Every time you turn on a faucet, flip a light switch, or fill your car with gas, there’s a good chance a SCADA system is working behind the scenes. SCADA (Supervisory Control and Data Acquisition) systems are the backbone of modern industrial operations, but what makes them truly powerful is how they process vast amounts of data from field devices in real-time. This article breaks down the data processing mechanisms inside SCADA systems-from database structures to quality flags and supervisory control.
Table of Contents
- The SCADA database: how industrial data is structured
- Analog data processing
- Status data processing
- Pulse accumulator processing
- Data quality: knowing what you can trust
- Data source flags
- Detailed quality flags
- Composite quality flags
- Calculated values and quality inheritance
- Topology processing: understanding system connectivity
- Supervisory control: commanding field equipment
- Control command types
- Command verification and safety
- From raw signals to operational intelligence
The SCADA database: how industrial data is structured
At its core, a SCADA system uses a tag database containing data elements called tags or points. Each tag corresponds to specific instrumentation or actuators within the monitored process. This structured approach allows operators to track thousands of individual data points-from temperature sensors to valve positions-across an entire industrial facility.
The SCADA database, often called the real-time database, stores three primary types of data: analog data, status data, and pulse accumulator data (also called count data). Each data type has its own processing rules, conversion methods, and storage requirements.
Analog data processing
Analog values represent the state of variable-state devices such as power lines, transformers, and pumps. When a physical variable is measured by a transducer in the field, the output passes through an analog-to-digital (A/D) converter in the RTU (Remote Terminal Unit). The SCADA Front End then converts all analog values to floating-point numbers and adjusts them to represent physical measurements in engineering units like megawatts or pounds per square inch.
Once received, the system performs several critical checks on analog values:
Limit checking: The system compares values against predefined thresholds. Normal limits define acceptable operational ranges. Reasonability limits filter out impossible values that indicate sensor faults or transmission errors. Rate-of-change limits detect abnormal fluctuations between consecutive scans. If any limit is violated, the system can trigger an alarm.
IEEE validity check: Before storing any value, SCADA validates it against IEEE standards to ensure the data can be safely processed and displayed.
Status data processing
Status values represent discrete-state devices such as circuit breakers, valves, and switches. Unlike analog values that exist on a continuous scale, status points indicate specific states-open or closed, on or off, normal or abnormal. All status point values are stored in POINT records within the database.
When an indicator changes its status, the system compares the new state against the previously stored value in the real-time database. If the new state differs from the expected normal value, the system generates an alarm event. Modern RTUs can also provide sequence-of-events (SOE) data with precise timestamps, allowing operators to reconstruct exactly what happened during complex system disturbances.
Pulse accumulator processing
Pulse accumulators measure the total amount of energy, liquid, or gas that passes a specific location. A detection device opens and closes a contact each time a unit of the measured substance passes by, and the accumulator counts these contact changes.
The processing involves differencing consecutive readings and scaling the result to engineering units. For example, a water meter might count pulses that represent gallons, and the system calculates flow rates by comparing accumulator values over time. Some RTUs support “freeze” commands that simultaneously capture accumulator values across multiple points, ensuring synchronized measurements for billing or reconciliation purposes.
Data quality: knowing what you can trust
Raw data means nothing if operators cannot trust it. SCADA systems employ sophisticated quality flagging mechanisms to indicate the reliability and source of every data point. This transparency allows operators to make informed decisions even when dealing with imperfect information.
Data source flags
Every measurement carries a source indicator showing where it originated. Values might be telemetered from RTUs (NREMOTE), calculated by the system (NCALC), manually entered by operators (NMANUAL), supplied by state estimation software (NESTIM), or provided by external applications (NEXTERN). When a value comes from a secondary source rather than its normal source, a replacement flag indicates this deviation.
Detailed quality flags
The system assigns quality indicators based on various conditions:
Un-initialized (U): No data has been retrieved for this point yet. These values should be considered meaningless until scanning brings in actual measurements.
Old (O): The displayed value could not be retrieved in the last scan cycle. This might indicate communication errors, reasonability check failures, or points placed out of service.
Bad (B): The data is unreliable due to calculation errors or hardware problems in field devices.
Not In Service (NIS): The point has been deliberately removed from service by an operator, preventing new values from being processed.
Composite quality flags
For rapid assessment, SCADA provides composite flags that summarize reliability: GARBAGE (un-initialized data), SUSPECT (old, bad, or questionable data), REPLACED (values overridden by operators or estimators), and GOOD (reliable current data). These composite indicators appear on operator displays, enabling quick visual assessment of system-wide data quality.
Calculated values and quality inheritance
User-defined calculations enable operators to derive new values from existing data points through arithmetic, logical, and comparative functions. A calculated point might sum power flows across multiple feeders or compute efficiency ratios from temperature and pressure readings.
Critically, calculated values inherit the quality of their input data. If any input carries a suspect or bad quality flag, the calculated result will also be flagged accordingly. This quality propagation ensures that derived information accurately reflects the reliability of its underlying sources. When input points are un-initialized, calculations are typically suspended until valid data becomes available.
Topology processing: understanding system connectivity
In electrical power systems, knowing which devices are connected is just as important as knowing their individual states. Topology processing determines the electrical connectivity of the network in real-time by analyzing the status of switching devices-circuit breakers, disconnectors, and switches.
SCADA systems collect data from field devices and process it to present operators with an accurate picture of system configuration. When a breaker opens, the topology processor recalculates which buses and lines remain energized and which are isolated. This information drives coloring on one-line diagrams, helps state estimators model power flows, and supports contingency analysis.
The topology model updates continuously as status changes arrive from the field. Operators see visual confirmation of switching operations, and the system can automatically detect inconsistent states-such as a breaker showing both open and closed-that might indicate measurement errors or equipment problems.
Supervisory control: commanding field equipment
SCADA’s supervisory control function allows operators and applications to control devices in the monitored system. From a central control room, operators can issue commands to change the state of equipment like pumps, valves, and circuit breakers located miles away.
Control command types
Control operations fall into several categories. Single-step commands perform immediate actions-opening a valve or starting a pump. Multi-step commands execute sequences of operations, such as energizing a transformer by closing a series of switches in the correct order.
Modern SCADA systems expose control functionality through Application Programming Interfaces (APIs), enabling automated applications to issue commands based on process conditions. For instance, an energy management system might automatically adjust generator outputs in response to changing load conditions.
Command verification and safety
When a command is issued, a timer is initiated to supervise the operation. If no feedback confirming the action arrives before the timer expires, an alarm alerts operators to the potential problem. This command supervision ensures that control actions actually reach their targets and produce the expected results.
Many systems implement interlocking logic that prevents dangerous operations-for example, blocking attempts to close a breaker onto a faulted line or refusing to start a pump when inlet valves are closed. These safety checks happen at multiple levels: in the SCADA host, in RTUs, and often in local PLCs at the equipment itself.
From raw signals to operational intelligence
The journey from field sensor to operator screen involves numerous transformations, validations, and calculations. Raw voltage signals become engineering values. Individual measurements combine into calculated quantities. Quality flags provide transparency about data reliability. Topology models reveal system connectivity. And control commands flow back to field equipment, closing the loop.
SCADA systems have evolved from simple monitoring tools into sophisticated platforms that handle thousands of data points with sub-second response times. Modern systems integrate with databases, analytics platforms, and enterprise applications, turning operational data into strategic intelligence.
Understanding how SCADA processes data helps engineers design better monitoring systems, operators interpret displays more effectively, and managers appreciate the infrastructure supporting critical operations.
What do you think? How might advances in edge computing and artificial intelligence change the way SCADA systems process and validate industrial data? What new quality indicators might become important as industrial systems become more autonomous?
References
- https://en.wikipedia.org/wiki/SCADA
- https://nathaniellindsey.gitbooks.io/scada-user-s-guide/content/overview_of_scada/how_scada_acquires_and_processes_data.html
- https://control.com/technical-articles/event-and-alarm-handling-and-processing-in-process-control-system/
- https://nathaniellindsey.gitbooks.io/scada-user-s-guide/content/
- https://www.dpstele.com/scada/how-systems-work.php
- https://inductiveautomation.com/resources/article/what-is-scada
- https://cedalo.com/blog/scada-system-guide/
Leave a Reply