W. DANIELS, TrendMiner, Limburg, Belgium
As companies in the manufacturing process industry explore sustainable ways to manage their energy use, engineers are finding that the deadline for drastically reducing carbon emissions is approaching rapidly.
Sustainability has reached the top of every organization’s agenda. Process manufacturers find that social pressure—even from their investors and employees—can be as strong as regulatory measures to reduce carbon dioxide (CO2) emissions. There is an increased awareness among all generations regarding the negative effects of carbon emissions on climate change. Governments have put teeth to regulations by setting deadlines for CO2 elimination. While the European Union strives to be emissions-neutral by 2060, China has set 2030 as its target. At the time of this publication, more than 40 countries had a carbon emissions tax.
In addition to the environmental initiatives, and with fewer than 30 yr until the deadline, companies must become sustainable to stay ahead of the curve. According to Deloitte, 56% of oil and gas companies tie their executive compensation to decarbonization efforts. Investments in solar, wind and other energy alternatives now top $350 B/yr. Additionally, during times of crisis (e.g., the COVID-19 pandemic), companies have learned that a diversified package of energy assets helps achieve energy needs.
About 20% of emissions savings come from improving energy efficiency. Process experts must clearly visualize emissions and their associated processes to make these improvements. They also must be able to flag excess emissions and investigate the root causes of such anomalies. Engineers must be capable of monitoring emissions, flaring and energy consumption as they identify areas for operational improvement. Organizations can go the extra mile by making step changes, such as electrification, photovoltaics, carbon capture and using green energy.
The low-hanging fruit of emissions reduction hides inside operational improvements. Energy efficiency goals are linked directly with reduced operational expenses, and improving processes lowers energy consumption. Tackling operational improvements also enables organizations to take an incremental approach. Companies do not face large investment costs, and improvements typically rely on process knowledge already in place within the organization.
Achieving operational improvements becomes easier with advanced self-service analytics.a An advanced analytics platform empowers process experts to make data-driven decisions that improve energy efficiency. As a result, engineers make substantial contributions to an organization’s sustainability efforts and bottom line.
The following examples include two situations where self-service analytics improved operations, increased energy efficiency and helped engineers meet corporate sustainability goals.
Eliminating a source of flaring. A dashboard following the overall emissions status of an asset showed a sudden increase in CO2 emissions. Although it is not a critical event on its own, it could lead to more severe incidents if the root cause of the flaring is not determined. A more frequent occurrence of similar events could result in a significant loss.
Engineers used their advanced analytics platform to review 1 hr of time-series flow data going to a flare. While the trendlines for such a feed should be constant, the data selected shows a sudden spike. Process experts need to determine why the spike is occurring and how frequently it previously occurred. Determining how often they occur is the best first step, as this often can help determine the root cause.
Process experts can use a pattern recognition search to find periods when spikes occur. After selecting the period where the analysis is most important—a stable period followed by a sudden spike—the advanced analytics solution determined a similar period had occurred 38 times before, in this case. Engineers can overlay the flaring incidents and compare the layers to determine if the behavior follows the same pattern.
To learn more about the spike, engineers can calculate the maximum value of the flares to determine which flare was the most severe. They can determine this by adding the calculation on the same tag within the advanced analytics platform. The calculations can then be run for all 38 incidents to rank the flaring events in severity. With the most severe events as a starting point, process experts can move on to determine the root cause of the flare.
In the advanced analytics platform, engineers can use the high-throughput recommendation engine to search for correlations in process behavior in all other tags. In this case, it found one tag with a 94% correlation with the spikes in flaring. It was an early indicator, making it more likely to be the root cause of the CO2 spike. More specifically, the identified correlation showed that a valve closed quickly right before each flare spike, as shown in FIG. 1. Process experts confirmed this correlation equated to causation and made changes to the control system to prevent the valve from closing quickly under normal operation.
The company reduced overall emissions and resolved a safety concern by correcting the root cause of the flare spikes. This also helped prolong the valve’s life and resolved problem that could interfere with an engineer’s other work.
This case nicely illustrates the power of self-service analytics. Through a simple sequence of steps, process experts can leverage their knowledge to identify root causes and make data-driven decisions quicker and easier.
Operational improvement of an integrated solar combined cycle. An advanced self-service analytics platform can also assist in retrofitting and step changes. Frequently, process experts have limited in-house experience with new technology. Companies must quantify how new technology will affect existing processes and must be able to demonstrate the added value.
With advanced self-service analytics, engineers can build knowledge faster, more easily quantify process influences and expedite the rollout of process monitors to more users.
An example of this can be found in retrofitting a combined-cycle plant—separate gas and steam turbines—to an integrated solar combined cycle (ISCC). After converting to an ISCC, engineers must determine the effectiveness of the step-change investment. Additionally, they must monitor the performance and identify any root causes of recent production losses.
In this case, engineers wanted to see if the new ISCC was living up to expectations and determine if any anomalies needed to be corrected or losses in production that would need to be addressed.
Engineers viewed a dashboard that measured the plant’s key performance indicators (KPIs) and noticed that the facility’s production had decreased (FIG. 2). Process experts wanted to determine if the decrease resulted from retrofitting the combined cycle to include a solar turbine.
Engineers began by loading data that showed power production and consumption over time. They then searched for data periods that showed the process before and after the retrofitting. Because the project had taken place within a year of the analysis, process experts could use the 1-yr mark to compare process behavior.
The company’s advanced analytics platform included a comparison table that showed various production values. Since the addition of the solar generator, power consumption had decreased, but power production had increased. This meant its retrofitting was successful, but the information did not explain why production decreased. Despite the success of the retrofitting, engineers still had to determine the root cause of this anomaly.
In one of the dashboards, engineers can find other clues to help them determine production loss. Process experts can see the production output of each power turbine within the ISCC. From the values on this dashboard, process experts determined that the steam turbine was not producing as much power as it should, and the solar field had a performance issue.
Engineers tackled the steam problem first. They used the advanced self-service analytics platform to view the steam production. The addition of the solar turbine created a trendline with peaks and valleys. Process experts can use the daily production average to smooth out the trendline and make it fit for analysis.
Formulas can then be built to calculate various measurements. Engineers can look for similar periods to see when the steam turbine power production fell below normal operation and compare the data. In this case, engineers determined that the steam turbine started losing power before adding the solar turbine.
Process experts then began to search for the root cause of the power loss from the steam turbine. They used the advanced analytics system’s recommendation engine to suggest root causes. They learned that the steam turbine pressure increased when the blowdown decreased. Furthermore, they determined that there was a 2-d window when the blowdown started to decrease before the steam turbine increased. This also warned that the process would shutdown entirely within 3 wk.
When searching for similar incidents and potential root causes, it often helps to filter out periods with known but unrelated deviations from normal operations. This is called contextual data, which process experts can integrate with time-series data to enhance their analysis.
With a decrease in blowdown determined to be the root cause of the power loss, engineers used a value-based search to discover when the blowdown was below a certain threshold for at least 1 d. They discovered this was unique to this anomaly, so they decided to save the information as contextual data for future use and to activate a monitor to receive immediate warnings for future decreases in blowdown.
There was one more peculiarity on the dashboard of the advanced analytics platform: a cooling problem in the new solar field. Using a linear temperature graph over time, engineers determined the southwest field of solar panels reached 675°K. They used this figure to find similar periods of interest where the temperature was that high or greater.
Engineers then created a Gantt chart to show relevant process events by the asset. Similar events are displayed apart from normal operations, which creates a concise graphical overview of plant performance. Furthermore, monitors were set up to provide early warnings when solar field process limits are exceeded. This allows faster resolution of process upsets and higher process efficiency in general.
Operational efficiency equals energy savings. Reducing energy consumption and improving energy efficiency are project assurances to a retrofit. However, opportunities for improvement do not lie solely within the area of reducing carbon dependency.
An advanced self-service analytics platform offers opportunities for operational improvements at every stage of production. When operational efficiency is improved, energy efficiency is improved. Companies will find they use less energy and reduce the discharge of CO2 emissions, and they can meet their sustainability objectives, but produce the same while using fewer resources.
With advanced self-service analytics, engineers and their companies can hit upcoming emissions target goals and keep their production schedules on track. HP
NOTES
a TrendMiner NextGen
WOUTER DANIELS has worked for TrendMiner as a data analytics engineer for more than 3 yr. Daniels helps train and advise companies to get the most out of their stored time-series process data and uses customer feedback to help improve the advanced self-service data analytics software. He provides custom solutions for more complex cases. Daniels earned his PhD in chemical engineering at the University of Leuven, where he studied data analysis and modeling of biochemical systems to increase bioproduct yields.