Process-intensive Industrial Operators are increasingly wading into the Digital Transformation revolution to streamline production in the face of globalization, a shifting workforce, and heightened cyber security threats. Delivering safe, efficient, reliable, and environmentally friendly operations while maintaining a competitive advantage over peers is virtually unachievable without intelligent adoption of technology and cultural acceptance of changes to long standing processes.
Bently Nevada has been partnering with customers to help solve these challenges. Through user research in 20 countries with more than 400 end users, we have studied our customers’ team dynamics, site processes, and technology suites to determine how System 1 can best support Plant-Wide Machinery Management. The resulting platform meets eleven unique customer use cases, with each use case leveraging a subset of System 1’s Connectivity, Analytics, and Visualization capabilities. This Orbit Article explores Decision Support, a key System 1 offering that delivers advanced insights for all operational assets.
Expectations for business improvements continue to grow, often spurred by the promises of digital transformation and big data. However, many challenges limit the ability of industrial operators to achieve these productivity goals while minimizing downtime and protecting the integrity of operations. Bently Nevada has conducted discovery sessions with leaders in multiple industrial markets to identify the challenges that have become increasingly important over time and are now a top business priority. These challenges include:
Overwhelming quantity of data, making it impossible to continuously monitor and analyze assets for known failure modes, anomalies, and tracking KPIs
Lack of central knowledge base for tracking and retaining corporate knowledge related to machinery operations and process improvements
Difficult to acquire and deploy proven analytics to support risk mitigation, as outlined in plant reliability strategies and/or OEM documentation
Aging workforce means losing machine and plant knowledge
Complexity of available tools for creating, modifying, and deploying custom algorithms or KPIs to proactively detect evolving asset issues
Limited ability to correlate analytic results with machine/process conditions for root cause analysis, due to disparate historian monitoring platforms
System 1 Solution
To address these challenges, Bently Nevada has expanded the System 1 Condition Monitoring Platform to offer an analytic solution called Decision Support. While Decision Support was first released in the early 2000’s for System 1 Classic, the new version of Decision Support has been redesigned from the ground up with enhanced capabilities and is tightly integrated with System 1 (starting with System 1 version 20.1). Decision Support harnesses the rich data set acquired by System 1, processes this data through configured analytic rules, and populates the System 1 application with the resulting enhanced insights, enabling more proactive asset management.
Decision Support contains a collection of proven Bently Nevada analytics, which can be tuned to suit the unique operational application of a machine or asset. Decision Support can also be utilized to create and deploy custom rules that help to capture, disseminate, and leverage business knowledge of industrial equipment and processes.
The combination of Connectivity, Analytics, and Visualization capabilities that enable enhanced insights through the combination of System 1 + Decision Support are detailed in the following sections.
Pillar 1: Connectivity
The System 1 Platform provides comprehensive connectivity to machine data sources on the edge and collects high resolution vibration, process, and control system data. This data can be collected from Bently Nevada devices, such as 3500, Orbit 60, Ranger Pro Wireless Sensor, and SCOUT, with Trend, Waveform, and Device-generated alarms all imported. In turn, process tags can be collected from a Programmable Logic Controller (PLC) or historians. Up to 15,000 OPC and Modbus tags are supported per server, allowing customers to add process, first-out, and permissive data to enhance their machinery management programs.
The Decision Support engine has access to the high-resolution data within System 1; data sets include (but are not limited to) vibration, speed, pressure, flow, temperature, performance, and emissions. The data flowing into System 1 can be used as rule inputs, which then flow through logic and mathematical steps. The outputs of the rules (called results) are shared with the System 1 Platform for visualization and further analysis. Depending upon the configuration of the analytic, rules can be configured to execute as fast as once per second.
System 1 and Decision Support reside on the same server. To meet additional use cases and customer requirements, Bently Nevada is targeting Decision Support integration with Replicated (Rx) databases on Level 4 of the end users network in 2021.
Pillar 2: Analytics
Decision Support automates diagnostics and provides early detection of mechanical, operational, instrumentation, auxiliary, and process anomalies. The application acquires high resolution data from System 1, analyzes it with physics and model-based logic rules, and returns results to System 1 for notification, trending, visualization, and root-cause diagnostics.
Decision Support can be used by your staff to design and deploy custom rules that help you capture, disseminate, and leverage knowledge of your equipment, processes, and business solutions. Custom rules preserve operational knowledge in a usable format that can be broadly applied in an easily repeatable and manageable way. In addition, end users may elect to acquire proven Bently Nevada analytics, called Decision Support Analytics, which are engineered to detect a wide variety of failure modes.
In the Build environment, users can graphically construct rules by leveraging the included library of steps. Examples of steps include math, logic, and timers.
Custom rules can be created to detect a change in machine and/or process response, machine running time, count of startup/shutdown events, performance data, failure mode detection, and many other applications where automated analytics are required. The Build environment has been designed to allow end users, with a variety of backgrounds, to create rules and understand those rules created by others. The simplicity of creating, deploying, and modifying rules does not require a background in programming languages.
Key features of the Build process include:
Ability to copy and paste entire rules or sub sections of rules to expedite rule development
Ability to have multiple rules open at the same time for comparison or modification
Ability to define input step data types to ensure inputs are properly assigned.
Example: An input step may be defined as a temperature tag. During the deployment process, other tag types (e.g., flow) cannot be mapped to it as the engineering units are different.
Ability to define rule result parameters.
Measurement properties of the results are set in Build to enable integration into the System 1 database for trending and visualization without any additional configuration in System 1.
Annotations may be applied to the rule to convey information to other team members.
Example: A rule may calculate efficiency. The rule builder may apply an annotation to include details on the math expression or to highlight where coefficients maybe modified.
Ability to build rules online or offline. If Decision Support is connected to a System 1 database, an online connection is present. However, end users may install Decision Support and create rules when not connected to System 1. Rules created offline can later be imported to an online Decision Support connection.
For the first time, Decision Support includes a library of pre-configured rules called extractions. The extraction rules use the same Bently Nevada methodology taught to our customers and machinery diagnostic personnel, combining measurements, statuses, reference values, and configured properties, to produce a simple set of values that physically represent the behavior of the asset. These rules are available to the end user to gain new insights into the behavior of one or many machines, or to use as inputs into your own custom rule to save time in rule creation.
The Extraction Library includes the below mechanical health rotating asset extractions for use in custom rule creation and diagnostic analysis. Bently Nevada has additional extractions for API reciprocating compressors and dry gas seals, and will be adding to our offering for additional assets, processes, and auxiliary systems in subsequent releases.
Rule Deployment & Rule Maintenance
Once a rule is built, the rule may then be deployed to the System 1 Platform. The deployment process includes associating the rule to its target location(s) within System 1 (e.g.. Machine, Process System, Auxiliary System). Deployment also includes the mapping of measurement data to the input of the defined rule. Inputs for the rule may come from the targeted machine, or from another location within the System 1 hierarchy, or even from other Decision Support rules. In the Deploy mode, Decision Support connects to a System 1, exposes the configured hierarchy, and provides the data sets to feed the rule.
Once the rule is deployed, the analytic results will automatically populate the System 1 Database (no further System 1 configuration is needed). The rule results are logically grouped (associated to the machine or process area where deployed) for ease of navigation, visualization, and diagnostics.
Decision Support uses simple color coding for rule deployment status conditions:
Green indicates that the deployment configuration is complete and valid to execute
Red indicates that the deployment configuration is not complete (Example: a required measurement has not yet been mapped to an input)
Green & yellow indicates a change in the master rule logic has taken place, but that deployment has not yet been upgraded to use the latest revision
If a rule has been applied to many machines and/or process locations within System 1 and that master rule is changed, the end user (with proper security privileges) can quickly update all instances of the rule or actively select which individual deployments shall be updated.
Decision Support Analytics
Bently Nevada has 60+ years of experience solving customer problems with a focus on machinery and process conditions. This deep knowledge of rotor dynamics has been embedded into a packaged offering called Decision Support Analytics. Decision Support Analytics are designed to detect failure modes associated with various equipment types. End users may elect to deploy Decision Support Analytics on one or many machines, create their own custom rules to provide new insights, or combine both solutions to solve business problems.
Bently Nevada is initially offering Decision Support Analytics for rotating fluid film bearing assets, API reciprocating compressors, and auxiliaries like dry gas seals used for centrifugal type compressors.
The initial Decision Support Analytics for rotating assets include:
Speed Increasing/Decreasing Gearboxes
Integral Gear Compressors
Industrial Gas Turbines
Aeroderivative Gas Turbines
With the appropriate instrumentation, the diagnostics included are:
Fluid induced instability
High synchronous vibration
Sub and Super synchronous rubs
High Exhaust/High Differential exhaust temperatures
Electric motor non-uniform airgap
The Decision Support Analytics for reciprocating assets will initially cover API type compressors in services of hydrogen, flare, LDPE, or natural gas. With the appropriate instrumentation, the initial diagnostics included are:
Crosshead pin loading
Pressure packing leaks
Suction/Discharge valve leaks
Leak-Cylinder to low pressure
Leak-High pressure to cylinder
Initially the Decision Support Analytics for Auxiliary systems will include two types of dry gas seals used in centrifugal compressors. Both the Tandem and Tandem with Intermediate Labyrinth will be covered. The diagnostics for these two types include:
Bearing oil migration
Low seal gas DP
Low seal gas temperature
Low secondary seal gas temperature
Low separation seal gas DP
Primary seal failure
Seal gas flow problem
Secondary or Separation seal failure
Secondary seal failure
Secondary seal gas flow problem
Separation seal gas flow problem
Seal gas booster fouled filter
Seal gas fouled filter
Secondary seal gas fouled filter
Separation seal gas fouled filter
In the future, there will be additional diagnostics covering a growing range of assets, auxiliary, and process systems. Stay tuned for additional Extractions and Decision SupportAnalytics. Decision Support Analytics will be commercially available in Q4 2020.
Pillar 3: Visualization
The rule results from Decision Support are integrated and stored in the System 1 Historian, enabling users to leverage System 1’s rich feature set to extract value from these new insights. These visualization capabilities are highlighted through the example below of a rub condition on an LP1 Compressor. The rub condition was detected by consuming real-time data from System 1 and processing the data though a Bently Nevada Decision Support Analytic, the result of which was integrated back into System 1 for visualization and further diagnostics.
The HMI Builder workspace enables customers to visually represent machine health across an entire facility, within the context of its sister machines and supporting sub-systems. These screens can incorporate key rule results from Decision Support, such as process indicators, fault conditions, and operational KPIs. Navigational links can be configured to quickly jump to supporting HMI views or directly to the associated data within the Plots workspace.
System 1 Notification Plans may be applied to rule results based on defined conditions across multiple datasets. Upon annunciation of these conditions, the appropriate personnel can be notified with customized messaging to take appropriate action.
Within the System 1 Events list, the Decision Support result is documented as an Alarm event, allowing quick access to the supporting data within the Plots workspace.
Correlated Data Analysis
System 1 allows users to view vibration, process, and control data within a single plotting workspace, with a multitude of plotting formats available. As data is stored at up to 1-second rates, measurement values can be easily correlated. This enables users to mix and match data presentations for root cause diagnostics.
Connectivity + Analytics + Visualization = A Complete Solution!
By combining its Connectivity, Analytics, and Visualization capabilities, System 1 is positioned as the premier Edge Historian and Condition Monitoring platform of all Industrial Operators. To recap:
System 1 collects data from any asset within a facility. Collection rates of up to once per second are achievable during steady-state operations, while sub-second data can be stored during alarm and startup/shutdown events (when available from the device).
The stored data can be analyzed to derive insights using a rich set of tools. The core System 1 application can configure threshold-based alarms, while Decision Support and Bently Performance layer on additional analytic capabilities for earlier detection and greater understanding of machine health and operational conditions.
Replication frees data from the secure confines of the plant network, allowing more users easy access to data. Up to eight Transmitter databases can be replicated to one Receiver database on the Business Network.
Once on the Business Network, users can access S1’s extensive visualization capabilities, allowing for efficient investigation of abnormal machine conditions. Finally, standard interfaces (such as OPC UA) serve up the rich System 1 data to external systems, populating data lakes and feeding machine learning algorithms. All provided in a fully integrated solution to serve your current and future condition monitoring goals.
What’s Planned for v20.2
System 1 v20.2 will be released in November 2020. A complete list of targeted capabilities is captured below.
As a Software Product Manager, Jeff listens to customer needs and establishes the direction for System 1 Analytics. While working for Bently Nevada, Jeff has accumulated 30 years of experience in condition-based maintenance and asset management.