✕

Following the 26th United Nationsʼ Climate Change Conference and the numerous pledges from countries and companies for net-zero emissions, the acceleration towards decarbonization is tangible. Process manufacturers have increased urgency to improve their environmental impact. Escalating fuel prices along with market demand for greener products also have added pressure to evaluate alternative sources of energy for heating processes.

When taking action to evaluate these alternative energy sources, the industry must consider the reality of energy production sources worldwide. Global energy use for heating accounts for 129 EJ (exojoules), with 79 EJ used by industry and 50 EJ used by households. Approximately 75 percent of this demand is met with fossil fuels, with 40 percent of the energy supplied by natural gas and 20 percent each by oil and coal, respectively. Sustainable renewable energy in the industrial sector only accounts for 10 percent of the total 79 EJ consumed.^[1]

Process heating accounts for more than two-thirds of the total global industrial energy consumption. Half of that demand occurs at temperatures levels below 752°F (400°C). For example, industry sectors such as food and beverage, textiles and pulp-and-paper each have a heat demand below 482°F (250°C) in 60 percent of their operations.

Depending on heating demand, which varies depending on the industry and geography, the case for using renewables to provide energy for heating or cooling technologies differs. Process stages like water preheating, washing, boiler makeup water and preheating air for drying or combustion can benefit from renewable sources of energy. such as solar. Renewable sources are well suited to supplying heat in the range from 482 to 752°F (200 to 400°C).

One of the biggest challenges of the energy transition for process manufacturing is adapting existing technologies and assets to use renewable energy. This necessitates connecting to the large volumes of operational and equipment data made accessible by Industrial Internet of Things (IIoT) projects and then analyzing the data to achieve valuable insights for driving towards a more sustainable energy future.

It is reasonable to assume that the energy transition will happen in phases, considering current dependencies on fossil fuels, different geographical economies and country needs. Changes will take time, but there are ways to accelerate the transition. This is where operational time-series data collected via IIoT projects along with analytics can help.

For years, operational leaders in the process industries have been implementing IIoT projects to provide real-time access to time-series data. This data is stored in process historians, analyzed and used to optimize processes. Ideally, the results of these analyses also are shared with suppliers and service providers to provide further insights into the efficiencies of their operations.

Prediction models created using advanced analytics applications help companies understand and reduce energy use. Image provided by Seeq

One obstacle with this use of data is that it is often stored in silos. For example, in a fiber line, engineers might only look at the data that affects the fibers: woodchip preheating, cooking process and pulp drying. Rarely do they look at information from the systems delivering the steam needed to preheat, cook and dry the fibers. So, the questions are, how do process manufacturers merge related but isolated pieces of information? How do they make the data readily accessible to engineers for better decision-making to reduce energy consumption and, therefore, the carbon footprint of their processes?

Going forward, the solution likely will not be found in spreadsheets. They are a general-purpose tool and are not designed for time-series data analysis of large amounts of data. Although many engineers still use spreadsheets, their use creates cumbersome tasks filled with barriers. Often, the limits of the program may prevent them from analyzing data in the broader business context necessary to improve performance.

Analytics Provide Insights

Recognizing the drawbacks of spreadsheets, many process manufacturers are now empowering their frontline employees with self-service advanced analytics applications that provide a streamlined approach and interface for accelerating improvements to their production outcomes. These types of applications allow process manufacturers to integrate and connect to disparate data sources, providing a better picture of the data generated by IIoT and other projects. Using automated data cleansing and contextualization, advanced analytics applications empower engineers to see relationships and correlations among data for more efficient analysis.

Manufacturers across many process industries — from power generation to chemicals — are already using self-service advanced analytics to reduce energy consumption and carbon footprint in their heating processes. Lessons learned from the solar and wind power generation companies using advanced analytics applications, like in the use cases discussed below, can be leveraged by process manufacturers starting their energy transition.

By filtering operational data using meaningful conditions describing different operating modes, advanced analytics application regression models can accurately predict energy consumption. They also can be used to predict process upsets before they occur such as upcoming required boiler maintenance, or heat exchanger downtime for cleaning based on fouling conditions, for example.

As companies formulate their energy transition strategies and explore renewable sources for their energy mix, wind and solar are viable options. In these scenarios, advanced analytics applications can help optimize wind turbine rampup. For solar power generation, advanced analytics applications can help identify soiled panels or axis faults by using predefined formulas.

The following use cases show how companies are using advanced analytics applications to create insights from IIoT and other data.

Process Optimization Using Dynamic Energy Models

Allnex, a specialty chemical company, leverages analytics technology to create energy models that are used to continuously improve their operations.

A leading challenge with energy consumption patterns is that related data is noisy and, therefore, difficult to model in a spreadsheet. And, when models are created, they are often simplistic and rarely updated, making them obsolete quickly.

Prediction models. created in advanced analytics applications — using either linear or non-linear regressions of the total steam demand based on instrumentation and equipment data — empower engineers to isolate the impact of specific equipment like an individual steam valve. Engineers can build their own models quickly and establish regular reviews using dashboards, process models and other data.

Using this approach for more than three years, the company has achieved a 30 percent increase in steam recovery and a 15 percent reduction in steam consumption. The greenhouse emissions prevented by these improvements equate to reducing vehicle miles driven by nearly 38 million per year.

Predictive Maintenance at Scale

Another chemicals manufacturer, Covestro, implemented a predictive maintenance approach, but, in their case, for heat exchangers. Using operational data from its analytics suite, the company calculated the heat transfer coefficient of its heat exchangers. With these values, it could predict when fouling in the equipment would degrade production quality.

Company engineers used linear regression analyses in the software to predict when to remove built-up solids from the heat exchanger. In the past, these types of calculations took days to run, but they can be performed quickly with the software.

Wind Turbine Ramp-Up and Quantifying Revenue Losses

A major energy company uses analytics software to optimize turbine ramp-up as market demand changes. This is done by comparing potential versus actual power to identify a ramp-up period and to quantify the cost and footprint of the time each wind power turbine is not operating. This real-time production potential (RTPP) varies depending on the age of components, wind turbulence and other factors.

Using analytics software, the energy company modeled the RTPP potential based on periods when the turbine was communicating data, producing power and not curtailed by the grid operator. At 5-min intervals, the grid operator was provided with the company’s RTPP calculation based on wind strength and direction. This information improved the grid operator’s ability to maintain system reliability and balance load. It also allowed the company to account for the revenue losses of power not generated during the curtailed and ramp-up time.

In this scenario, the energy company was experiencing curtailments roughly once every two days at a cost of several hundred dollars per day, adding up to lost revenue of roughly $3,000 per month per windfarm. Using analytics software, they created a scalable and accurate way to quantify the period of curtailment and ramp-up to ensure accurate chargeback to the grid operator.

Today, the energy company calculates its commercial losses during ramp-up across its entire fleet of wind turbines, enabling them to pass charges to the grid operator. It was able to recapture more than $30,000 per windfarm per year thanks to accurate chargeback calculations. The company is scaling these calculations across additional fleets of turbines.

This information improved the grid operator’s ability to maintain system reliability and balance load. It also allowed the company to account for the revenue losses of power not generated during the curtailed and rampup time. Image provided by Seeq

Condition-Based Maintenance of Solar Panels

With solar energy generation using heating processes, soiled panels or axis faults reduce total power output. Developing models to account for these kinds of power losses is difficult with spreadsheets. Or, it requires specific software with no capability for tuning or ad hoc analysis of plant performance.

With the analytics software, periods that exemplify optimal performance such as when panels are clean and there are minimal tracking deviations from setpoint can be identified. These periods can be combined to create a training window. A model of plant performance based on the identified training window and plant meteorology instruments then can be developed to monitor the impact of axis tracking faults and soiling losses.

The developed plant model allows for continuous monitoring of soiling and axis tracking losses. It also can provide a dollar quantification report to determine when the monetary cost of maintenance is justified by the losses associated with required downtime to address issues.

Conclusion

The use of operational data and advanced analytics is beneficial in diverse ways. It can help companies understand their energy consumption and ways to optimize it. It can help extend the life of critical pieces of equipment and cut maintenance costs. It also can provide a way to measure the carbon footprint and emissions from the use of fossil fuels, helping companies justify the integration of renewal sources of energy for suitable heating processes.

The amount of energy consumed by the industrial sector will continue to increase, so improvements must be made to meet the global goals for net-zero emissions by 2050. The power industry is leading the way in terms of efficiently operating renewable sources of energy, but other industries, including pulp-and-paper and metals and mining, are not far behind.

The use of digital technologies such as IIoT and self-service analytics software can help engineers to make data-driven decisions regarding the best ways to diversify the energy mix.