Introduction to Direct Digital Control Systems
Purpose of this Guide:
The purpose of this guide is to describe, in generic terms, the various architectures, hardware components and software associated with Direct Digital Control (DDC) systems. To accomplish this goal, a generic framework of the various components and configurations used in current DDC systems has been defined. This framework is used as a yardstick for several DDC manufacturers so readers may compare the relative features and benefits.
Due to the complexity and proprietary nature of DDC systems, it has become difficult to stay current with the designs, installations, operation and maintenance of DDC systems. This guide was developed specifically to help building owners and consulting/specifying engineers with these issues.
What is an Energy Management System?
For the purposes of this guide, an energy management system (EMS) is defined as a fully functional control system. This includes controllers, various communications devices and the full complement of operational software necessary to have a fully functioning control system. This guide addresses approximately twenty of the DDC vendors who serve the institutional and commercial marketplace in the United States. Vendors who supply a complete line of all the necessary hardware and software are included. This guide does not cover specialty markets (retail grocery, hotels), nor does it cover industrial or process controls.
What is Control?
The process of controlling an HVAC system involves three steps. These steps include first measuring data, then processing the data with other information and finally causing a control action. These three functions make up what is known as a control loop. An example of this process is depicted in Figure 1.
Basic Control Loop
The control loop shown in Figure 1 consists of three main components: a sensor, a controller and a controlled device. These three components or functions interact to control a medium. In the example shown in Figure 1, air temperature is the controlled medium. The sensor measures the data, the controller processes the data and the controlled device causes an action.
The Figure 1 could be an example of a pneumatic or electronic control system, where the controller is a separate and distinct piece of hardware. In a DDC system, the controller function takes place in software as shown in Figure 2.
The sensor measures the controlled medium or other control input in an accurate and repeatable manner. Common HVAC sensors are used to measure temperature, pressure, relative humidity, airflow stateand carbon dioxide. Other variables may also be measured that impact the controller logic. Examples include other temperatures, time-of-day or the current demand condition. Additional input information (sensed data) that influences the control logic may include the status of other parameters (airflow, water flow, current) or safety (fire, smoke, high/low temperature limit or any number of other physical parameters). Sensors are an extremely important part of the control system and can be the first, as well as a major, weak link in the chain of control.
The controller processes data that is input from the sensor, applies the logic of control and causes an output action to be generated. This signal may be sent directly to the controlled device or to other logical control functions and ultimately to the controlled device. The controllers function is to compare its input (from the sensor) with a set of instructions such as setpoint, throttling range and action, then produce an appropriate output signal. This is the logic of control. It usually consists of a control response along with other logical decisions that are unique to the specific control application. How the controller functions is referred to as the control response. Control responses are typically one the following:
- Proportional (P only)
- Proportional plus Integral (PI)
- Proportional plus Integral plus Derivative (PID)
Controlled Device or Output
A controlled device is a device that responds to the signal from the controller, or the control logic, and changes the condition of the controlled medium or the state of the end device. These devices include valve operators, damper operators, electric relays, fans, pumps, compressors and variable speed drives for fan and pump applications.
Two-position control compares the value of an analog or variable input with instructions and generates a digital (two-position) output. The instructions involve the definition of an upper and lower limit. The output changes its value as the input crosses these limit values. There are no standards for defining these limits. The most common terminology used is setpoint and differential. The setpoint indicates the point where the output pulls-in, energizes or is true. The output changes back or drops-out after the input value crosses through the value equal to the difference between the setpoint and the differential.
Two-position control can be used for simple control loops (temperature control) or limit control (freezestats, outside air temperature limits). The analog value can be any measured variable including temperature, relative humidity, pressure, current and liquid levels.
Time can also be the input to a two-position control response. This control response functions like a time clock with pins. The output pulls-in when the time is in the defined on time and drops out during the defined off time.
Figure 3, shows an example of two-position control in a home heating system, where the thermostat is set to energize the heating system when the space temperature falls below 70 F and turn off when the temperature rises to 72 F in the space. This is an example of a setpoint of 70 F with a two-degree differential.
Floating control is a control response that produces two possible digital outputs based on a change in a variable input. One output increases the signal to the controlled device, while the other output decreases the signal to the controlled device. This control response also involves an upper and lower limit with the output changing as the variable input crosses these limits. Again, there are no standards for defining these limits, but the terms setpoint and deadband are common. The setpoint sets a midpoint and the deadband sets the difference between the upper and lower limits.
When the measured variable is within the deadband or neutral zone, neither output is energized and the controlled device does not change - it stays in its last position. For this control response to be stable, the sensor must sense the effect of the controlled device movement very rapidly. Floating control does not function well where there is significant thermodynamic lag in the control loop. Fast airside control loops respond well to floating control. An example of floating controls is shown in Figure 4.
A proportional control response produces an analog or variable output change in proportion to a varying input. In this control response, there is a linear relationship between the input and the output. A setpoint, throttling range and action typically define this relationship. In a proportional control response, there is a unique value of the measured variable that corresponds to full travel of the controlled device and a unique value that corresponds to zero travel on the controlled device. The change in the measured variable that causes the controlled device to move from fully closed to fully open is called the throttling range. It is within this range that the control loop will control, assuming that the system has the capacity to meet the requirements.
The action dictates the slope of the control response. In a direct acting proportional control response, the output will rise with an increase in the measured variable. In a reverse acting response, the output will decrease as the measured variable increases. The setpoint is an instruction to the control loop and corresponds to a specified value of the controlled device, usually half-travel. An example is shown in Figure 5.
In a proportional control system, the value of the measured variable at any given moment is called the control point. Offset is defined as the difference between the control point and the desired condition. One way to reduce offset is to reduce throttling range. Reducing the throttling range too far will lead to instability. The more quickly the sensor feels the effect of the control response, the larger the throttling range has to be to produce stable control.
Proportional plus Integral (PI) Control
PI control involves the measurement of the offset or error over time. This error is integrated and a final adjustment is made to the output signal from the proportional part of this model. This type of control response will use the control loop to reduce the offset to zero. A well set-up PI control loop will operate in a narrow band close to the setpoint. It will not operate over the entire throttling range (Figure 6).
PI control loops do not perform well when setpoints are dynamic, where sudden load changes occur or if the throttling range is small.
Proportional plus Integral plus Derivative (PID) Control
PID control adds a predictive element to the control response. In addition to the proportional and integral calculation, the derivative or slope of the control response will be computed. This calculation will have the effect of dampening a control response that is returning to setpoint so quickly that it will overshoot the setpoint.
PID is a precision process control response and is not always required for HVAC applications. The routine application of PID control to every control loop is labor intensive and its application should be selective.
Definition of Direct Digital Control (DDC)
DDC control consists of microprocessor-based controllers with the control logic performed by software. Analog-to-Digital (A/D) converters transform analog values into digital signals that a microprocessor can use. Analog sensors can be resistance, voltage or current generators. Most systems distribute the software to remote controllers to eliminate the need for continuous communication capability (stand-alone). The computer is primarily used to monitor the status of the energy management system, store back-up copies of the programs and record alarming and trending functions. Complex strategies and energy management functions are readily available at the lowest level in the system architecture. If pneumatic actuation is required, it is accomplished with electronic to pneumatic transducers. Calibration of sensors is mathematical; consequently the total man-hours for calibration are greatly reduced. The central diagnostic capabilities are a significant asset. Software and programming are constantly improving, becoming increasingly user-friendly with each update.
Benefits of DDC
The benefits of direct digital control over past control technologies (pneumatic or distributed electronic) is that it improves the control effectiveness and increases the control efficiency. The three main direct benefits of DDC are improved effectiveness, improved operation efficiency and increased energy efficiency.
DDC provides more effective control of HVAC systems by providing the potential for more accurately sensed data. Electronic sensors for measuring the common HVAC parameters of temperature, humidity and pressure are inherently more accurate than their pneumatic predecessors. Since the logic of a control loop is now included in the software, this logic can be readily changed. In this sense, DDC is far more flexible in changing reset schedules, setpoints and the overall control logic. Users are apt to apply more complex strategies, implement energy saving features and optimize their system performance since there is less cost associated with these changes than there would be when the logic is distributed to individual components. This of course assumes the user possesses the knowledge to make the changes.
DDC systems, by their very nature can integrate more easily into other computer-based systems. DDC systems can integrate into fire control systems, access/security control systems, lighting control systems and maintenance management systems.
Improved Operational Efficiency
Operational improvements show the greatest opportunity for efficiency improvements in direct digital controls. The alarming capabilities are strong and most systems have the ability to route alarms to various locations on a given network. The trending capabilities allow a diagnostic technician or engineer to troubleshoot system and control problems. They also allow the data to be visualized in various formats. These data can also be stored and analyzed for trends in equipments performance over time.
Run-times of various equipment can be monitored and alarms/messages can be generated when a lead/lag changeover occurs or if it is time to conduct routine maintenance.
The off-site access/communication capability allows an owner/operator to access their system remotely. Multiple parties can also be involved in troubleshooting a problem. The control vendor, design engineer and commissioning authority can use these features to more efficiently diagnose and visualize problems.
Increased Energy Efficiency
There are many energy-efficient control strategies employed in pneumatic logic that can be easily duplicated in DDC logic. Due to the addition of more complex mathematical functions (easily obtained in software), there are many additional energy-efficient routines that can be used with DDC.
Strategies such as demand monitoring and limiting can be more easily implemented with DDC systems. The overall demand to a facility can be monitored and controlled by resetting various system setpoints based on different demand levels. If a DDC system is installed at the zone level, this could be accomplished by decreasing the requirement for cooling on a zone-by-zone basis.
By storing trends, energy consumption patterns can be monitored. Equipment can also be centrally scheduled on or off in applications where schedules frequently change.
Elements of a Direct Digital Control System
The word points is used to describe data storage locations within a DDC system. Data can come from sensors or from software calculations and logic. Data can also be sent to controlled devices or software calculations and logic. Each data storage location has a unique means of identification or addressing.
Direct digital controls (DDC) data can be classified three different ways - by data type, data flow and data source.
Data type is classified as digital, analog or accumulating. Digital data may also be called discrete data or binary data. The value of the data is either 0 or 1 and usually represents the state or status of a set of contacts. Analog data are numeric, decimal numbers and typically have varying electrical inputs that are a function of temperature, relative humidity, pressure or some other common HVAC sensed variable. Accumulating data are also numeric, decimal numbers, where the resulting sum is stored. This type of data is sometimes called pulse input.
Data flow refers to whether the data are going into or out of the DDC component/logic. Input points describe data used as input information and output points describe data that are output information.
Points can be classified as external points if the data are received from an external device or sent to an external device. External points are sometimes referred to as hardware points. External points may be digital, analog or accumulating and they may be input or output points. Internal points represent data that are created by the logic of the control software. These points may be digital, analog or accumulating. Other terms used to describe these points are virtual points, numeric points, data points and software points.
Global or in-direct points are terms used to describe data that are transmitted on the network for use by other controllers. These points may also be digital, analog or accumulating.
Analog input points typically imply an external point and represent a value that varies over time. Typical analog inputs for HVAC applications are temperature, pressure, relative humidity, carbon dioxide and airflow measurements. Typical analog outputs include control signals for modulating valve positions, damper positions and variable frequency drive speed.
Typical digital inputs for HVAC applications represent the status (example: whether or not the motor is running) of fans, pumps, motors, lighting contactors, etc. A temperature high limit is considered a digital input because, although it is monitoring an analog value (temperature), the information that is transmitted to the controller is a digital condition (whether or not the temperature has exceeded a defined value). Digital outputs are typically motors or other devices that are commanded on or off. Digital outputs include fans, pumps, two-position (solenoid) valves, lighting contactors, etc.
A true analog output (voltage or current) is a varying DC voltage or milliamp signal that is used to drive a transducer or controlled device. Another type of analog output is pulse width modulation (PWM). PWM is accomplished by monitoring a timed closure of a set of contacts. The amount of time the contacts are closed is proportional to a level of performance for the controlled device.
There are basically three common approaches used to program the logic of DDC systems. They are line programming, template or menu-based programming and graphical or block programming.
Line programming-based systems use Basic or FORTRAN-like languages with HVAC subroutines. A familiarity with computer programming is helpful in understanding and writing logic for HVAC applications.
Menu-driven, database or template/tabular programming involves the use of templates for common HVAC logical functions. These templates contain the detailed parameters necessary for the functioning of each logical program block. Data flow (how one block is connected to another or where its data comes from) is programmed in each template.
Graphical or block programming is an extension of tabular programming in that graphical representations of the individual function blocks are depicted using graphical symbols connected by data flow lines. The process is depicted with symbols as on electrical schematics and pneumatic control diagrams. Graphical diagrams are created and the detailed data are entered in background menus or screens.
System architecture is the term used to describe the overall local area network or LAN structure, where the operator interfaces connect to the system and how one may remotely communicate to the system. It is the map or layout of the system.
The network or LAN is the medium that connects multiple intelligent devices. It allows these devices to communicate, share information, display and print information, as well as store data. The most basic task of the system architecture is to connect the DDC controllers so that information can be shared between them.
A control loop requires a sensor to measure the process variable, control logic to process data, as well as calculate an instruction, and a controlled device to execute the instruction. A controller is defined as a device that has inputs (sensors), outputs (controllable devices) and the ability to execute control logic (software) (Figure 7).
Communications between devices on a network can be characterized as peer-to-peer or polling. On a peer-to-peer LAN, each device can share information with any other device on the LAN without going through a communications manager (Figure 8).
The controllers on the peer-to-peer LAN may be primary controllers, secondary controllers or they may be a mix of both types of controllers. The type of controllers that use the peer-to-peer LAN vary between manufacturers. These controller types are defined later in this section.
In a polling controller LAN, the individual controllers can not pass information directly to each other. Instead, data flows from one controller to the interface and then from the interface to the other controller (Figure 9).
The interface device manages communication between the polling LAN controllers and the higher levels in the system architecture. It may also supplement the capability of polling LAN controllers by providing the following functions: clock functions; buffer for trend data, alarms, messages; and higher order software support.
Many systems combine the communications of a peer-to-peer network with a polling network. In Figure 10, the interface communicates in a peer-to-peer fashion with the devices on the peer-to-peer LAN. The polling LAN-based devices can receive data from the peer-to-peer devices, but the data must flow through the interface.
Controllers can be categorized by their capabilities and their methods of communicating (controller-to-controller). In general, there are two classifications of controller - primary control units and secondary control units
Primary controllers typically have the following features:
- Real-time accurate clock function
- Full software compliment
- Larger total point capacity
- Support for global strategies
- Buffer for alarms/messages/trend & runtime data
- Freeform programming
- Downloadable database
- Higher analog/digital converter resolution
- Built-in communication interface for PC connection.
Secondary controllers typically have the following features:
- Not necessarily 100% standalone
- Limited software compliment
- Smaller total point count
- Freeform or application specific software
- Typically lower analog-to-digital converter resolution
- Trend data not typically stored at this level
- Typical application is terminal equipment or small central station equipment.
The next critical element in the system architecture is an operator interface. Operator interfaces are required to:
- See data
- Program the system
- Exercise manual control
- Store long term data
- Provide a dynamic graphical interface.
There are five basic types of operator interfaces. They include:
- Desktop computers which act as operator workstations
- Notebook computers which act as portable operator workstations
- Keypad type liquid crystal displays
- Handheld consoles/ palmtops/ service tools
- Smart thermostats
Desktop computers are centralized operator workstations where the main function is programming, building and visualizing system graphics; long term data collection; and alarm and message filtering.
Notebook computers may connect to the LAN through a communication interface that stands alone or is built into another device. The notebook computer connected to the LAN at a particular level may not have the same capability as a computer connected to the LAN at a higher level.
Keypad liquid crystal displays typically are limited to point monitoring and control. They may have some limited programming capability, such as changing a set point or time schedule.
Handheld consoles, palmtops and service tools are proprietary devices that connect to primary controllers or secondary controllers. Typically they allow point monitoring and control, controller configurations (addressing and communication set-up), and calibration of inputs and outputs.
Smart thermostats are sensors with additional capabilities. They connect to secondary controllers and have a service mode to allow for point monitoring, control and calibration. They also have a user mode that allows point information to be displayed, setpoint adjustment and an override mode.
The communications interface shown in the Figure 11 is a communication interface device. It provides the path between devices that do not use the same communications protocol. This includes computers, modems and printers.
It may be a stand-alone component or it may be built into another device as shown in Figure 12.
Each communications interface on Figure 12 may:
- Translate protocol
- Provide a communication buffer
- Provide temporary memory storage for information being passed between the network and the external PC, modem or printer (mailbox function)
Larger System Architectures
When systems become larger than the capacity of a single sub-network, a higher level of architecture is added to allow the use of multiple sub-networks.
The site LAN wide area network or WAN is used to connect multiple sub-networks and site computers. Multiple sub-networks can be connected to a single site LAN/WAN that allows information sharing between devices on different sub-networks (Figure 13). There may be a limitation on the number of site computers. The site LAN/WAN may include routers if TCP/IP is used. If no routers are used, the protocol can be totally proprietary. If TCP/IP is used, the EMS site LAN/WAN can be the information system backbone within the facility or between facilities.
Multiple site computers can be added to the site LAN/WAN. They can connect the site LAN/WAN via a communications interface, which may be a router. Site LAN/WAN computers can send and receive information from the entire system. Information can be received by each of the site computers, but can not be subsequently shared from one computer to another. Sub-network computers may only be able to see their own sub-network.
Site LANs allow multiple computers to communicate with each other. They may use commercially available computer network software and hardware. Messages, alarms and other data can be re-routed to other computers on the primary site LAN. Information stored in other computers can be remotely accessed. This includes graphics, programming and stored trend and operational data.
Some vendors combine multiple functions into a single device. In the following system architecture, Figure 14, the communication interface is built into the primary controller. A peer-to-peer LAN or sub-network is connected directly to the device.
In Figure 15, the key component in the system consists of a communication interface, a primary controller and an interface to the secondary polling network.
The addition of a site LAN allows a system to gain size in terms of the number of devices that are served, but in some applications, the location of the devices, rather than the number of devices, is the bigger challenge. In this situation, modem-based communication is used to expand the geography of the system.
Auto-Answer/Auto-Dial System Architecture
In auto-answer/auto dial systems, a specialized communication interface is substituted which introduces a modem and phone lines into the standard architecture. These communication interfaces are made with built-in modems or use external commercial modems. Auto-answer/auto-dial configurations are used to provide monitoring and access to remote buildings. They are used where traditional direct-wiring methods are impractical; and where central site monitoring is desired; or where remote access to controllers is desired.
In an auto-answer/auto-dial system, the central communications interface may call the individual sites or vice versa. Information and data can be passed to and from the layer above the central communications interface (Figure 16).
The auto-answer/auto-dial LAN architecture is typically used by installations with multiple facilities where control and monitoring needs to be centralized. Multiple LANs are used to maintain the groupings of devices, or to separate controllers into defined groups.
Multiple Dial LAN Support
In a systems architecture, the local sites have the ability to call an alternate communication interface, if the primary is not available (Figure 17).
One-Way Dial System Architecture
One-way dial systems, Figure 18, are typically used to enable system owners to access their systems from a remote location, such as their home. It is used where auto-dial monitoring is not required. It can also be used by the installation and service company or by the commissioning authority to troubleshoot and program from remote locations. One-way dial can also be used to dial into remote site LANs or sub-networks.
Two modems are required, one located at the remote computer and one at the system site. Typically, the DDC operating software must be installed on the remote computer.
Communication between two different devices controlling equipment, requires a common protocol, a common communication speed and known data formatting. Vendors build their devices around these criteria, so communication between devices by the same manufacturer is routine.
Third Party Interfaces
In many installations, it is desirable for a proprietary building DDC system to communicate with other proprietary DDC systems controlling pieces of equipment. Examples would include a building DDC system and a chiller DDC system (Figure 19) or a fume hood DDC system. Communication between the two systems will require an interface or gateway, due to different proprietary protocols, communication speeds and data formatting.
The gateway or interface translates protocol between the two proprietary systems. The proper operation of the gateway is dependent on the continued use of the specific revised levels of software on both systems. It typically requires the support of the manufacturer at the corporate level to implement and cooperation between the manufacturers. In addition, the costs can vary widely.
In the DDC world, there are the three classifications of protocols: closed protocol, open protocol and standard protocol.
A closed protocol is a proprietary protocol used by a specific equipment manufacturer. An open protocol system uses a protocol available to anyone, but not published by a standards organization. A standard protocol system uses a protocol available to anyone. It is created by a standards organization.
An open system is defined as a system that allows components from different manufacturers to co-exist on the same network. These components would not need a gateway to communicate with one another and would not require a manufacturer specific workstation to visualize data. This would allow more than one vendors product to meet a specific application requirement.
The sole use of an open or standard protocol does not guarantee that a DDC system will be an open system. A manufacturer has the ability to use open or standard protocols, yet create a closed system, thus continuing a building owners dependence on a single manufacturer. This can be accomplished by using unique communication speeds, unique data formatting and by not adopting the full range of an open protocol.
Note: A building owner/engineer should thoroughly research a manufacturers claim of an open system.
BACNET is a standard protocol published by a standards organization (American Society of Heating, Refrigerating and Air-conditioning Engineers or ASHRAE). It is a specification for a protocol. DDC vendors create a communication protocol that complies with this specification.
BACNET is a relatively complex standard. The standard defines protocol implementation conformance statements (PICS) that define different levels of compliance. A given vendor may or may not support the level required for a given application. In other words, a vendor could meet a very low level of compliance and be BACNET-compatible. The key question is, At what level?
In Figure 20 the chiller control units DDC will communicate with the building DDC system if each has a BACNET gateway and their PICS match.
If a vendor states their product is native BACNET, they are using the BACNET protocol in lieu of a proprietary protocol on their LAN. In Figure 20, a native BACNET building system would be able to communicate to the chiller control DDC with one less gateway.
An overlay system is a high-end workstation that communicates with multiple manufacturers proprietary EMS systems. An overlay system supplier creates drivers to talk to the different systems. The vendors must have a cooperative relationship and revision control is important for continued successful use. The workstation typically displays data, allows manual control and setpoint changes, and handles alarms and messaging. Any detailed editing of the control sequence will still require knowledge of the underlying proprietary software.
The Echelon Corporation has created an open protocol that uses a standard processor and a set of standard transceivers, which allows components from different manufacturers to co-exist on the same LAN. The protocol is available to anyone and is called LONTALK. A unique chip is required for any device that uses LON. Standard network variable formats have been established to allow the transfer of data from one device to another regardless of origin.
Presently, various vendors are competing to become the defacto standard for the network database structure. The network database is a map of the components and the relationship of the data moving between them. The operator workstation needs this structure to visualize the data.
Software suppliers providing the software for the operator workstation may be independent of those providing the software for the database structure and the EMS vendors.