Scenarios in SAInt
A scenario is a case study performed on a network. It is characterized by a scenario type (e.g., steady-state or dynamic gas simulation, ACPF, ACOPF, DCUCOPF, etc.), a time window (e.g., a starting and ending time), a definition of settings, controls, and constraints, and their behavior over time (e.g., how they change utilizing events and profiles). The mathematical model describing the type of scenario can be distinguished between:
-
Simulation: Compute a solution for the continuous variables that fulfill a set of (non-)linear equations.
-
Optimization: Compute a solution for the continuous or discrete decision variables that minimize an objective function that is subject to inequality and equality constraints.
SAInt offers the following scenario types for electric (Table 1), gas (Table 2), thermal (Table 3), and combined electric-gas (Table 4) networks.
There are five types of electric scenarios in SAInt:
-
Balanced AC-Power Flow Simulation (ACPF),
-
Balanced AC-Power Flow Optimization (ACOPF),
-
Unbalanced AC-Power Flow Simulation (UACPF),
-
Linearized Unit Commitment AC-Power Flow Optimization (DCUCOPF/PCM),
-
Capacity Expansion Model Optimization (CEM).
Network | Scenario | Description |
---|---|---|
Electric |
Balanced AC-Power Flow Simulation (ACPF) |
Single- or multi-time step (quasi-dynamic), phase-balanced, steady-state, alternating current (AC) power flow simulation. Simulation of power flows in an electric network considering:
The resulting mathematical model is a set of non-linear equations solved using a Newton-Raphson linearization algorithm. |
Electric |
Balanced AC-Power Flow Optimization (ACOPF) |
Single- or multi-time step (quasi-dynamic), phase-balanced, steady-state alternating current (AC) power flow optimization. Optimization of active power generation dispatch costs (objective function) in an electric network subject to:
The resulting mathematical model is a non-linear optimization model (NLP) solved with Gurobi’s LP-Solver using a sequential linear programming algorithm (SLP). |
Electric |
Unbalanced AC-Power Flow Simulation (UACPF) |
Single- or multi-time step (quasi-dynamic), single-, two- or three-phase unbalanced, steady-state, alternating current (AC) power flow simulation. Simulation of power flows in an electric network considering:
The resulting mathematical model is a set of non-linear equations which is solved using a Newton-Raphson linearization algorithm. |
Electric |
Linearized Unit Commitment AC-Power Flow Optimization (DCUCOPF, PCM) |
Multi-time step optimization of the decisions on unit commitment and economic dispatch considering generator, storage, and transmission constraints using a linear approximation of the non-linear AC power flow equations ( i.e., (a) nodal voltage = nominal voltage, (b) resistance << reactance, (c) negligible voltage phase angle difference across lines and transformers). Each optimization time horizon can have a look-ahead period to inform decisions that influence the network’s state beyond the end of the optimization window. The resulting mathematical model is a linear, mixed-integer optimization model (MIP) solved with Gurobi’s MIP-Solver using a rolling time horizon with look-ahead method. |
Electric |
Capacity Expansion Model Optimization (CEM) |
Long-term planning tool to model the installation of generation, storage, and transmission capacities in an energy system. CEM seeks to find optimal values for the capacities of these generation, storage, and transmission assets to meet forecasted demand for one or several future years. It can incorporate policy restrictions, such as emission goals, prices, or governmental subsidies, to study their effects in a perfect market environment. The resulting mathematical model is a linear optimization model (LP), which is solved with Gurobi’s LP-Solver considering discrete investment years and multiple representative periods for each investment year. |
There are two types of gas scenarios in SAInt:
-
Steady-State Hydraulic Gas Pipeline Network Simulation (SteadyGas), and
-
Dynamic Hydraulic Gas Pipeline Network Simulation (DynamicGas).
Network | Scenario | Description |
---|---|---|
Gas |
Steady State Hydraulic Gas Network Simulation (SteadyGas) |
Steady-state, single-time-step, hydraulic gas network simulation with gas quality, composition, and temperature tracking and considering:
The resulting mathematical model is a set of non-linear equations solved using a Newton-Raphson linearization algorithm. |
Gas |
Dynamic Hydraulic Gas Network Simulation (DynamicGas) |
Dynamic hydraulic gas network simulation. Simulates the operation of a gas network under time-varying demand profiles, control settings, and set points considering:
The resulting mathematical model is a set of nonlinear, implicit finite difference equations solved using a Newton-Raphson linearization algorithm. |
There is one type of thermal scenario in SAInt:
-
Thermal-Hydraulic District Heating and Cooling Pipeline Network Simulation (Steady/Quasi-Dynamic Thermal).
Network | Scenario | Description |
---|---|---|
Thermal |
Thermal-Hydraulic District Heating and Cooling Pipeline Network Simulation (Steady/Quasi-DynamicThermal) |
Single- or multi-time step (quasi-dynamic), steady-state thermal-hydraulic simulation of district heating and cooling pipeline networks considering:
The resulting mathematical model is a set of non-linear equations solved using a Newton-Raphson linearization algorithm. |
There are two types of combined gas-electric scenarios in SAInt:
-
Steady-State Hydraulic Gas Network Simulation (SteadyGas) combined with Balanced AC-Power Flow Simulation (ACPF)
-
Dynamic Hydraulic Gas Network Simulation (DynamicGas) combined with a Balanced AC-Power Flow Simulation (ACPF)
Network | Scenario | Description |
---|---|---|
Electric & Gas |
Steady State Hydraulic Gas Network Simulation (SteadyGas) combined with Balanced AC-Power Flow Simulation (ACPF) |
Steady (single-time step) combined simulation of alternating current power flow and a steady hydraulic gas network simulation. The resulting mathematical model is a set of non-linear equations solved using a Newton-Raphson linearization algorithm. |
Electric & Gas |
Dynamic Hydraulic Gas Network Simulation (DynamicGas) combined with a Balanced AC-Power Flow Simulation (ACPF) |
Combined simulation of a succession of (steady AC-power flow) simulations of power flows in an electric network using AC power flow equations and a transient hydraulic gas network simulation under time-varying demand profiles, control settings, and set points. The resulting mathematical model is a set of non-linear equations solved using a Newton-Raphson linearization algorithm. |
1. Electric Scenarios
1.1. Power Flow (ACPF & UACPF)
Understanding how current flows across an alternating current (AC) electric network, especially as generators energize it, is crucial for maintaining the network’s safe and stable operation. The AC power flow (ACPF) problem addresses this by identifying the active and reactive power flows, nodal voltage magnitudes, phase angles, and generator outputs characterizing various operating conditions. These conditions are influenced by fluctuating demands from residential, commercial, and industrial consumers, requiring matching variations in generation sources. At any moment, the power supplied must equal the power consumed and lost within the system.
The ACPF problem uses the network’s physical properties - such as transmission line impedances and susceptances - to create a mathematical model that solves these operating conditions, assuming the system is in a steady state. This means that, at any given operating condition, the supplied and consumed powers are balanced and unchanging. This is the basis of the SteadyACPF scenario, which considers the ACPF problem for a single instance.
System operators employ various tools and equipment to enhance network operations, especially during instances of high voltages or losses. SAInt provides core components for the true emulation of most electric networks. This includes transformers with automatic tap controls for local and remote voltage regulation, line drop compensation, and reactive power control. Shunt devices, capable of providing or consuming reactive power for voltage regulation, are also available with automatic switching control.
Distribution systems, which interface consumers with the transmission system, operate at lower voltages and are prone to unbalanced loading. This necessitates individual phase modeling rather than assuming phase balance. SAInt’s unbalanced ACPF (UACPF) capabilities allow for such detailed modeling, accommodating spatial and load diversity. Users can model both distribution and transmission systems simultaneously, assessing their interplay - an advanced feature not commonly available in commercial electric network simulation platforms.
Understanding how operating conditions evolve over time is as important as steady state analysis. SAInt offers Quasi-Dynamic scenarios for previously mentioned steady power flow scenarios. These scenarios perform a sequence of power flow simulations over time, allowing for the evaluation of how operating conditions change with varying loads and generation. This complements SAInt’s production cost modeling capabilities by facilitating a direct analysis of a wide range of dispatches.
1.2. Optimal Power Flow (ACOPF)
Unlike the power flow problem that assumes a unique solution based on equal numbers of unknown variables and equations, the optimal power flow (ACOPF) problem introduces optimization. In ACOPF, constraints replace set values (e.g., a range of operating voltages instead of a specific setpoint), transforming it into an optimization problem where the solution seeks to fulfill a particular objective, such as minimizing generator operating cost.
Given constraints like minimum and maximum voltages, active and reactive power limits, and transmission line capacities, the ACOPF problem optimizes the operating conditions. SAInt can execute ACOPF for a single time step or extend it to multiple time steps in quasi-dynamic scenarios.
By addressing both steady-state and quasi-dynamic scenarios, SAInt provides a comprehensive platform to model, simulate, and optimize electric network operations, offering unmatched capabilities in the energy industry.
1.3. Production Cost Modeling (DCUCOPF)
SAInt offers a state-of-the-art module to run production cost models through its Direct Current Unit Commitment Optimal Power Flow (DCUCOPF) scenario. As the name indicates, the DCUCOPF scenario solves unit commitment and economic dispatch problems using a DC approximation of the power flow. The network flows can also be represented using a transport model. The DCUCOPF scenario in SAInt is designed to be highly flexible and modular. The user controls how the production cost model is set up and solved. For example, settings like the duration of the optimization horizon, length of timesteps, and the parameters of an optional look-ahead period can be chosen to model day-ahead, intra-day, or multi-year electricity market models. Unit commitment is an inherently hard optimization problem. The increasing need to model complex interactions between different technologies (e.g., modern storage technologies, renewable generation, flexible demand, etc.) renders it even harder to model and solve planning problems involving unit commitment and economic dispatch. To assist the user in understanding the mathematical equations taking place in the model, the mathematical models prepared by the software are available for the user to inspect in an LP file format.
The DCUCOPF scenario offers several operational constraints, like flow limits on transmission lines, upper and lower limits on generation units, state-of-charge limits on storage devices, and volume limits on hydroelectric plants. In addition to the preset operational constraints, the user can create user-defined constraints. User-defined constraints enable the user to include complex features on top of the standard unit commitment and economic dispatch model. For example, it is possible to specify the combined power output of two generators should be within a specified threshold or the fuel consumption of a specific fuel generator should stay within a pre-defined envelope. This feature gives the user improved control over the model and solution process. Other capabilities of DCUCOPF include:
-
The ability to model ancillary services.
-
The ability to specify detailed cost parameters (e.g., piece-wise linear fuel consumption curves for thermal units).
-
The ability to control parameters of electric storage objects (e.g., charging, discharging, and state-of-charge with events and user-defined constraints).
-
The ability to model hydropower plants as a separate entity from the hydropower generators. This allows the representation of complex interactions between hydro plants and generators). For example, the cascading flow (spilling, turbinating) from multiple linked hydro plants can be accurately modeled.
-
The ability to use demand objects to represent various demand models. The user can model fixed and price-responsive demands by varying the magnitude of penalty prices. Features like the ability to control the demand pattern using an averaged profile can also be utilized to model flexible demands. Additionally, SAInt has a prosumer object representing demands that can consume and supply energy (e.g., grid-connected rooftop solar).
1.4. Capacity Expansion Modeling (CEM)
Capacity expansion modeling (CEM) is a long-term planning tool to model the installation of generation, storage, and transmission capacities in an energy system. CEM seeks to find optimal values for the capacities of these generation, storage, and transmission assets to meet forecasted demand for one or several years. It can incorporate policy restrictions, such as emission targets, carbon fees, or governmental subsidies, to study their effects in a perfect market environment.
Common use cases of CEM include the study of how an energy system might look in the future, how it should look in the future, which emissions and costs it will entail, and how variations in input parameters and information shocks might affect these outcomes. It allows resource planners to develop strategies to address future renewable portfolio standards and emission regulations. The resources (and alternatives) that can be analyzed in the CEM scenario include curtailable demand, fuel generators, storage technologies, renewable generators (solar and wind), and transmission lines. The solution uses linear programming (LP) algorithms to solve for the optimal future mix of these technologies while respecting an array of operational and policy constraints. Some of the salient features of the CEM scenario in SAInt include:
-
Supports both greenfield and brownfield models, where the brownfield model can include existing network assets.
-
Offers several investment options for assets. For example, an asset can be newly built, expanded, retired, or rendered inactive for potential future use (mothballing).
-
Enables solutions for both single-stage and multi-stage problems. In multi-stage models, investment decisions can be made over several investment years.
-
Provides a user-friendly mechanism to model demand and renewable forecasts using representative periods.
-
Allows the specification of investment constraints (minimum and maximum limits on installed capacity or storage volumes) and policy restrictions (caps on emissions, renewable quotas, etc.).
The CEM scenario is designed to offer control over the investments' spatial and temporal resolution. The user can specify most constraints at the object, node, zone, group, or network level. From a temporal sense, the constraints can be included on a per-year or per-horizon basis.
2. Gas Scenarios
Diverse challenges need to be addressed when it comes to simulating the behavior of the gas flow in transmission and distribution networks. Some of these challenges include:
-
Gas systems consist of hundreds or even thousands of elements such as pipelines, compressors, regulators, supplies, demands, and storages that are connected and directly or indirectly affect the behavior of the other elements.
-
Gas is a compressible fluid which means its thermo-physical properties depend on pressure, temperature, and composition.
-
Multiple physical and operational constraints for each gas network element must be considered in a simulation.
-
Because of the varieties of gas compositions from various sources (e.g., injecting hydrogen into a gas network), the quality of the gas can vary dramatically in different parts.
-
This interconnected system is rarely in its steady mode since the parameters of the gas networks' elements can change over time. Analyzing this system’s transient behavior is necessary for planning and operation purposes.
A user of SAInt can overcome and address these challenges. SAInt provides detailed models of various gas transmission and distribution network elements and their interconnections. SAInt’s gas network simulation builds the model based on the conservation of mass, momentum, and energy, as well as using a state equation for taking into account the non-ideal behavior of the gas in high-pressure networks. SAInt considers a wide range of physical and operational constraints and finds a solution within bounds by taking advantage of its control mode switching algorithm. The quality and temperature tracking capabilities of SAInt allow the users to simulate and analyze the changes in gas quality and temperature in different network parts. Therefore, the injection and blending of different gas mixtures (e.g., high and low calorific natural gas, biogas, hydrogen, etc.) at different supply points can be modeled.
3. Thermal Scenarios
The thermal network module can model district heating and cooling (DHC) networks in SAInt. A significant part of primary energy consumption is dedicated to heating or cooling needs. The optimization of the operation and the planning of such networks comes with a huge benefit. Furthermore, the potential to effectively store energy inside the water flow of DHC networks can lead to efficient integrated energy systems operation. The main inputs to the model are the pipe characteristics, the downstream temperatures, and the demand and supply heat exchange rates. The outputs of the model are the mass flows, the pressure profile, and the previously unknown temperatures. The user can specify the heat exchanged into or out of the demand and supply flow, so two sets of equations, hydraulic and thermal, are required in the modeling. The modeling focuses mainly on the pipe, the external, and the node.
In real-world operation, the supply information is most likely incomplete because of the unknown amount of heat loss along pipes. The user can set the loss to be compensated by multiple supplies located dispersedly in the network. Such a feature has been considered in the temperature drop equation and the energy balance equation, so no additional equation is added to the system, and there is no need for extra iterations. The performance is thus not compromised.
Finally, there is a comprehensive validation of inputs and a friendly log system to help the user identify the problematic inputs. One important validation is ensuring the demand and the supply are not over or under-determined.
4. Combined Scenarios
The installations of gas-fired power plants around the world, significant replacement of the traditional gas turbines with electric drivers to operate facilities in the gas system-like LNG terminals, underground gas storages, and compressor stations-and the considerable development of the power to gas technology, have dramatically increased the coupling and interconnections between gas and electricity networks. This increased coupling between gas and electricity networks brought economic and environmental benefits. However, it made the planning and operation of these two systems much more challenging and necessitated using tools capable of modeling the interdependencies between them.
A combined scenario describes the operation of two or more networks of different types according to conditions specified in each participating scenario and to the type and properties of the coupling objects. One of the unique features of SAInt is that it allows users to perform combined gas and electric network simulation. In the combined simulation, the system of equations that describe the behavior of each of the networks will be integrated by adding coupling equations that model the interconnection between gas and electric facilities. Thus, simultaneously solving this integrated system of equations represents the behavior of these two combined interconnected systems. SAInt simulates various interconnections between gas and electricity networks as a "hub object". Hub objects include:
-
The gas offtake from gas networks to generate electricity in gas-fired power plants connected to electricity networks.
-
The power offtake from the electric network to operate the electric engines in gas compressor stations.
-
The power offtake from the electric network to operate LNG regasification terminals.
-
The power offtake from the electric network to operate gas storages.
-
The power offtake from the electric network to operate Power-To-Gas facilities and to inject hydrogen and synthetic natural gas into the gas network system.
5. State Transfer
The state transfer process is conceptually the extraction of operating conditions from one type of simulation to be used as initial conditions or to adjust constraints, on other types of simulations. This process leverages the unique approach of the SAInt software of establishing a network with as much detail as the user can provide, from which a slew of different simulation types can be applied, which will only use the network properties relevant to the simulation of interest. In this way, multi-timescale/detail simulations can be executed on the same network, making data handling and control far more efficacious.
The predominant application of this process in the SAInt software to date is as an interface between DCUCOPF and AC(O)PF simulations. The DCUCOPF simulation results consist of active power set points and commitment status. These results can be transferred within the SAInt framework to an AC(O)PF scenario, where the increased detail of the power flow simulations is applied to determine the network’s voltage and reactive power operating conditions. This transfer can be used to execute steady and quasi-dynamic simulations.