How to build a “needle-finding” machine Today, the rules are changing. The entry of the IoT world into buildings has not yet been noticed by everyone.

Krzysztof Jaczewski, Key Account Manager, Global Control 5

Someone would say that to use a strong magnet is enough. What if the needle is not made of metal? Should we burn the haystack to get our needle?  

Data surrounds us, and the amount grows very fast within every second. Sensors, machines and other devices which are inbuilt in buildings produce a lot of chaotic and unsorted information. In addition, data on the use of commercial buildings is still often proprietary and difficult to access. Unfortunately, the construction industry is more conservative than others and has lagged behind. As a result, often overused term BIoT means only a mixture of wheat and chaff. In this case, collected data causes only significant costs of information storage.

First of all, the data set should be cleaned and sorted in order to make it easier to find valuable and reliable information. To use a more colloquial turn of phrase, we have to make more needles in a haystack than hay. It is obvious that in this case, the use of magnet or matches will be ineffective. So that is where Project Haystack comes.
    
Project Haystack is an open source initiative to develop tag naming conventions and taxonomies for modelling and easily utilizing data from various sources for analysis and decision making. The metadata methodology enables understanding of data and its interpretation, both by people and machines. Members of the Project Haystack community have created software plugins to allow different systems to communicate in the haystack protocol. As an example, we can use the nHaystack module for Niagara® systems.

It means that Haystack is not just a tag definition but also a protocol which brings more information than the BACnet standard. nHaystack is a Niagara 4 module that enables Niagara stations to act as servers or clients in the Project Haystack format, via a RESTful protocol. When acting as a server, nHaystack automatically generates standard tags for all ControlPoints.

iSMA-B-MAC36NL as a controller and Haystack device in one

The above solution also allows you to convert serial RS-485 buses to high-speed Ethernet.  Thanks to the use of the MAC36NL (or Tridium JACE) the old systems, unsuitable so far for IoT, which are based on legacy, closed protocols such as N2, C-Bus, TAC Xenta etc., now can be unlocked and “internet enabled.”

We have just received the first part of our data machine that makes the data useful and understandable for the IoT world. This was a great step forward but not the last one.

With so much data out there, humans can’t review them all. Buildings also have some level of intelligence. The sad fact is, that is not enough these days.

As a result of that, the facility manager’s actions are often limited only to responding to events that have already occurred.
Today, it is possible to get more from building data and ultimately, make better decisions. Our data collection must be analyzed in order to obtain information that is in it, but it is not directly visible. To do this, we need an advanced analytical tool.

Let’s use SkySpark as the substance of the second part of our data machine.

SkySpark by SkyFoundry is revolutionary technology which transforms building data into a business solution. This tool automatically analyzes data and finds information which allows to:

• detect the most important factors responsible for the issues, deviations, equipment faults and incorrect execution of control sequences
  
• find the best setting values based on records from different subsystems
• predict ahead failures and errors (predictive maintenance)
• estimate the future consumption of energy and its costs
• perform normalization against weather, building size, baseline data and any other self-prepared normalization parameters
• find profiles and patterns that are specific to a given system
• generate views on detected issues (automatically)
• benchmark with the multi-site comparison.

[an error occurred while processing this directive]SkySpark is an open and flexible analytical platform. It means that each user can implement his own rules which fit a specific object. Possibility to create your own APIs and dashboards, allows you to customize that tool to your own needs.

However, the strongest point of this framework is the built-in machine learning module. With this functionality, it is easy to spot patterns and solve problems that would have otherwise taken years to solve.

SkySpark includes a huge library of built-in analytics functions (over 500) and set of applications to visualize data and analytics result which allows going ahead with the project without having to start from creating complex models.

Today, the rules are changing. The entry of the IoT world into buildings has not yet been noticed by everyone. IoT does not tolerate imposed artificial barriers. Some big players in the building automation market will be surprised by the fact that the strategy of blocking the customer with their technology has not been effective.

 

About the Author

Krzysztof Jaczewski, Key Account Manager, Global Control 5, is responsible for the development of the iSMA brand in Poland and northern Europe.  Previously, for twenty years, associated with the international concern Honeywell.

He introduced the CentraLine brand to the Polish market and the Saia Burgess Controls (SBC) brand to the local structures of the company. Krzysztof has many years of experience in the market of HVAC products, automation systems and BMS/BAS.