From steam-powered mechanical production to the introduction of electricity and mass assembly line in manufacturing, and most recently, the rise of computerized means of production using robotics, we have come a long way. According to industry experts, currently, the world is going through its fourth industrial revolution, otherwise known as Industry 4.0 (I4.0). Industry 4.0 has now become the umbrella term for a range of big data and manufacturing concepts, advanced robotics, and other emerging technologies such as the cloud and cognitive computing, Industrial Internet of Things (IIoT), digital twins and predictive analytics. Amongst these, digital twins—one of the critical aspects of I4.0—has gained significant attention in the past few years with many enterprises leaning toward a virtual environment or simulation that can replicate their existing processes. This helps to monitor process performance, reduce maintenance costs, train their workforce in a risk-free environment, and maximize the returns on investment. MEL Group, an industrial automation solution provider from Atlanta, GA, fulfils the need for such simulation solutions and integrated controls among organizations in power, oil and gas, pulp and paper, and various other industries.

With over 30 years of experience, MEL group optimizes performance, improves reliability, and enhances efficiency throughout their customers’ plant lifecycles. The company’s in-house digital twin (or triplets) and simulation solution combines digital power plant simulation (DPPS), AI (neural nets)-based decomposition and FFTD (Fast Fourier Transform Digitization), and data analytics to monitor the performance of different types of equipment. Igor A. Berkovich, CEO of MEL group, says, “Our THINK Digital Twins solution, Thermal-Hydraulic Integrated Network (THINK) combines the power of DPPS, AI-based sound decomposition, FFTD, and data analytics to monitor and predict the performance of equipment across an organization.”

THINK is a first principle simulator that performs acoustic monitoring through microphones and decomposes a sound into its fundamental modes via FFTD and neural nets (TensorFlow). This decomposition facilitates instantaneous comparison of performance records of earlier waveforms to determine if the equipment is functioning correctly. As a result, it can simulate the machines at a microscopic level to provide a granular view of processes. The tool is fitted with a GUI and an easy-to-use visual interface, and it can perform non-equilibrium two-phase flow in real or faster than real-time simulation, using finite element type solution methods. Additionally, THINK has components for boiling and condensation, which enables it to model heat transfer, turbines, boilers, and combustors. Besides, a collation of the collected data and digital data history of a plant can be used to train neural nets on power histories, FFT decomposition, equipment health, and operations. Acoustic, vibrational, and thermal resonance identification can reduce downtime during replacement, predict equipment failures and diminish the amount of time taken for startup and shutdown of a power plant. Phillip Wang PE, President and group member of the managing board of MEL, says, “With THINK, power plants can significantly improve their operational efficiency and reduce maintenance cost.” The digital twin solution offers an unparalleled simulation of power plant processes that helps companies to stay ahead of the competition. Additionally, organizations can use MEL digital twins for “Load follow.” The electrical grid or any energy production distribution is a “Load” on the production plant. Load following for electricity producers is becoming more difficult as companies use a different means of electricity/ energy production. A solution to this can be training a neural net to “load follow” to aid the operators.

This allows forecasting the load to avoid going offline as these types of shutdowns can cost the electricity/ energy producers millions of dollars in downtime.

At present, MEL group is working with some of the nation’s largest utility enterprises to provide digital twins to various power generation facilities for operator training, control system upgrades, and operational improvement studies. Two of these key customers include Dominion Energy and SIEMENS. The companies also are developing digital twins for one of the largest power and desalination complexes in the Middle East, in conjunction with a major power-generation equipment supplier and provide solutions for global energy Market.

Our digital twin simulation solution, Thermal Hydraulic Integrated Network (THINK) combines the power of DPPS, AI-based sound decomposition, FFTD, and data analytics to monitor and predict the performance of equipment across an organization

Industry thought leaders are predicting the use of AI tools such as the MEL digital twins to examine data trends and search for patterns. MEL’s future roadmap is closely aligned with these recent industry trends in the energy sector. Scott Lucas PhD, Chief Digital Technology Officer (CDO) says, “MEL’s immediate plans include enhancing the THINK-based simulation to produce turbulence, sound (acoustics), and heat (infrared) to help train TensorFlow (training data set) on normal and off-normal plant events at different power levels.” Therefore, the company is currently modifying THINK to address the growing needs of the market and drive Industry 4.0 forward.

Neal Wunderlich, President and group member, Wunderlich—Malec Engineering (WME), said that he foresees considerable growth for Smart Power Generation and Hybrid Energy businesses. WME, an ENR top 200 engineering firm with over 30 local US offices and over 450 engineers provides MEL and their proprietary and complimentary Power Digital Twin applications and local engineering resources and support. The group’s strategic Vision 2025 is to be in the Top 5 Power Plant Digital Technology Solution Providers and Energy System Integrators.

For more information regarding MEL please visit and for WME solutions please visit