated in Fig. 2 View and modify easily energy architectures with GUI, as well as the ability / capability of capitalizing on models.

Cairn is a tool that allows building a model of a multi-energy system and its associated data, and to turn it into an optimization problem. Cairn is used to :

solve optimization problems of so-called unit-commitment problems with size optimization trying to find optimal energy mix for an energy system submitted to one or several time load profiles.

simulate the operation of the system. The set of equations describes the state of the system (with accounts for all constraints), depending on the past and on the boundary conditions.

What are optimization and simulation modes?

Optimization and simulation are two commonly used approaches in system modeling and analysis, but they serve different purposes.

Optimization aims to find the best possible solution to a given problem, by maximizing or minimizing an objective function. For example, this could involve minimizing costs, maximizing profit, or achieving the best trade-off between multiple criteria.

Simulation, on the other hand, is used to model the behavior of a real or hypothetical system under various conditions without necessarily seeking the optimal solution. It allows understanding how a system works, assessing the impact of different scenarios or conditions, as well as testing hypotheses.

After the optimization, Cairn allows analyzing the results. The Fig. 1 Overview of the steps of Cairn’s operation. illustrates the global steps of the Cairn operation.