Slash your energy bill by 35% using Device-level Analytics and Machine Learning

Author: : Saurabh Jhamb, Anil Chhikara

0 33

United Nations Industrial Development Organization (UNIDO) data points out 30-35% of average Industrial energy wastage in emerging economies such as India, China. Overall, Indian manufacturing sector consumes ~200 Billion Units of energy growing at 10% CAGR while output is comparatively stagnate at 2%, as per Energy Ministry Report 2013-2017 Govt. of India. Thanks to advancements in miniaturization, connectivity, data storage & smart sensors, Internet of Things (IoT) has opened a huge opportunity for smarter energy utilization worldwide.

Recent developments in IoT and energy management systems have enabled visibility into real-time device-level energy consumption. Such operational data holds incredible value related to operational and energy efficiency of production sites.

Processing this data to classify, predict and compare the operational states of different field assets is being tipped as a game changer. It allows the organization to not only reduce energy consumption by identifying leakages as they happen but also calibrate processes & dependencies towards greater productivity, while engaging in preventive maintenance at the earliest pre-signs of malfunction.

In the past, energy managers had to periodically review and analyze energy consumption at device, block or plant level, accompanied with redundant shutdowns for routine-checks and/or proactive maintenance. With increasing number of devices and need for greater details, processing and implementation of conclusions, field monitoring was a herculean task. It was therefore crucial to automate this analysis and provide managers with actionable insights.

IoT enables possibility of self-powered&maintenance-free wireless sensors (patent-pending design) &Advance Analytics, such as URJA, maximizing ROI by equipping managers with real-time insights about their plants’ energy consumptions, and derived intelligent insights to avoid unexpected failures at first sight of malfunctioning or a projected catastrophe.Smart sensing makes virtually any device such as manufacturing equipments, HVAC components, Lighting systems, a connected smart device to enable smart maintenance through Robust analysis of sensor data combined with asset & production data

Textile Factory, Field Asset over time

Fig. 1 – Textile Factory, Field Asset over time

Asset Operational Characteristics
Different devices have different operational states that varies with production and timing of the day / month of the year. Simple systems such as lighting solutions have on/off states, while other more complex equipments can cycle, idle with different combinations of compressors/motors in them. Even machines with exact same make and model will share different states, therefore manufacturer specifications are expressed as a range (or not to exceed values) as opposed to absolute numbers.

Steel Forging Unit, Field Asset over time

Fig. 2 – Steel Forging Unit, Field Asset over time

Such variation reinforces need for device-level analysis. While energy consumption patterns for one asset may reflect normal or even ideal performance, for other asset it may reveal an anomaly, potentially leading to future breakdown.

Baselining & Profiling of Assets
With granular data at periodic intervals collected in realtime by sensors and communicated wirelessly to the cloud every 5 seconds, coupled with machine learning enabled data science algorithms, developed by experienced engineers.

With an initial baseline from manufacturer guidelines, data from machine’s performance over next few weeks is pumped into algorithms to generate follow-on profiles specific at device & time periods. The earliest signs of deviations from baseline and profiles are flagged as it happens. Advanced algorithms are applied to extract the averages and standard deviations of timeseries states, enabling determination of the device state, in real time.

Fig. 3 - Oil & Gas OEM, Field assets –Operational Timeseries View (URJA)

Fig. 3 – Oil & Gas OEM, Field assets –Operational Timeseries View (URJA)

Climb the Fast Growth Curve
Next generation energy monitoring systems such as URJA offer intelligently automated data science procedures for in-depth visibility to facility management and maintenance professionals. Scalable deployment of such sensors and analytics systems allow for increased operational and energy efficiency of production sites and at a lower cost.

Indian manufacturing industry burns around 192 billion units of energy annually to make a contribution of <5% to Indian GDP and has < 1% penetration of energy standards such as ISO 50001. To bridge this gap, device-level energy management systems, such as URJA, offer huge opportunity to lift your business above competition, favoring your bottom-line and growth.

After all, time has come to turn the tables around on energy spend by hearing what your equipments have to say.

- Advertisement -

Leave A Reply

Your email address will not be published.