Skip navigation
All Places > All Things PI - Ask, Discuss, Connect > Blog > 2017 > August

          This post will contain an overview of an intern project designed to show the value of the PI System in a Facility context. The project was undertaken at the Montreal OSIsoft office, by two interns Georges Khairallah and Ali Idrici – both of whom are studying Mechanical Engineering.




OSIsoft Canada ULC - Montreal is located on one of the floors of an old building in Montreal downtown. The building management team does not have any building management system (BMS) to manage energy and thus they do not have data regarding where most of the energy is being used. In fact, only the lighting systems can be controlled by the different offices. Other control systems, such as the HVAC system, are controlled by various master switches that regulate the entire facility.

OSIsoft Montreal’s office has expanded throughout the past several years with a wider working area and two distinct spaces located on the same floor.

Despite not having a BMS, OSIsoft Montreal would like demonstrate that they can measure and manage their office’s energy consumption from lighting and track human presence with their real-time operational data software, the PI System, in order to lower energy usage and reduce false alarms.


Project Description

OSIsoft Montreal is looking to reduce energy consumption from unnecessary lighting as well as reduce the number of false alarms caused from absence of occupancy presence information. Currently, there are no systems in place to monitor and track real-time usage or presence. However, the company is looking to implement a number of IIoT (Industrial Internet of things) devices and sensors across their workspace, then send the collected data to one PI System, where real-time analyses will be generated and displayed.


Summarizing the two critical business issues defined by OSIsoft Montreal:


Scope Overview – Situation 1: Energy Efficiency

  • Critical Business Issue

      Unable to cut down the energy consumption from unnecessary lighting.

  • Problem/Reason:

Tedious to walk across all work areas before leaving the office. Lack of visibility of every room’s lighting conditions.

Need real time visibility on every room’s lighting status with metrics (energy usage, costs) across all working areas to determine where appropriate action is required.

o   Delta (Benefit):

       - Need to cut travel time before heading out of the office by 80%

- Need to save 25% of energy consumption from lighting

  • Target Date:

Mid of August 2017– Presentation to OSIsoft Montreal Team


Scope Overview – Situation 2: Alarms Management and Delayed Departure

  • Critical Business Issue:

      Unable to reduce the number of false alarms.

  • Problem/Reason:

Disturbs other colleagues with unnecessary agitation, and a 10 minute delay to deal with the alarm company and building surveyor if triggered. The absence of occupant presence information leads to inaccurate assumptions on whether to activate the alarm at time of departure.

Need real time occupant presence monitoring to activate the alarm system only when one person remains.

o   Delta (Benefit):

                            Need to reduce false alerts to 1/4 of current occurrences (currently: once a quarter)

  • Target Date:

Mid of August 2017– Presentation to OSIsoft Montreal Team


While OSIsoft Montreal does currently have a workaround to their business issues, they do not have a time saving solution, nor an effortless process to locate lightning statuses in each room, and no easy method to detect human presence in the office at all times.



Project Goals

The business goals of this project are to leverage OSIsoft’s real-time data infrastructure, the PI System, in order to:

  • Increase awareness on lighting consumption to the Montreal team
  • Highlight the zones which contribute the most to the energy bill
  • Track presence in the office at any time of the day
  • Indicate when there is only one person in the office to let them activate the alarm and easily locate where remaining lights must be shut off
  • Automate the procedure necessary before heading out of the office



How Are We Collecting the Required Data?


The collection of light intensity data is achieved by an Electric-Imp.

The collection of human presence data uses an Omron D6T 1x8 low-resolution thermal camera temperature sensor, that is connected to a DragonBoard410c.

Both of these devices send the raw data as secured HTTPs Requests to a REST EndPoint in a Docker located in an Azure Cloud.


Project Architecture

The following figure represents the target data flow for the generation of lighting and human presence reports:


      Figure 1 – OSIsoft Montreal System Architecture Used For This Project



Using Analytics on PI AF to Analyze Incoming Raw Data


o   Data coming from Electric imps:

ON/OFF light status is obtained by performing a periodic comparison of the current light intensity to a threshold value. Later on, energy consumption can be obtained knowing how many watts each light is consuming when turned on. Other statistics such as daily and weekly light energy consumption are also computed on Analytics.


     Figure 2 – Asset Framework display of lighting status analysis


o   Data coming from Thermal sensor:

Temperature profiles are retrieved 4 times a seconds and each temperature profile contains 8 values. The following graph displays the raw data coming from the thermal sensor as someone walks past the sensor and then comes back.


    Figure 3 – Temperature profile variation with respect to time (PI Vision display)


Using Pi analytics, an attribute named “Polarity” (see yellow curve on fig. 4) is computed for every temperature profile. Polarity quantifies how shifted to one side or another the temperature values are. For simplicity’s sake, the remaining of the analysis focuses on polarity variation.


     Figure 4 – Polarity variation with respect to time (PI Vision display)


When the polarity switches continuously from negative to positive, it means someone walked passed the sensor . Conversely, when the polarity switches from positive to negative, someone passes by the sensor in the other direction. Further analytics can be performed on the Polarity attribute to extract it’s sign only.


     Figure 5 – Polarity sign variation with respect to time (PI Vision display)


Switching polarity sign from -1 to +1 triggers a positive increment in the counter whereas polarity sign switching from +1 to -1 triggers a negative increment.




Using PI Vision 2017 to Display the Processed Data


o   Dashboard for Lights:


      Figure 6 – Live Lighting status for every room in the Montreal office (PI Vision display)


      Figure 7 – Light status for every Electric Imp (PI Vision display)


o   Dashboard for human counter :


      Figure 8 – Counting number of individuals in the Montreal office (PI Vision display)

Filter Blog

By date: By tag: