HVAC Controls & Automation Systems

Modern HVAC systems rely on sophisticated control strategies to maintain occupant comfort, minimize energy consumption, and ensure reliable operation. Control systems measure process variables through sensors, compare measured values against setpoints, and adjust actuators to eliminate control errors. Understanding control fundamentals, instrumentation, and automation architectures enables engineers to design responsive, efficient, and maintainable systems.

Control Theory Fundamentals

Feedback Control Principles

Feedback control forms the foundation of HVAC automation. A closed-loop control system continuously monitors a controlled variable (process variable), compares it to a desired value (setpoint), and manipulates an output to eliminate the difference (error).

graph LR
    A[Setpoint] --> B[Controller]
    B --> C[Actuator]
    C --> D[Process]
    D --> E[Sensor]
    E --> F[Measured Value]
    F --> B
    G[Disturbance] --> D

    style A fill:#e1f5ff
    style D fill:#ffe1e1
    style B fill:#e1ffe1

The control error at any time $t$ is:

$$e(t) = SP - PV$$

Where:

$SP$ = setpoint (desired value)
$PV$ = process variable (measured value)

The controller output $CO(t)$ is a function of this error:

$$CO(t) = f(e(t))$$

Control system performance is evaluated by:

Rise time: Time to reach setpoint
Settling time: Time to remain within acceptable tolerance
Overshoot: Maximum deviation beyond setpoint
Steady-state error: Persistent offset from setpoint

PID Control Algorithm

The Proportional-Integral-Derivative (PID) algorithm is the most common control strategy in HVAC applications. The controller output combines three control actions:

$$CO(t) = K_p \cdot e(t) + K_i \int_0^t e(\tau) , d\tau + K_d \frac{de(t)}{dt}$$

Where:

$K_p$ = proportional gain (dimensionless or %/unit)
$K_i$ = integral gain (1/time)
$K_d$ = derivative gain (time)

Proportional (P) Action:

Output is proportional to current error
Fast response to load changes
Creates steady-state offset (droop)
Throttling range $TR = 100/K_p$ (percentage)
Example: 10% throttling range → $K_p = 10$

Integral (I) Action:

Output based on accumulated error over time
Eliminates steady-state offset
Slower response, potential for overshoot
Reset time $T_i = 1/K_i$ (minutes)

Derivative (D) Action:

Output based on rate of error change
Anticipates future error trends
Reduces overshoot and improves stability
Sensitive to sensor noise
Rate time $T_d = K_d$ (minutes)

Control Mode	Response Speed	Offset Elimination	Stability	HVAC Application
P only	Fast	No (droop)	Good	Two-position dampers
PI	Moderate	Yes	Good	Most HVAC loops
PID	Fast	Yes	Best	Critical temperature control
PD	Fast	No	Better than P	Rare in HVAC

Tuning Methods:

Ziegler-Nichols Ultimate Cycle Method:
- Set $K_i = 0$ and $K_d = 0$
- Increase $K_p$ until sustained oscillation occurs
- Record critical gain $K_u$ and oscillation period $P_u$
- Calculate tuning parameters:

$$K_p = 0.6 K_u$$ $$K_i = \frac{1.2 K_u}{P_u}$$ $$K_d = 0.075 K_u P_u$$

Lambda Tuning:
- Specify desired closed-loop time constant $\lambda$
- More conservative, reduces overshoot
- Suitable for slow processes (temperature control)

Advanced Control Strategies

Cascade Control:

A master controller’s output becomes the setpoint for a slave controller. This configuration improves disturbance rejection and response time.

graph TD
    A[Zone Temperature Setpoint] --> B[Master Controller<br/>Zone Temperature]
    B --> C[Discharge Air Temperature<br/>Setpoint]
    C --> D[Slave Controller<br/>Discharge Air Temp]
    D --> E[Cooling Valve Position]
    E --> F[Cooling Coil]
    F --> G[Zone]
    H[Zone Sensor] --> B
    I[Discharge Air Sensor] --> D

    style A fill:#e1f5ff
    style B fill:#e1ffe1
    style D fill:#ffe1e1

Example: Room temperature (master) controls discharge air temperature setpoint, which controls cooling valve position (slave).

Feedforward Control:

Measures disturbances before they affect the process variable and adjusts the manipulated variable preemptively.

$$CO(t) = K_{ff} \cdot D(t) + CO_{feedback}(t)$$

Where:

$K_{ff}$ = feedforward gain
$D(t)$ = measured disturbance
$CO_{feedback}(t)$ = feedback controller output

Example: Outdoor air temperature directly adjusts preheat valve position before affecting discharge air temperature.

Reset Schedules:

Supply temperature or pressure setpoints vary based on system conditions to minimize energy consumption while maintaining comfort.

Supply water temperature reset based on outdoor air temperature:

$$T_{supply} = T_{design} - m \cdot (OAT - OAT_{design})$$

Where:

$m$ = reset ratio (°F supply / °F outdoor)
Typical heating: $m = -1.5$ (supply decreases 1.5°F per 1°F OAT increase)
Typical cooling: $m = 0.5$ (supply increases 0.5°F per 1°F OAT increase)

Sensors and Instrumentation

Temperature Measurement

Temperature sensors convert thermal energy into an electrical signal for controller input.

Resistance Temperature Detectors (RTDs):

Platinum element (Pt100, Pt1000 most common)
Linear resistance-temperature relationship
Accuracy: ±0.15°C (±0.27°F) for Class A
Range: -200°C to 850°C
Applications: Critical measurements, laboratories, pharmaceutical facilities

Resistance change with temperature (simplified):

$$R(T) = R_0[1 + \alpha(T - T_0)]$$

Where:

$R_0$ = resistance at reference temperature (100 Ω at 0°C for Pt100)
$\alpha$ = temperature coefficient (0.00385 Ω/Ω/°C for platinum)

Thermistors:

Negative Temperature Coefficient (NTC) most common in HVAC
Nonlinear, high sensitivity
Accuracy: ±0.1°C to ±0.2°C
Range: -40°C to 150°C
Applications: Room sensors, duct sensors, outdoor sensors

Steinhart-Hart equation for NTC thermistors:

$$\frac{1}{T} = A + B \ln(R) + C (\ln(R))^3$$

Where $A$, $B$, and $C$ are calibration constants.

Thermocouples:

Voltage generated at junction of dissimilar metals (Seebeck effect)
Self-powered, no excitation required
Type T (copper-constantan): -270°C to 400°C, ±0.5°C
Type J (iron-constantan): -210°C to 1200°C
Applications: Flue gas, boiler exhaust, high-temperature processes

Sensor Type	Accuracy	Range	Cost	HVAC Application
RTD	±0.15°C	-200 to 850°C	High	Critical processes
Thermistor	±0.2°C	-40 to 150°C	Low	Standard HVAC
Thermocouple	±1°C	-200 to 1200°C	Medium	High temperature
Infrared	±2°C	-50 to 500°C	High	Non-contact measurement

Pressure Sensors

Pressure measurement enables airflow calculation, filter monitoring, and system diagnostics.

Differential Pressure Transmitters:

Measure pressure difference across filters, coils, duct sections, or flow elements. Diaphragm deflection creates capacitance or resistance change.

Airflow calculation from differential pressure:

$$Q = K \sqrt{\Delta P}$$

Where:

$Q$ = volumetric flow rate (CFM)
$K$ = flow coefficient (depends on duct size and element geometry)
$\Delta P$ = differential pressure (in. w.c.)

For averaging Pitot tube arrays:

$$K = 4005 \times A \times C$$

Where:

$A$ = duct cross-sectional area (ft²)
$C$ = calibration coefficient (typically 0.95-1.05)

Pressure Sensor Types:

Gauge pressure: Measures relative to atmospheric pressure
Absolute pressure: Measures relative to perfect vacuum
Differential pressure: Measures difference between two points

Range selection for HVAC applications:

Filter status: 0-2 in. w.c.
Duct static pressure: 0-10 in. w.c.
Airflow stations: 0-0.5 in. w.c.
Hydronic systems: 0-100 psi

Humidity Sensors

Relative humidity sensors measure moisture content in air for humidity control and economizer operation.

Capacitive RH Sensors:

Polymer film capacitance varies with moisture absorption
Accuracy: ±2% RH (typical), ±1% RH (precision)
Range: 0-100% RH
Response time: 15-30 seconds
Applications: Room humidity, duct humidity, outdoor air

Dew Point Sensors:

Chilled mirror (optical detection, ±0.2°C accuracy)
Capacitive polymer (±2°C dew point accuracy)
Applications: Compressed air dryers, critical process control

Relationship between relative humidity and moisture content:

$$W = 0.622 \frac{RH \cdot P_{ws}}{P_{atm} - RH \cdot P_{ws}}$$

Where:

$W$ = humidity ratio (lb water/lb dry air)
$RH$ = relative humidity (decimal)
$P_{ws}$ = saturation pressure at dry-bulb temperature
$P_{atm}$ = atmospheric pressure

Flow Measurement

Differential Pressure Flow Measurement:

Averaging Pitot tubes, flow grids, or venturi sections create pressure drop proportional to flow squared.

$$v = C \sqrt{\frac{2 \Delta P}{\rho}}$$

Where:

$v$ = velocity (ft/min)
$C$ = discharge coefficient
$\Delta P$ = differential pressure (lb/ft²)
$\rho$ = air density (lb/ft³)

Thermal Dispersion Flow Sensors:

Heated element loses heat proportional to mass flow
Suitable for low velocities (50-4000 fpm)
Applications: Fume hood face velocity, lab exhaust verification

Ultrasonic Flow Meters:

Transit time difference between upstream and downstream signals
Non-invasive (clamp-on) or inline installation
Accuracy: ±1-2% of reading
Applications: Chilled water, condenser water, heating hot water

Air Quality Sensors

CO₂ Sensors:

Non-Dispersive Infrared (NDIR) technology
Range: 0-2000 ppm (typical HVAC applications)
Accuracy: ±50 ppm
Applications: Demand-controlled ventilation per ASHRAE Standard 62.1

Outdoor air requirement based on CO₂:

$$V_{oa} = \frac{N \cdot G \cdot 10^6}{C_{space} - C_{oa}}$$

Where:

$V_{oa}$ = outdoor air ventilation rate (CFM)
$N$ = number of occupants
$G$ = CO₂ generation rate (0.005 CFM/person typical office)
$C_{space}$ = space CO₂ setpoint (1000 ppm)
$C_{oa}$ = outdoor CO₂ concentration (400 ppm typical)

VOC Sensors:

Metal Oxide Semiconductor (MOS) sensors
Detects total volatile organic compounds
Applications: IAQ monitoring, demand ventilation in residential

Particulate Matter Sensors:

Laser scattering technology
Measures PM2.5 and PM10 concentrations
Applications: Filter performance verification, IAQ display

Actuators and Final Control Elements

Electric Modulating Actuators

Electric actuators position dampers and valves based on analog control signals (0-10 VDC or 4-20 mA).

Key Specifications:

Torque rating: 35-8000 in-lb for dampers
Actuation time: 30-180 seconds for 90° rotation
Position feedback: 0-10 VDC or 4-20 mA
Power: 24 VAC typical

Spring Return vs. Non-Spring Return:

Spring return: Returns to fail-safe position (open or closed) on power loss
Non-spring return: Maintains last position on power loss

Damper sizing requires torque calculation:

$$T = A \cdot P \cdot L \cdot K$$

Where:

$T$ = required torque (in-lb)
$A$ = damper area (ft²)
$P$ = differential pressure (in. w.c.)
$L$ = damper length perpendicular to shaft (in.)
$K$ = blade shape factor (1.5-3.0)

Control Valves

Control valves modulate fluid flow in hydronic systems.

Valve Characteristics:

Linear valve: Flow proportional to stem position

$$Q = C_v \cdot x \sqrt{\frac{\Delta P}{SG}}$$

Equal percentage valve: Each equal increment of travel produces a percentage change in existing flow

$$Q = Q_{max} \cdot R^{(x-1)}$$

Where:

$Q$ = flow rate (GPM)
$C_v$ = valve flow coefficient
$x$ = valve position (0-1)
$\Delta P$ = pressure drop across valve (psi)
$SG$ = specific gravity
$R$ = rangeability ratio (typically 50:1)

Valve Authority:

Ratio of valve pressure drop to total system pressure drop at design flow:

$$\beta = \frac{\Delta P_{valve}}{\Delta P_{total}}$$

Optimal valve authority: 0.3-0.5 for stable control. Low authority (<0.25) causes unstable control.

Control Sequences and Strategies

VAV System Control

Single-Zone VAV with Reheat:

Cooling mode: Airflow increases from minimum to maximum
Dead band: Airflow at minimum
Heating mode: Reheat valve modulates (airflow remains at minimum)

Zone airflow calculation:

$$CFM_{zone} = \frac{Q_{sensible}}{1.08 \times (T_{supply} - T_{zone})}$$

Where:

$Q_{sensible}$ = sensible cooling load (Btu/h)
$T_{supply}$ = supply air temperature (°F)
$T_{zone}$ = zone temperature (°F)

Static Pressure Reset:

Reduce supply fan static pressure setpoint when VAV dampers are not fully open:

Monitor all VAV damper positions
If no damper >90% open, reduce static pressure setpoint
Typical range: 1.0-2.5 in. w.c.
Energy savings: Fan power ∝ (pressure)^1.5 approximately

Economizer Control Strategies

Air-side economizers use outdoor air for free cooling when conditions are favorable.

Control Strategies Comparison:

Strategy	Enable Condition	Advantages	Disadvantages
Fixed dry-bulb	OAT < 55°F	Simple, no humidity sensor	May introduce high humidity
Differential dry-bulb	OAT < RAT	Prevents heating outdoor air	Ignores humidity
Fixed enthalpy	$h_{oa}$ < 28 Btu/lb	Controls moisture	Requires enthalpy calculation
Differential enthalpy	$h_{oa}$ < $h_{ra}$	Most efficient	Complex, requires sensors

Enthalpy calculation from temperature and humidity:

$$h = 0.24 \times T_{db} + W(1061 + 0.444 \times T_{db})$$

Where:

$h$ = enthalpy (Btu/lb dry air)
$T_{db}$ = dry-bulb temperature (°F)
$W$ = humidity ratio (lb water/lb dry air)

Chilled Water Plant Optimization

Primary-Only Variable Flow:

Modulate pump speed to maintain differential pressure at critical (farthest) coil.

Pump affinity laws relate speed, flow, pressure, and power:

$$\frac{Q_2}{Q_1} = \frac{N_2}{N_1}$$ $$\frac{H_2}{H_1} = \left(\frac{N_2}{N_1}\right)^2$$ $$\frac{P_2}{P_1} = \left(\frac{N_2}{N_1}\right)^3$$

Where:

$Q$ = flow rate
$N$ = pump speed (RPM)
$H$ = head (pressure)
$P$ = power

Reducing pump speed to 70% design:

Flow: 70% of design
Pressure: 49% of design
Power: 34% of design (66% energy savings)

Building Automation System Architecture

Modern BAS systems employ distributed intelligence with hierarchical control.

graph TD
    A[Enterprise Server<br/>CMMS, Analytics, Reporting] --> B[Building Management Server<br/>Graphics, Alarms, Trends, Scheduling]
    B --> C[Network Backbone<br/>BACnet/IP, Ethernet]
    C --> D[Building Controller<br/>Supervisory Control]
    C --> E[Building Controller<br/>Supervisory Control]
    D --> F[Application Controller<br/>AHU-1]
    D --> G[Application Controller<br/>Chiller-1]
    E --> H[Application Controller<br/>AHU-2]
    F --> I[Field Devices<br/>Sensors, Actuators]
    G --> J[Field Devices<br/>Sensors, Actuators]
    H --> K[Field Devices<br/>Sensors, Actuators]

    style A fill:#e1f5ff
    style B fill:#ffe1e1
    style C fill:#fff4e1

System Layers:

Management layer: Operator interface, graphics, trending, alarming
Automation layer: Building controllers, supervisory control, scheduling
Field layer: Application controllers (AHU, VAV, chiller), local control loops
Device layer: Sensors, actuators, variable frequency drives

BACnet Protocol

ASHRAE Standard 135 defines BACnet (Building Automation and Control Networks) as an open, interoperable protocol.

BACnet Objects:

Analog Input (AI): Temperature sensor, pressure sensor
Analog Output (AO): Valve position command, damper position command
Binary Input (BI): Status switch, alarm contact
Binary Output (BO): Start/stop command, enable/disable
Analog Value (AV): Calculated value, virtual point
Schedule: Time-based control sequences

BACnet Services:

ReadProperty: Request current value
WriteProperty: Change setpoint or command
COV (Change of Value): Automatic notification when value changes
Trend Log: Historical data storage

Energy Optimization Strategies

Optimal Start/Stop:

Calculate equipment start time to reach setpoint exactly at occupancy:

$$t_{start} = t_{occupancy} - \frac{T_{desired} - T_{current}}{R_{heating}}$$

Where:

$t_{start}$ = equipment start time
$R_{heating}$ = heating rate (°F/hour), learned from previous starts

Demand Limiting:

Shed non-critical loads when approaching peak demand threshold:

Monitor building electrical demand
Compare to target limit (e.g., 90% of contract demand)
Shed loads in priority order: lighting, plug loads, HVAC setback
Restore loads when demand decreases

Trim and Respond Chiller Sequencing:

Adjust chiller loading based on efficiency curves:

Load most efficient chiller first
“Trim”: Incrementally load/unload chillers
“Respond”: React to rapid load changes
Minimize total kW/ton across all operating chillers

Reference Standards

ASHRAE Standard 135: BACnet Protocol
ASHRAE Guideline 13: Specifying Building Automation Systems
ASHRAE Guideline 36: High-Performance Sequences of Operation for HVAC Systems
NIST SP 800-82: Guide to Industrial Control Systems Security
ISA-95: Enterprise-Control System Integration

Effective HVAC control transforms mechanical systems into intelligent, responsive environments that adapt to occupant needs while minimizing energy consumption and operational costs.