Reliability HotWire: eMagazine for the Reliability Professional
Reliability HotWire

Issue 62, April 2006

Hot Topics

Modeling Periodical Inspection and Early Failure Detection in BlockSim

 

To improve availability, integrity, productivity and safety, many organizations implement regular monitoring of critical parts of the system, enabling regular feedback about the status of the equipment and the degradation over time before reaching failure. This allows quick response to disturbances and helps avoid unplanned shutdowns or safety problems.

 

There is an increased trend toward introducing modern censoring and early failure detection techniques into the design of advanced control centers in various applications such as manufacturing, navy ships, process industry and nuclear plants. The role of these computerized systems is to detect faults and degraded performances and alert the operators to anomalies in the system to help them manage the problems as quickly as possible with minimal adverse consequences. This article presents an approach to modeling periodical monitoring of a system and early detection of failures using BlockSim.

 

Let us consider the example of a Shutdown System in a nuclear power plants control system. The simplified Reliability Block Diagram (RBD) of the system  is shown next.

 

 

 

The channels are represented by the following RBD.

 

 

The Reactor component of the system is considered to be a critical component that needs regular monitoring and improved visibility to allow early detection of degradation that indicates an immanent failure. The following table summarizes the failure distributions and the repair distributions and cost of repair for the blocks in the above two RBDs.

 

 

Sensors are used in the Reactors to detect degradation in their parameters. The sensors monitor the Reactor daily. In BlockSim, this can be modeled using the Inspection Policy by clicking the Create New Inspection Policy button, , in the Inspection tab of the Reactor's Block Properties, selecting Upon fixed time interval based on: and System Age and entering the value that represents the frequency (every 24h) of the sensors.

 

 

The implemented sensor system can detect degradation in the Reactor and predict when a failure is about to happen. The sensitivity of the sensor is estimated at 90% from failure time, i.e. at about 90% of the failure time, the impending failure is detected. This is quantified in BlockSim using the the Failure Detection Threshold (FDT), which is a number from 0 to 1 that indicates the percentage of an items life that must elapse before an approaching failure can be detected. For example, if the FDT is 0.9 and the item will fail at 1000 days, the approaching failure can begin to be detected at 900 days. Note that even though the sensor will be measuring degradation, the threshold to trigger a preventive maintenance action in BlockSim is based on simulated failure times. From the relationship between time and degradation, the degradation threshold can be transferred to a triggering time threshold. The Failure Detection Threshold (FDT) field is found in the Inspection tab of the Reactor's Block Properties, as shown in the next figure.

 

 

When a failure is anticipated, the controller issues an order to the on-site operators to perform a preventive action. The cost and preventive repaire time are described in the following table.

 

 

This preventive inspection plan is defined in BlockSim by clicking the Create New Preventive Action Policy button, , and setting the appropriate Preventive Action Policy. Make the selections shown in the next two figures.

 

 

 

Usually, preventive maintenances are performed on a regular basis (either based on item age or system age) or when the system is down. In this example, neither of these preventive maintenance types is performed. In the above figure, the Upon Maintenance of another Group Item was selected; however, since no groups of blocks were created in this RBD, this selection will not have an influence on the analysis. The goal of this preventive maintenance policy is to associate a preventive repair action to the time when the Failure Detection Threshold is reached.

 

If the inspection plan is not implemented (i.e. no settings are made under the Preventive tab or the Inspection tab in the Reactor's Block Properties), then for 5 years (43,800h) of operation, the average availability is estimated to be 0.9951 and the cost of maintaining the Shutdown System is $2,581,671 as shown next in the simulation's results summary.

 

 

Before simulating the entire Shutdown System with the Reactor's preventive and inspection settings, the Reactor (one Reactor) is simulated by itself for 5 years of operation. The following is an overview summary of the Reactor's simulation results if the regular inspection plan is implemented.

 

 

The above table indicates that about 5 preventive maintenances are expected to be initiated (based on the early detection of failures).

The system is simulated for 5 years of operation.

 

The following is a summarized overview of the system's performance with the Reactor's preventive and inspection settings.

 

 

 

The above table shows that for 5 years of operation, the average availability is estimated to be 0.9997 and the cost of maintaining the Shutdown System is $183,153 + $500,000 = $683,153 (after adding the sensor equipments and the computerized system costs) which is a major improvement compared to the results obtained if no inspection and preventive maintenance is implemented at each Reactor. The next table summarizes the comparison analysis between implementing the inspection plan and not implementing it.

 

 

Shown next is a plot of the components' FCI (Failure Criticality Index) values, which is obtained from the number of system downing failures divided by the number of system failures.

 

 

The next plot is a plot of the components' DECI (Downing Event Criticality Index) values, which is obtained from the number of system downing events divided by the number of system failures.

 

 

These two plots show that even though the Reactor component is less reliable in comparison with the other components in the system, its contribution to the overall system unreliability is not significant, which is due mainly to the preventive maintenances that are implemented once a Reactor failure is anticipated. This analysis illustrates the benefit of implementing regular inspection policies and the great advantages gained when failures can be anticipated and dealt with in a timely fashion.

Copyright 2006 ReliaSoft Corporation, ALL RIGHTS RESERVED