INSTALLATION AND CONFIGURATION
How do I supply power to the wireless modules?
The sensor modules are battery-powered, with battery life expected to be on the order of two years. When battery power is running out, they automatically notify the server so that the batteries can be changed. The control modules that connect to physical devices such as VFDs and valves require 120V AC power and communicate to the server over your company's network.
What hardware do you recommend for uninterruptible power supply (UPS)?
How much time is required for your system to complete its start-up learning cycle and develop an intial behavior profile?
The AI engine typically requires 20 minutes per AHU/CRAC to complete its initial learning process. During this time all fans are spun up to maximum speed, and each unit is then sequentially throttled back to 50% speed for 20 minutes. This allows the thermal impact on every sensor node to be determined, as it relates to each cooling source. This is then fed into the dynamic model that evolves over time.
NETWORKING AND COMMUNICATIONS
What wireless technology does Vigilent use for its network?
Vigilent energy management systems are built upon a sophisticated, wireless mesh network using technology developed by Dust Networks®, the leader in industrial wireless networking. With over a 70% share of that market, Dust Networks technology is a standard that has been adopted by Emerson, General Electric and others who rely on its resilience, reliability and ultra-low-power capabilities. This implementation is designed for the most demanding industrial applications, in harsh environments where packet delivery is critical.
How do the sensor nodes communicate with the server? Will I need to run cable in our facility?
All sensor nodes use Dust Networks® wireless technology to communicate with the server. This network is what is known as a "mesh network", known for its low-power consumption and ability to automatically self-configure. Every node in the network is not only a source, but also a repeater, in fact. The nodes wirelessly connect to a gateway, which is then connected via an Ethernet cable to the Vigilent server.
What communication protocols are used by your system?
Our systems uses native BACnet for machine control, and native web services for communications and administrative control. For other protocols, such as SNMP, a protocol translator is used. Interfacing to CRAC/CRAH units is accomplished via a direct software BACnet connection (or through a protocol translator), or wirelessly with a 0-10V DC analog signal.
How does the Vigilent server communicate with cooling units?
The server communicates with AHUs and CRACs using your local area network connection, using standard protocols such as BACnet.
Will we need to purchase a server to install your software?
No, Vigilent intelligent energy management systems include either a rackmount or workstation server, preloaded with all of the necessary software.
ROBUSTNESS AND RELIABILITY
What happens when there is a power outage?
Our server and gateways are installed on UPS power, so they will never go down due to a power outage. Our CRAC control kits will lose power when the CRAC loses power, but immediately switch to battery power and continue to communicate with our gateway and server. Our wireless sensors are always on battery power. So our entire system will stay up and running during a power outage.
If it's a short outage, the CRACs that were on before the outage will turn on as soon as power is restored. CRACs usually have a staged start sequence that can be configured to minimize the inrush current that might otherwise overload the panel that supplies the CRACs.
If it's a longer outage, and the room temperatures have risen, our system will ensure that additional CRACs are turned on to deliver needed cooling as soon as power is restored.
Is your system able to self-diagnose failure conditions in its components?
Yes. There is an always-running monitor process that constantly verifies that all system/network components are functioning properly, including all sensor/control nodes. It will set alarms or initiate restarts when necessary.
Does your system have a failsafe? What will happen if that occurs?
Yes. There is extensive capability throughout our system to protect against a wide range of failure modes. All wireless control modules have a SafetyNet™ technology implemented in firmware to automatically set analog outputs to pre-configured (and configurable) levels, should they ever lose contact with the network or from the server heartbeat. The AI engine and server has a guard mode that will lower set points, raise fan speeds and turn cooling units on if an alert threshold is exceeded. There is also a manual control prominently featured in the Administrator Console that resets the entire system, if needed, turning all units on and throttling fan speeds to maximum. In addition, every device we control (i.e. VFDs, AHUs, CRACs, etc.) have their own manual override that can be used to manually control its performance, when needed.
Isn't it enough to only monitor temperatures in my facility? I want to use this information to control temperature manually, rather than have it under computer control.
Monitoring temperatures with a sensor network is a good start, and essential, in fact. However, it's only a start. For effective control of your cooled environment, it's important that cooling become a managed resource. In a typical data center or building, there may be hundreds or even thousands of sensor nodes and many AHUs or CRACs under control. The sheer volume of data from these devices almost instantly becomes overwhelming, and tracking and making changes in real-time is impractical, if not impossible. A managed environment needs to rely on sophisticated processing technology to handle this dynamic control, typically with sub-second response to balance the interactions between these many nodes.
SYSTEM PERFORMANCE & USE
Should I be concerned about your system allowing my facility to get too hot by over-aggressively reducing cooling output?
No, you do not need to be concerned about that. Our system is designed to work in manual mode until a complete functional test has been performed, after which time it switches to automatic mode. You can easily set high and low thresholds to govern the system's behavior. And the system has failsafes that automatically shift to "all on" cooling mode when network, controller or sensor outages are detected.
Wouldn't slowing down airflow have a negative effect on server fan usage and the internal temperatures of the servers?
It does seem counterintuitive that less airflow wouldn't have a negative impact, but we've found time and time again that server fan speeds and internal server temperatures aren't affected as long as inlet rack temperatures are maintained. Since the Vigilent system controls cooling units to maintain inlet air temperatures, reducing the airflow in a data center doesn't have an effect on the internal server temperatures. In our experience, we've actually seen inlet air temperatures decrease as a result of less mixing due to the reduced airflow. Typically, the Vigilent system controls to the ASHRAE recommended thermal range of 64.4 - 80.6F, although operators can adjust these limits.
Can different weightings be assigned to each input from wireless sensors?
Yes, different setpoints can be assigned to each sensor to meet your requirements.
What machine capabilities is your system capable of controlling?
Our energy management systems can control any point addressable via either a BACnet or 0-10V DC analog interface, including CRAC setpoint, on/off status, VFD speed and valve position.
What are the "learning algorithms" you refer to?
Our systems are built around a core artificial intelligence engine that resides in the server. It contains sophisticated algorithms and AI technology that learns over time. This begins from the time the system is commissioned and an initial behavior profile is developed during a multi-hour period of perturbation, where responses are provoked and responses measured. It continues throughout the regular use of the system, learning as it simultaneously controls the devices in its network. We believe that closed-loop control is essential to effectively implement dynamic cooling, and our M3™ technology is the power behind this achievement.
What is used as the control element in determining thermal loads: Cold aisle temperature? Or return temperature to a CRAC?
The control element is the cold aisle inlet air temperature. The return air setpoint can be limited, if there is an SLA on that variable, but we generally find that most sites place little emphasis on return air temperature once uniform cold aisle temperatures have been achieved.
Will your system save energy even if the set point for inlet temperature is too narrow?
If cooling capacity exceeds the load, then there are opportunities to save even if the inlet temperature set point is low and the range is narrow. Of course, the opportunity is greater if the upper limit is higher and the range is wider. The wider the range in setpoints, in fact, the less set point overages will occur, which means more opportunity to save energy.
How does your system determine the appropriate range for temperature adjustment?
These limits are determined through empirical observation, as it throttles fan speed up or down between off and maximum speed. The minimum level can be set, however, to whatever is desired (e.g. for a minimum fan speed of 50%, the system will establis the zero point at 51%).
What trigger initiates an automatic restart of AHUs/CRACs?
For DX units with no VFDs, the trigger is a cold aisle temperature exceeding a predefined threshold. The AI engine uses its predictor module to determine which air handler to start or stop. Start-stop is one of the primary mechanisms by which we achieve energy savings for installations with DX units that have staged compressors. Out-of-range performance can be avoided by lowering the start threshold and limiting the time between stops and restarts.
What is the minimum offtime for AHUs/CRACs?
The minimum offtime can easily be configured using the Administrator Console.