Environmental control systems solvers

Overview

[overview approach(s) used to represent and solve environmental control systems...]

[if a network approach taken describe the topology of the network...]

The environmental system components included with ESP-r represent a variety of approaches to how components are represented, their thermophysical resolution and the level(s) of control which can be implied. There are thus multiple radiators, boilers and accumulators to choose from. Some components are represented as a single control volume while others may have more than a dozen. The number of attributes ranges from 4-132. Some components also include data for associated mass flow entities.

Below are a few examples of components which might be used in HVAC networks. A full list is included in the standard database are found (here).

ESP-r does not use pre-defined templates for environmental control system networks. Users are expected to design the system (from the available components).

Heat mass and control paths between components

[Discussion how heat, mass, control signals and electrical power paths within and between components are handled (and solved)...]

[If templates are used briefly discuss their creation....]

[If templates are used briefly discuss how they can be adapted by users....]

[Discuss how relationships between components components are defined. Use one or two examples to clarify....]

Component discretisation

[Discuss how / whether sub-sections of a component, for example a boiler are represented - is is a black box or a collection of discrete components being solved....]

Component attributes

[Discuss user editable attributes of component, for example the number of and type of attributes for one or two examples e.g. a boiler....]

[Link to a list of available components (where readers can explore)...]

[If multiple representations of a component could exist (e.g. a simple early design stage vs a very detailed version which could be used to fine-tune performance) discuss the implications for solving diverse networks...]

Communication with other solvers

[Discuss how system components and/or templates gather information from the building zone domain as the solution progresses...]

Control of components

[Discuss how control of components tends to be implemented. Give examples...]

Environmental control sensors

[Discuss supervisory control i.e. of collections of components. Give examples...]

[Discuss sensors associated with environmental controls - are these implied/implicit in the component(s) or explicitly defined by the user. Give examples...]

Environmental control logic

[Discuss control logic that can be associated with environmental controls - are they explicitly defined by the user? Can the user tune a PID controller? Real devices often have complex control diagrams with multiple AND OR states. Give examples...]

The following is taken from a 2007 paper by Clarke, Kelly and Tang:

The solution of the plant matrix is dictated by control interaction, where the state of the system components is adjusted to bring about a desired control objective. The sensed node for plant control may be another plant component or node in the building model, e.g. the flow rate of cold water into a chiller coil may be controlled based on the relative humidity in a thermal zone.
As a general rule, the plant-side matrix equation is substantially smaller than its building-side counterpart. For example, within ESP-r the total number of equations for a domestic central heating system is approximately 150, while a building-side model for an average-sized house will require approximately 1000 equations. It is therefore possible to process the plant model as two equation-sets for energy and mass balance (up to two phases are permitted) without the application of partitioning to accommodate sparsity.

Solution timesteps

The solution timestep follows the users directives. By default this is fixed for the duration of the assessments. It is possible to over-ride this with a timestep control regime. Such a regime could detect changes in boundary conditions as well as rapid changes in internal conditions which could trigger shorter timesteps. This would appear to be a rarely used facility.

Available performance data

There are, essentially two paths for recording performance data. The most common is to set a save level which dictates to the solver the scheme of performance data to write to the domain results files. This is discussed further HERE

In addition the simulator includes a trace facility where intermediate data used in the various sub-systems of the solver can be written to file. The trace facility allows for specific sub-system topics to be identified as well as the period of the assessment when the additional information should be captured and exported.

[Discuss the computational resources implied by system solution and options for limiting resource use....]

Dependencies on the data model

[Discuss how templates and components are held in the data model.... And how the solver access this information....]

Numerical approach

The ESP-r approach to systems solution is summarised in a 2007 paper by Clarke, Kelly and Tang:

As with the zones that make up the building model, the control volume technique can also be used to derive the equation sets describing a building’s plant network. Within ESP-r the plant network consists of a coupled group of plant component models, each described by one or more control volumes. Mclean (1982), Tang (1985), Hensen (1991) and Aasem (1993) have applied the control volume modelling technique to a variety of plant components. The resulting library of models allows the simulation of air conditioning systems, hot water heating systems and mixed air/hot water systems as a network in ESP-r.

©Copyright 2017 Energy Systems Research Unit, Glasgow, Scotland. License: GPL V2. Last edited by JWH, 22 March 2017