Intelligent alarms create actions from noise
Inside process: Alarm management software can decipher raw data, identify critical alarms, locate alarm sources, and provide role-based intelligence for faster resolution.
Industrial facilities and manufacturing plants are busy places with numerous pieces of equipment sending raw data, information and alerts from many systems and machines. According to a report from human-machine interface (HMI)/supervisory control and data acquisition (SCADA) experts at GE Digital, about 75% of all alarms are noise. Operators face the difficult task of deciphering the raw data, identifying which alerts are critical, and locating where the alerts are coming from. This time-consuming, costly process is complicated by a shifting workforce of new and temporary staff who may not have the experience or familiarity with plant systems to differentiate noise from important data.
Critical assets, smarter alarms
Intelligent alarming has been used over the past decade to consolidate data from SCADA systems, creating clear direction and driving corrective actions. With the increasing push for the adoption of the Industrial Internet of Things (IIoT) and Industry 4.0, intelligent alarming is giving plant operators an opportunity to gain efficiency and effectiveness with SCADA monitoring as well as enhance critical asset reliability for boilers, chillers, compressed air and other systems.
Intelligent alarming supports the operability of various systems including HVAC, which allows plant operators to maintain environmental parameters such as temperature and humidity, ensure staff safety and comfort and increase process reliability.
Advances in IIoT, real-time monitoring, and predictive analytics have helped intelligent alarming evolve into a multivariable alert system that looks at customized conditions and compares related data points. It provides a mobile safeguard for plant operations that can be kept in an operator’s pocket.
Four benefits for operators
By establishing a platform that transforms equipment noise from all systems across a plant into immediate corrective action, operators can:
- Preserve the longevity and operability of assets
- Reduce allocated costs for electrical and maintenance efforts
- Maintain regulatory compliance
- Take a more proactive approach to plant management.
Bringing intelligence to real-time monitoring
Every plant will have different requirements for operations and compliance based on the space, size and the industry they’re working in. The needs of biopharmaceutical manufacturing plants, for example, may vary from those of a manufacturer of home goods, requiring more air turns, tighter temperature ranges and reduced humidity. Pre-determined parameters create a series of alerts when operations fall outside of those parameters for temperature, energy use or other measurements.
A large quantity of alerts can overwhelm operators, creating multiple notifications for the same problem without identifying the issue or its source. For example, a valve leak may result in a spike in temperature, sending dozens of alerts in minutes. Combined with other monitoring efforts, operators may face hundreds of alerts per minute. It’s impossible for staff to keep up with all the alarms, which can result in alerts being ignored, which increases overall risk. Incorporating intelligence into alarming provides the ability to determine what the issue is, where it’s coming from, and create a faster alert to resolution process.
Creating condition-based alerts
The reason industrial facilities like manufacturing plants have so many alerts is because many alerts are “triggered-based.” This design sends an alert when a data point hits a pre-set threshold that acts as a trigger. Plant operators can set these triggers to send alerts at different levels as the data point hits different thresholds, which creates multiple alerts for one issue. Triggered-based alerts also are not subject to time span, data quality or frequency (sometimes referred to chattering). These alarms tend to be noisy and distracting to operators.
In the IIoT-driven world, new technologies like fault detection are using processed data and advanced analytics to consolidate alerts and determine what is critical and a priority before sending actionable alerts and information about a situation to the operator. This new technology has built-in intelligent alarming capabilities to eliminate false positives, minimize the chattering effect, and the reduce the overall effort to get to the cause and resolution. Time to resolution saves money and increases reliability in the process and product.
How it works: 3 steps
An intelligent alarm is a “condition-based” alert that accesses multiple data points through a layered analysis approach instead of creating an alert once the data point hits a certain trigger.
- In Layer 1, a data point coming in to the system is cleaned for quality, weeding out bad quality data from detection including gaps, controller sampling issues, and sensor health. If data quality is bad, the operator is alerted to the already pre-assessed situation, which further adds to the chatter.
- Next, the data point or multiple points are used for fault detection, forming a current condition or prediction of a condition. In a typical fault detection system, the data is assessed for cause prior to sending the alert, giving the operator actionable insight and knowledge of what could cause the situation.
- Lastly, the “condition-based” alert is tracked for chattering, trending, and timespan analysis so it doesn’t become noise. Conditions of similar assessments are not alerted multiple times, but instead triggered on the rules set by the operator, who also can assess the condition of the equipment after reviewing these aggregate results. For example, picture a motor sending an alert that it failed to start several times in a matter of minutes. Now, think about only seeing one intelligent alert with good data quality assessed, the condition analyzed down to the cause, and the chatter of continued attempts to start minimized.
By making these alert systems more intelligent with real-time monitoring and customizable analytics for multivariable analysis of condition-based alerts from multiple systems, operators can identify the most urgent conditions and provide suggested preventative or corrective action. This approach also streamlines all data into a single cloud-based platform, which consolidates critical alerts and suggested actions. These alerts can also be sent to an operator’s mobile device, driving higher response times when staff are on- or off-site.
Caters to varied staff experience
Another key concern for operators is ensuring newer staff members have the tools and direction to be as effective at identifying and addressing critical issues as existing staff members. Operators who have been at a facility for a while often have a better understanding of the equipment’s voice, knowing what data to monitor, what’s regular noise and which alerts are the most urgent. New and temporary staff will eventually accumulate this understanding, but the ramp-up period can be minimized with the use of intelligent alarming systems.
Because intelligent alarming and analytics provide such detailed alert context and visualization at the machine and process level, faster response time for new and existing staff is possible. Time to resolution can be reduced by up to 30% as management and engineering teams don’t have to sort through raw data and guess where an issue is coming from, which can save hours or days depending on the plant’s size. A prioritized list of alerts enables staff to save time identifying and assigning tasks for corrective actions and helps them address the most critical issues first to reduce allocated costs.
Issues and corrective actions can also be assigned through intelligent alarming, ensuring different team members are alerted only to the problems they need to address. This customization helps staff focus only on the alerts relevant to them and makes them seem less overwhelming by reducing their number. For example, an operator may get an alert about piece of equipment requiring adjustment while a plant engineer may get alerts about potential product quality issue. Individuals can receive customized, role-based view.
Shift to proactive maintenance
Using intelligent alarming and analytics provides an opportunity for a plant to shift from a reactive approach to maintenance to a more proactive approach. Teams with intelligence information can drive faster resolutions. Predictive analytics helps teams identify potential problems that could occur and address them proactively to prevent potential equipment failures.
Continuous fleet analysis of equipment and systems can reveal critical issues and key performance and reliability gaps. By understanding how the equipment is operating and what maintenance may need to be performed, staff can enhance operational efficiency and preserve the reliability of equipment — saving on allocated costs for maintenance and replacements.
Analytics also enables energy savings and reduces allocated costs for electricity as operators can identify target areas to maximize facility resource utilization and reduce energy waste.
Intelligent alerts also are critical to maintaining compliance. In many plants, compliance with certain standards from various regulatory bodies is absolutely necessary. If an issue jeopardizes compliance, operators will need to show they’ve done due diligence, which the collection of data on corrective actions can reveal. Alerting staff to conditions that could lead to a problem enables them to address the issue before it becomes critical.
Analytics and intelligent alarming are necessary for those who want to get the most out of their plants. It provides a tool that can make the lives of plant staff easier while also being more efficient.
Intelligent alarms can pull the corrective actions from equipment noise, allowing operators to reduce allocated costs maintain compliance and ultimately become a more reliable, efficient and proactive operation.
KEYWORDS: Intelligent alarms, alarm management
Alarm floods make operators less effective and increase risk.
Smart software assesses alarms in layers, providing intelligent information.
Lower risk, increased proactive maintenance can result.
Could intelligent alarms lower risk, save time, and improve your maintenance?