Long-established companies are transforming from their historically industrial, machine-focused roots into “digital industrial” companies, connecting people, machines and data via tools like networked sensors and software. By merging big data and cloud-based analytics with industrial machinery, this digital industrial approach is expected to push business operations in new directions. The ultimate is optimization: operating at the highest possible level through greater efficiency and increased operating performance.
As the adoption of digital industrial strategies and tools accelerates, companies are using analytics — targeted metrics based on specific key performance indicators (KPIs) — to uncover meaningful patterns in the data collected. But, while data provides the foundation, the analytical review tells the story of an operation. For an industrial company, using analytics can:
Using Analytics in Reverse Osmosis (RO) Applications
Analytics can help balance operational and financial business goals, thereby maximizing returns. Early adopters of the technology are already seeing real-time benefits of analytics in their water and process applications. For example, when analytics are applied to a RO (reverse osmosis) plant, they provide a tool that can be used to maximize the useful life of the RO membranes. As deposits form on membranes, their efficiency is reduced, which significantly raises the energy costs to process the same amount of water. Some deposits cause irreversible fouling and can compromise membrane integrity, thereby shortening membrane life, causing earlier than expected replacement.
Leveraging predictive analytics — centered on normalization of flow and salt passage, time-series statistical clustering and forecasting techniques — can build analytics models to predict the remaining time to clean and remaining useful life. By implementing analytics for RO, plant operators can better manage membrane cleaning schedules to obtain the most value from assets by optimizing energy use and achieving maximum membrane life.
The model-based predictive analytics is based on membrane anomaly detection through the lens of historical and current signatures of normalized permeate flow (NPF) and normalized salt passage (NSP) curves. The trends of NPF and NSP are analyzed together and projections about membrane health are made based upon statistically comparing these trends with signature patterns in the database.
- Take data from multiple sources and convert it into new and meaningful information.
- Allow current conditions and their trajectory to be shown visually.
- Help identify and diagnose problems and discover opportunities for improvement.
- Draw attention to events or trends before they threaten asset production or integrity.
- Facilitate reporting on key performance indicators and their impact on business objectives.
While these outcomes are undeniably advantageous, achieving them is not as simple as just installing a sensor or launching a software program. In practice, leveraging analytics in water and process operations is a three-step process to get connected, get insight and get optimized. Using this process in tandem with an agile development program and a strong digital vision, businesses can develop predictive maintenance plans, optimize cost, increase employee value and enable growth.
Start with a Connected Plant
Getting connected is the prerequisite system architecture that allows for data to be acquired from a range of sources and modes of capture. This includes:
- Manual direct entry through computers and mobile devices.
- Wireless connectivity with controllers, human-machine interfaces (HMIs), programmable logic controllers (PLCs) and supervisory control and data acquisition (SCADA).
- Wireless connectivity with field sensors, including sensors for tank-level monitoring.
- Plant distributed control systems (DCS) and data repositories.
- Manufacturing execution systems (MES), enterprise resource planning (ERP) and customer relationship management (CRM) data.
Getting connected also requires a close look at data-ingestion architecture, including data hygiene and storage. Integration of the disparate data into a logical software system is important to ensure success. For example, time-series data may require looking beyond traditional databases for big-data enabled frameworks like Hadoop+, an open-source software for storing and processing large data sets. To determine overall storage and processing needs, start by analyzing critical first-tier data and then prioritize the balance with the relative value of second-tier and third-tier data sources.
Being connected also means that necessary information is being communicated to the right people anywhere, on any device.
Get Insight into Your Industrial Plant
Getting insight into your process and plant is about harnessing the knowledge that companies can gain by bringing together data from intelligent machines with people and technology. With insights into asset performance, one is able to take actions that will resolve potential issues before they lead to impairments. Being connected allows companies to leverage analytics to find meaningful patterns in data and apply them to inform decision-making that drives productivity and improves operational efficiencies.
Figure 1 illustrates how analytics can be used to gain intuitive information about an asset. Using a series of data transformations, from noise cancellation to physics-based reconciliation, one can see developing fault trends in data that are not otherwise visible through “eyeballing” of unprocessed data trends. This underscores the importance of using analytics to gain more insight about operating processes.
Optimize Your Process Based on Analytics
Finally, getting optimized takes analytics to the next level by creating continuous-improvement processes that generate “best-in-class” operations. Optimization is enabled by applying insight, analytics and models that sustain the most predictive and prescriptive analyses of a plant.
Aggregating and analyzing data across assets “within the fence” can establish operating levels for individual assets that optimize the entire plant system, especially when inter-relationships between individual assets are taken into consideration. In turn, as the plant operates more efficiently, its overall water and energy usage can be reduced, improving the plant’s environmental footprint and operating margins. Without these analytics, each asset would be individually configured, which might result in a suboptimal performance for the total plant.
Analytics can identify the sweet spot by optimally balancing operational and financial business goals, thereby maximizing returns. Early adopters already are seeing real-time benefits of analytics in their water and process applications.
Condenser Efficiency. One such example is that of a condenser in a power plant. Condenser performance is a critical factor in determining heat rate. An increase in condenser backpressure means the turbine has to do more work, thereby directly affecting its heat-rate efficiency and overall power generation efficiency.
The loss in revenue can be substantial — in the hundreds of thousands of dollars — from additional fuel spent to make up for loss in heat transfer. Figure 2 depicts the sensitivity curve of a 40 MW steam-based power plant using a steam surface condenser.
Though the backpressure is directly measureable, the underlying cause could be many factors (for instance, condenser tube scaling, leakages or the recirculating water). Regardless of the cause, backpressure often can increase beyond an acceptable threshold and may only be observed when it is already too late.
Analytics can help detect backpressure in advance. Techniques involve physics-based modeling and time-series data modeling to find troublesome signatures in advance and provide early warning to take evasive actions. One of the most basic analytics for condensers is the cleanliness index (CI). The CI is normalized to take out variation due to operating conditions and shows the true degradation due to scaling. The trending visualization of this is a simple but powerful way to understand developing fault conditions in the condenser (figure 3).
Often, a rise in backpressure creates pressure on the chemical treatment vendor. Unless the correct causes are identified, corrective action will not provide the intended benefit.
With an analytics-based approach, incipient fault conditions are detected well in advance, and corrective actions can be applied proactively. The actions can include:
- Adjusting chemical treatment.
- Rebalancing recirculating water flows.
- Investigating a possible leak.
- Determining when the next cleaning is needed.
A digital approach can provide real-time data normalization, providing valid efficiency data during peak production times. Using this data, plant operators can visualize current conditions, identify problems within the system and adjust as needed to achieve optimal performance.
This is just one of the many analytics that water and process operators can use to help achieve greater system efficiency, reduce operating costs and maximize production in industrial applications. (For another example, see sidebar.)
To survive in the new digital industrial world, it is no longer enough to keep the status quo when it comes to technology. The most successful industrial companies will be equipped with fully integrated, predictive systems that help them get the most out of their investments and remain competitive in an ever-changing global marketplace.