Skip to content
xpang edited this page Feb 20, 2014 · 6 revisions

Summary

HVAC system faults modeling plays key role in various aspects of the building technology development. A collection of component and operational faults will be developed in the Modelica Buildings Library. The aim is to provide basic models of common HVAC faults in building energy simulation to address the needs for: * Modeling a realistic HVAC system with faults * Fault detection and diagnosis * Training and education

Possible Implementation

The implementation should be realized using the Modelica language as it is an open-source language that is well-posed to become the de-facto standard for modeling of dynamic systems. Models formulated in this language can then be used by different `modeling and simulation environments for Modelica <http://www.modelica.org/tools>.

Leaking economizer dampers

Description

The economizer dampers include the return air damper, the outside air damper and the relief air damper. The leaks can be caused by the aged seals on the edge of the damper blade or the loose linkages between the damper and the actuator. A leaking return air damper indicates that when the system is operating at 100% outside air condition, a certain amount of return air leaks through the return air damper and mixes with the outside air in the mixing box of an AHU resulting in reduced economizing capability. A leaking outside/relief air damper indicates that when the damper is commanded to be fully close, a certain amount of outdoor air leaks through the outside/relief air damper into the AHU. There are two major consequences causing by this: (1)In cold climate, when the system is off during unoccupied hours, the cold outside air can flow into the unit through the leaking outside/relief air damper and trigger the freezestat, and (2)If the leaking area is large, the outside air intake could be more than needed during non-economizer period.

Modeling approach

The flow coefficient of the air damper in the Modelica Buildings Library is modeled as a piecewise function. It consists of three pieces: (1) for yL < y < yU, the damper characteristics is k=exp(a+b(1-y)) (y is the control signal, and y=0 means the damper is closed, and y=1 means the damper is open. yL and yU are the low and high bounds that the damper characteristics holds.), and (2)for 0 <= y <= yL and yU <= y <= 1, the damper characteristics is defined by a quadratic polynomial that matches the damper resistance at y=0 and y=yL or y=yU and y=1. The polynomials are such that k(y) is differentiable in y and the derivative is continuous.

For the loose linkage, a user input "minimum damper closing position" will be use as the minimum control signal y. For the aged seal, the user input of the damper resistance at y=0 should include the impact from the leak of the aged seal.

Reverse acting damper

Description

Economizer dampers respond to the control command in the opposite moving direction, i.e. outside/relief air damper is fully closed and return air damper is fully opened when controller calls for 100% outside air. This can be caused by the inconsistent control actions between the PID controller in the control program and the settings in the actuator.

Modeling approach

Change the control actions of the PID controller in the model.

Stuck damper

Description

Economizer dampers are stuck at a fixed position. This fault has different scenarios. For the case that the economizer dampers are interlinked with a shared actuator, the malfunction actuator will cause the three dampers are stuck in the same time. If the linkage between the actuator and the damper is broken, the damper can be stuck individually. For the case that each economizer damper has its dedicated actuator, the damper will be stuck individually if either the actuator is malfunction or the linkage is broken.

Modeling approach

We will model the fault by fixing the control signal sent to the dampers.

Leaking control valve

Description

A leaking valve is characterized as a higher-than-normal amount of fluid flowing through a fully closed control valve. The leaking often arises from worn-out valve seat, undersized actuator, leaking diaphragm in pneumatic actuator or debris in the valve.

Modeling approach

Three valve characteristics were modeled in the Modelica Buildings Library, i.e. linear, quick open and equal percentage. In each of the models, there is an input parameter, l, for valve leakage (l=Cv(y=0)/Cvs). For linear and quick open valves, this input parameter can be directly used for modeling the valve leakage. However, for the equal percentage valve, the use of the l is tricky since equal percentage valve has two related input parameters, l and R. R is the rangeability defined as R=maximum controllable flow/minimum controllable flow. The l should never exceed 1/R for a normal valve. For a leaking valve with l > 1/R, R = 1.0/l. This approach applies to both two-way and three-way valve models.

Fouled coil

Description

During normal heat exchanger operation with most liquids and some gases a dirt film gradually builds up on the heat-transfer surface. The deposit is referred to as fouling. Its value depends on several factors including the type of fluid, the type of heat exchanger, the fluid temperature, the fluid velocity, and the length of service of the heat exchanger. The fouling impact on the pressure drop is typically negligible since the fouling is normal thin.

Modeling approach

This effect can be treated by introducing an additional thermal resistance, termed the fouling factor, Rf(K/W). For the heating coil and cooling coil, the fouling factors in both the water pass (Rfw, m2K/W) and air pass (Rfa, m2K/W) will be considered. The UAfouling is calculated as

UAfouling = 1/(Rf + 1/UAcleaning)

Rf = Rfw/Ai + Rfa/Ao (where, Ai is the inner water tube surface area, Ao is the external surface area)

The existing heat exchanger models will have four new input parmaters, Rfw, Rfa, Ai and Ao.

Stuck control valve

Description

The control valve is stuck at a fixed position due to malfunction actuators or loosely coupled actuator with the valve stem.

Modeling approach

Introduce an additional input parameter,Pstu, in the existing valve models to represent a stuck valve position. If this input parameter is enabled in the model, the control valve will fix at the position regardless of the control signal it receives.

Sensor offset

Description

Only constant sensor measurement offset is considered.

Modeling approach

Add a fixed value (positive or negative) to the sensor measurement.

Actual Implementation

To be done.