Equipment Factors

The nature and state of plant equipment are major sources of abnormal situations. A broad classification of different equipment items in a plant is: process equipment and control equipment. By process equipment we refer to such equipment as pumps, compressors, tanks, reactors, columns, etc. while control equipment includes sensors, valves and controllers. A variety of equipment-related factors such as age (old vs. new), loading (equipment operating at or beyond original design limits), maintenance history (poor or excellent history of preventive maintenance), etc. can impact the nature of abnormal situations.

Process Equipment

  • Equipment failures—These refer to sudden breakdowns of equipment such as a pump or compressor shutdown. Similarly, if a tube inside a furnace ruptures, then the feed through the furnace may have to be stopped. Such failures can cause significant plant upsets and in an extreme case can lead to plant shutdown. Typically an area of a unit gets affected and some standby equipment may be available to be used as a backup, e.g. spare pumps or spare heat exchangers.

  • Degradation of equipment—This is the more common mode of equipment-related abnormal situations. Heat exchanger fouling is an example of this kind. These gradual failures result in loss of production, or product quality or loss of control and are often difficult to detect.

Control Equipment

Errors in sensor reading and valve positions can cause a significant burden on the operations team in dealing with an abnormal situation because a decision has to be made on whether to trust a sensor/valve based on other corroborative evidence available. Often the correlations between one process value and other variables are significantly complex, resulting in erroneous judgments with potentially catastrophic consequences.

  • Sensor failures—There are four modes of failures of sensors: (a) Sensor reading is outside the sensor and/or process limits, (b) Sensor reading is changing at a rate that is outside the physical limit of the process or is inconsistent with the sensor characteristic, (c) Sensor is stuck giving out a constant reading, (d) Sensor bias, that is, the sensor is providing an output within physical limits but its reading has drifted away from the actual value. The output of sensors is often used to implement automatic feedback control. A faulty sensor reading can result in faulty control actions which can then lead to process malfunctions of varying consequences. For example, measurements of air flow into a reactor were inaccurate because of drift in the calibration which produced false air to gas ratio calculations.

  • Valve failures—Similarly, valves that are used to implement control actions may fail. While the control algorithm commands the valve to open (or close) by a certain percentage, the actual valve position may be different. This can again cause significant process upsets, if the operators do not identify the problem in time.

  • Controller failures—The advent of automated distributed control systems has made it very difficult for the operations team to switch back to manual control if such a need arises. In the event of a power outage that results in a DCS failure or if the CRTs blank out, the operations team is typically at a loss to figure out how to keep the plant running smoothly. Not only does the operations team lack the experience of manually controlling the plant, but also the complexity of the process presents tremendous challenges to the information processing capacity of human operators.