Skip to content

Posts tagged ‘artificial intelligence’

Applying Artificial Intelligence to Commercial Refrigeration

Charles Larkin | Director of Data and Analytics, Cold Chain

Emerson’s Commercial and Residential Solutions Business

Over the past decade, artificial intelligence (AI) has become an ever-present aspect of everyday life. From e-commerce and smartphone functions to social media to modern industry, AI and advanced machine learning (ML) algorithms analyze continuous streams of data to derive predictive insights and optimize performance. Although these data science techniques are not new to commercial refrigeration, food retail and foodservice operators have been relatively slow to embrace AI’s vast potential. I recently participated in an ACHR The News article where we discussed AI’s barriers to adoption and how Emerson is helping to prove the value of AI to its customers.

AI is not a new concept for the food retail and foodservice industries. Many prominent retailers are already using AI techniques in customer-focused areas of their businesses, such as personalizing their consumer rewards and loyalty programs. In fact, several leverage in-house data science teams to champion these initiatives. But when it comes to turning AI’s focus toward refrigeration, very few have the domain expertise or experience applying AI to other critical facility systems — which can be significantly more complex and require a completely different knowledge base.

Another barrier to implementing AI in commercial refrigeration is the challenge of aggregating different sources and types of operational data into a useable format. Many food retailers already have some type of control system in place. Since different control system vendors collect and process data differently, it can be difficult to ensure the accuracy and consistency of the data. In addition, many vendor systems have proprietary constraints that don’t allow data to be shared easily.

Although the industry recognizes the potential of AI to deliver value in commercial refrigeration, food retailers and their servicing teams still have questions about its role in their operations. Demonstrating the value of AI across a wide range of food retail applications will be necessary in order to remove these doubts.

Engaging in proof-of-concept trials

At Emerson, one of the most important jobs we have is to provide the expertise and data science programs to build the business case for AI’s potential value to our customers. As a refrigeration controls, components and equipment manufacturer, we are focused on developing AI-enabled controls and integrated equipment that can deliver numerous benefits for operators and contractors alike.

Currently, we are engaging some of our customers in short-term, proof-of-concept trial periods. This gives us opportunities to demonstrate how our AI and ML solutions can integrate with their operations and deliver the potential for long-term, continuous refrigeration performance improvements. Once they see how quickly we’re able to deliver value and offer a return on investment (ROI), they’re much more interested in exploring a longer-term engagement.

The core of AI and ML technologies resides within the system control devices, which are typically incorporated into the equipment itself. By capturing data from sensors, modern equipment controls can perform a variety of key system optimization functions — from system fault protection and diagnostics to performance management and event scheduling. And in many instances, we can enable these capabilities without having to perform a significant retrofit.

Many of our existing customers already have a data-rich infrastructure — including sensors, controls and modems — that we can tap into and begin delivering insights. We often recommend installing additional sensors, which is relatively inexpensive compared to a full retrofit.

Adding up the advantages

As for the advantages that AI offers, not only can it deliver significant reliability and longevity benefits to commercial refrigeration equipment, but it can also address an ever-expanding variety of store operator and contractor concerns. For operators, we’re building data models that help them to optimize food quality and safety and reduce waste — in applicable case types and perishable food categories.

For contractors, we’re developing ML algorithms that are designed to detect asset health or condition issues. Over time, this data will allow retailers and their contractors to:

  • Implement more predictive maintenance programs
  • Reduce energy costs
  • Keep assets running in optimum condition

Today, Emerson is leveraging AI and ML to optimize critical aspects of our customers’ operations. Our solutions utilize sensors that deliver data to powerful control devices — such as the new Lumity™ E3 supervisory control — and integrate with advanced, cloud-based software. By leveraging the deep domain expertise of our refrigeration engineers, we’re able to create data models that maximize refrigeration performance and help our customers to achieve a variety of key food retail and foodservice objectives.

Building Blocks of Artificial Intelligence for HVACR

 

JohnWallace_Blog_Image John Wallace | Director of Innovation, Retail Solutions

Emerson Commercial & Residential Solutions

Emerson is applying our expertise in commercial refrigeration and AC toward building predictive models for a variety of applications and architectures, a foundation for the emerging artificial intelligence technologies in the HVACR industry. I recently discussed our work in ACHR News magazine, “The Impact of Artificial Intelligence on HVACR.” You can read the full article here.

The building blocks of artificial intelligence (AI)-enabled equipment and systems in HVACR are already well in development: next-generation sensors and controllers, increasingly sophisticated predictive analytics, and machine-to-machine learning (M2M) software, cloud data storage and the growing implementation of the internet of things (IoT). These tools are already providing opportunities to improve comfort, save energy, reduce maintenance costs and extend equipment life, all while helping end users better manage their operations.

But integrating these tools into true AI solutions — data- and algorithm-driven applications that will enable systems and equipment to learn and automatically perform critical tasks without human intervention — is a challenge that will require a deeper understanding of the complexities of equipment, HVACR architectures and building systems.

At Emerson’s innovation centers and in customer field trials, we’re tackling this challenge head on — but methodically. Rather than simply throwing more technologies into the mix, we’re leveraging our deep refrigeration domain expertise to simplify complexities and uncover insights into the industry’s most common refrigeration scenarios. We are in the process of understanding how deeply AI could be implemented into equipment and buildings, and how effectively it could help solve the industry’s biggest challenges.

As I stated in the article, Emerson is researching how some newer AI-related technologies can be utilized for more advanced services, such as detecting problems faster and pinpointing which actions need to be taken. For example, we are already incorporating some AI-related technologies into equipment when we learn they add value, such as sensors that warn of refrigerant leaks in supermarket refrigerants.

However, delivering on the promised value of AI — autonomous predictive analysis and control of HVACR equipment and even entire building environments — will require more than simply installing connected sensors and devices, transmitting clouds of data, and creating libraries of algorithms. As the automobile industry has learned, building a self-driving vehicle is a far more complex undertaking than it appears. This example is important to keep in mind when considering the inherent complexities and diversity of commercial refrigeration applications.

A typical commercial refrigeration system consists of many interdependent components — often from multiple suppliers — with potentially diverse data sources. The proliferation of system architectures and refrigerants has resulted in an ever-expanding diversity of applications. This makes data modeling and defining predictive algorithms difficult. At Emerson, we believe that the development of AI in HVACR will grow as an iterative process, via data processing performed at the equipment level — with tighter integration of sensors and controllers providing richer data to cloud- and IoT-based services. These services provide both real-time alerts and historical trends of equipment performance under a given set of conditions — including indications of potential failures.

These data sets are the foundation of the next level of AI, enabling predictive maintenance models that will anticipate problems and maintain optimum conditions across a defined range of variables. Reaching that point will require generating sufficient historical data detailing the operation, failures and problems of equipment and components. And while much of this data is available today, new sensors may also be required to provide more advanced predictive capabilities.

Relatively speaking, the use of AI in HVACR equipment and controls is still in its infancy. But we’re working to accelerate its advancement to help our industry reap its potential benefits, including: improved reliability, energy savings, prolonged asset life and, of course, predictive analytics. As more AI-related technologies arrive in the HVACR space, we’ll start to fully understand the significant benefits and valuable data they are capable of providing.

%d bloggers like this: