Air Flow Entities

Air (and/or water) flow networks in ESP-r are composed of flow nodes, flow components and flow connections. Two different topologies are shown below:

Node and component topologies

Nodes

In the left of the figure the two zone nodes are of type unknown pressure internal because we want to evaluate the pressure of the air in the zones. In the right of the figure there are two boundary nodes of type wind-driven boundary node. At simulation time the current wind speed and direction are converted into a pressure at the facade and used, along with the flow characteristics of the opening components to evaluate the flow. The node types available in ESP-r are:

  • Wind-driven boundary nodes derive their pressure from climate data (wind speed and direction) and pressure coefficient set that resolves the angle of the wind at the facade into a pressure.
  • Fixed pressure boundary nodes are equivalent to a virtual blower door test with a static pressure in Pa.
  • Unknown pressure internal nodes are associated with a body of air (e.g. a zone) and act as bookkeepping points where pressure is resolved. This is the most common node type used with zones.
  • Fixed pressure internal nodes are equivalent to a virtual blower door test with a static pressure in Pa. If you have a strong opinion that an air source is at a fixed pressure then this type of node is what you want to use.

    • Wind pressures

    ESP-r includes a database of pressure coefficient sets. Each set has values at eight directions (22.5 degree intervals) which characterise the pressure relationship for a specific device. The user is tasked with supplying these values - perhaps from a wind-tunnel test or a CFD run.

    Components

    In the figure above a crack component (see below) was used to represent an undercut at a door. The shape of the undercut is not a good fit for a common orifice component. Simple air flow components (see below) were used to represent the top and bottom of the sash window. We could also have used a common orifice component.

    Components which are based on curve fits:

  • Power law volume flow or power law mass flow - where flow is well described by a power law. e.g. m = rho.a.dP^b
  • Quadratic law volume and mass flow - where flow is well described via a quadratic fit e.g. dP= a.m/rho+b.(m/rho)^2 or dP= a.m+b.m^2

    • Components that induce or restrict flow (e.g. ducts, valves) are typified by the following:

  • Constant volume or mass flow (abstract fan representation which can be controlled to vary flow)
  • General flow conduit (classic duct definition) e.g. m = rho.f(D,A,L,k,SCi)
  • Laminar pipe volume flow (class pipe) m = rho.f(L,R,mu,dP)
  • General flow corrector valve or damper m = rho.f(comp, signal)
  • Corrector with polynomial flow resistance e.g. C = f(H/H100)
  • Conduit starting or ending in a 3-leg junction & Ccp = f(q/qc)
  • Conduit starting or ending in a 4-leg junction & Ccp = f(q/qc)

    • Components associated with architectural elements such as windows, cracks and doors are typified by:

  • Common orifice - openings as area and discharge coefficient e.g. m = rho.f(Cd,A,rho,dP)
  • Air flow opening - fixed discharge coefficient (0.65) so only area needed
  • Air flow crack - useful for linear openings up to ~12mm wide e.g. m = rho.f(W,L,dP)
  • Bi-directions flow - useful for openings (doors and windows) where temperature and pressure differences can result in flow in two directions
  • Compound component - specifies two components e.g. an opening and a crack with control directives

    • Linkages

    In the figure above flow paths are defined via the syntax of node X is connected to node Y via component Z where Z is D1 distance above/below X and Z is D2 distance above/below Y. The crack is below the two zone nodes and thus the D1 and D2 are negative. For the sash window, lets assume that the boundary nodes heights were set to be the centre of each opening. The D1 values are from the point of the height of the zone node and the D2 values will be zero.

    Control

    In ESP-r, most flow components can have control imposed (but we would not impose control a crack). Control follows the general ESP-r pattern of defining the location of the sensor, what it senses, defining the component to be actuated as well as a schedule (for each day type) of control laws (control logic). For example control applied to an air flow opening or orifice would inpose a different area and control applied to a bi-directional component would alter the width of the opening.

    Simple controls which act on a component can be associated with the following control actions:

  • ON/OFF based on a sensed condition with a directive for inverse operation.
  • Proportional controller with hysteresis (ideal proportional control)
  • PI controller (PID controller which can be extensively tuned)

    • Some control actions are more complex and one approach is to adjust the topology of the network so that controls can be placed in series as in the image below. Flow along the path is only possible when both the window control control and the additional component controller agree.

    Adapted topology

    An alternative is to use a multi-sensor flow controlM/i> in which multiple sensors and multiple AND OR logic states. It starts with a default position (either open or closed) and changes (i.e. close or open) the connection/component according to up to four logical AND/OR conditions as in the figure below.

    AND OR logic

    Relationship with other solvers

    Discussion ... In XX the flow solution [preceeds/follows] the zone solution and ...

    In XX the flow solution [preceeds/follows] the system solution and ...

    A network flow solution supersedes scheduled air movement defintions. Users can bein with scheduled flows (imposed) and then create a flow network.


    Back to top | Back to Welcome page
    ©Copyright 2017 Energy Systems Research Unit, Glasgow, Scotland. License: GPL V2. Last edited by JWH, 21 Feb 2017