Handbook of Reliability, Availability, Maintainability and Safety in Engineering Design - Part 53 ppt

4.4 Application Modelling of Availability and Maintainability in Engineering Design 503 Table 4.14 Preliminary design data for simulation model sector 1 Assembly Code Flow vol. Mass flow Liq. Solids Transfer conv eyor 1 C015141 575 1,508 106 1,403 Transfer conv eyor 2 C015241 575 1,508 106 1,403 Rev.shuttle conveyor 1 C024111M 575 1,508 106 1,403 Rev.shuttle conveyor 2 C024211M 575 1,508 106 1,403 Storage bin 1 U024111 192 503 35 468 Storage bin 2 U024211 192 503 35 468 Storage bin 3 U024311 192 503 35 468 Belt feeder 1 Y024121 192 503 35 468 Belt feeder 2 Y024221 192 503 35 468 Belt feeder 3 Y024321 192 503 35 468 Mill feed conveyor 1 C024121 192 503 35 468 Mill feed conveyor 2 C024221 192 503 35 468 Mill feed conveyor 3 C024321 192 503 35 468 a) Evaluation of Simulation Model Sector 1 A major characteristic oftheprocess flowdiagram(PFD) ofsector 1 is that it depicts material flow and indicates how inputs are generated and then transformed by each system (or assembly) into outputs that, in turn, become the inputs to the next system (or assembly), as depicted in the preliminary design data given in Table 4.14. These are specifically mass-flow volumes of solids. The PFD is systematically examined to analyse deviations in process flow and system performance and, in this case, to determine mass-flow balance through the integrated assemblies. Each assembly is graphically represented in the simulation model by a virtual prototype process equipment model (PEM). Each of the assemblies of the PFD depicted in Fig. 4.51 (i.e. the feeder, three storage bins, three chute conveyors,and three transfer conveyors)is a process equip- ment model. Each PEM contains selected model components that are linked together with logical flows. A model component is a modular design entity with a complete spec- ification describing how it is connected to other model components in a model con- figuration. Model configurations are created when two or more model components are connected to each other via their interfaces. Each model component has connec- tors that are the interface points of the component (or block). Connections are lines used to specify the logical flow between the connectors from one block’s output to another’s input. Thus, a process system or assembly can be represented either as a single model component or as a configuration of several components. Logical flow initiation—the random number generator The model components are selected, configured and assembled in such a way that the design specifications of each system are met through the component’s attributes, and the linked logical flows. Thus, the feeder assembly PEM, for example, has its own specific model con- figuration, in contrast to that of the storage bins, as depicted in the design details of 504 4 Availability and Maintainability in Engineering Design Fig. 4.51 Process flow diagram for simulation model sector 1 Figs. 4 .52 and 4.53. However, all simulation models, especially Monte Carlo simu- lation, have random number generators for ‘seed’ values initiation of the simulation model’s input flow variable(s) that constitute the initial flow of the linked logical flows thereafter. The model component’s attributes depicted in Fig. 4.52 generate random num- bers according to a statistical probability distribution, convert the outputs of a con- version function by modifying the component’s inputs through a selection of statis- tical functions, as well as calculating the mean, variance and standard deviation of the component’s inputs. Logical flow storage—the process equipment models (PEMs) Logical flow in the context of process systems simulation modelling represents upstream material feed that, in effect, cau ses the initiation of the process equipment model (PEM). Logical flow storage PEMs are process simulation models in which the model con- figuration incorporates a model component attribute of an output conversion func- tion that modifies the component’s inputs through a selection of statistical functions, and statistical pro bability distributions. As previously indicated, the PEMs incor- porate all the essential process analysis preliminaries for preliminary engineering designs of large integrated process systems. 4.4 Application Modelling of Availability and Maintainability in Engineering Design 505 Fig. 4.52 Design details for simulation model sector 1: logical flow initiation The application of dynamic systems simulation modelling incorporating the PEMs is primarily to determine the effect of logic flow in complex integrations of systems in large engineered installations. The model component’s attributes de- picted in Fig. 4.53 incorporate probability distribution modifiers of the logic flow within each PEM. Output performance results The Extend c  Performan ce Modellin g program pro- vides a powerfully flexible graphical output presentation through dynamic plotters. These plotters can be placed anywhere in the modelled system configuration, and connected between any of the PEM input/outputinterface connectors,or within each PEM b etween model component connectors. Figure 4.5 4 illustr ates a typical output document showing performance results of the storage bin assembly. These performance variables relate to system or assembly contents, input and output flow quantities, as well as flow surges. The flow surge gives an indication of m aterial flow balancing in the process, subject to upstream material feed. The storage bin PEM illustrated in Fig. 4.54 has a plotter connected to the output model components of the PEM. The plotted graph in the figure shows the trend of material flow through the storage bin from start-up to steady state. 506 4 Availability and Maintainability in Engineering Design Fig. 4.53 Design details for simulation model sector 1: logical flow storage PEMs b) Conclusion of Simulation Model Sector 1 Evaluation Table 4.15 below indicates the values o f a comparative analysis of preliminary de- sign data and simulation output data for simulation model sector 1. Column 2 of the table gives the specified preliminary design flow volumes, and column 3 gives the mean of the simulation model’s output data. On first scrutiny, these two values are identical with an expectation of a 100% correlation, resulting in the conclusion that the model’s output is a perfect match to the specified preliminary design flow volumes of the listed assemblies in simulation model sector 1. The evaluation of simulation model ou tput data is, however, not that simple, as other factorsmust be includedsuch as requirements for meeting the full design spec- ification inclusive of allowable tolerances, and determining whether the minimum and maximum values, i.e. the range of output variances for each simulation run of the model’s output data, fall within the expected confidence intervals of the design specification. The test of whether the simulation model’s output variances fall within the allowable design tolerances is set at a 99% level of confidence. The allowable design tolerance for throughput flow volumes is set at ±2.5% of the mean. Figure 4.55 indicates the simulation model’s output for simulation model sec- tor 1, including operational flow throughput (OPS), maximum and minimum flow (MAX) and (MIN), and mean flow output (MEAN). However, an acceptable lower 4.4 Application Modelling of Availability and Maintainability in Engineering Design 507 Fig. 4.54 Design details for simulation model sector 1: output performance resul ts Table 4.15 Comparative analysis of preliminary design data and simulation output data for simu- lation model sector 1 Assembly Design flow Model flow Model min. Model max. vol. vol. flow vol. flow vol. Transfer conv eyor 1 575 575 565 585 Transfer conv eyor 2 575 575 565 585 Rev.shuttle conveyor 1 575 575 565 585 Rev.shuttle conveyor 2 575 575 565 585 Storage bin 1 192 195 180 210 Storage bin 2 192 195 180 210 Storage bin 3 192 195 180 210 Belt feeder 1 192 195 180 210 Belt feeder 2 192 195 180 210 Belt feeder 3 192 195 180 210 Mill feed conveyor 1 192 195 180 210 Mill feed conveyor 2 192 195 180 210 Mill feed conveyor 3 192 195 180 210 tolerance limit (LL) and an upper tolerance limit (UL), against which the minimum and maximum values of the simulation model’s output data can be compared, need to be established to determine whether the range of variances of the model’s output data falls within these tolerance limits. 508 4 Availability and Maintainability in Engineering Design Fig. 4.55 Simulation output for simulation model sector 1 Table 4.16 Acceptance criteria o f simulation output data, w ith preliminary design data for simu- lation model sector 1 Assembly Design min. Design max. Model min. M odel max. Yes/no vol. 2.5% tol. vol. 2.5% tol. vol. vol. at 99% Transfer conve yor 1 565 585 565 585 Yes Transfer conve yor 2 565 585 565 585 Yes Rev.shuttle conveyor 1 565 585 565 585 Yes Rev.shuttle conveyor 2 565 585 565 585 Yes Storage bin 1 187 197 180 210 No Storage bin 2 187 197 180 210 No Storage bin 3 187 197 180 210 No Belt feeder 1 187 197 180 210 No Belt feeder 2 187 197 180 210 No Belt feeder 3 187 197 180 210 No Mill feed conv eyor 1 187 197 180 210 No Mill feed conv eyor 2 187 197 180 210 No Mill feed conv eyor 3 187 197 180 210 No Validation of the simulation model’s output data is thus not confined to a mere correlation of the mean values, whereby problems of autocorrelation can be sig- nificant, and the simulation model runs are not large enough to justify statistical spectral analysis of the output data (especially with very large, complex dynamic 4.4 Application Modelling of Availability and Maintainability in Engineering Design 509 simulation models), but the range or variance of the model’s output data is com- pared to acceptable lower and upper confidence limits within a specified exact prob- ability. The design specification is thus used as the mean, and the allowable design tolerance of ±2.5% of the mean is used as the square root of the variance, or stan- dard deviation in the statistical t-distribution, to determine a confidence range or interval with lower tolerance limit (LL) and an upper tolerance limit (UL) at a 99% level of confidence for ten simulation runs. The minimum and maximum values of the simulation model’s output data are then compared against this confidence range or interval. The last column of Table 4.16 indicates whether the model’s output is acceptable in meeting the design criteria within a 99% level of confi- dence. c) Evaluation of Simulation Model Sector 2 A major characteristic of the process flow diagram (PFD) of sector 2 is that it de- picts the conversion of solids to a solids and liquid slurry flow (through the action of the mills), and indicates how inputs are transformed into logical flow outputs that become modified inputs to the following assemblies (through the action of the Table 4.17 Preliminary design data for simulation model sector 2 Assembly Code Flow vol. Mass flow Liq. Solids Rod mill 1 X024131 307 654 186 468 Rod mill 2 X024231 307 654 186 468 Rod mill 3 X024331 307 654 186 468 Rod mill 4 X024431 307 654 186 468 Mill discharge tank 1 T024141 1,119 2,102 943 1,159 Mill discharge tank 2 T024241 1,119 2,102 943 1,159 Mill discharge tank 3 T024341 1,119 2,102 943 1,159 Mill discharge tank 4 T024441 Classifier feed pump 1/1 P024151 560 1,051 472 580 Classifier feed pump 1/2 P024152 560 1,051 472 580 Classifier feed pump 2/1 P024251 560 1,051 472 580 Classifier feed pump 2/2 P024252 560 1,051 472 580 Classifier feed pump 3/1 P024351 560 1,051 472 580 Classifier feed pump 3/2 P024352 560 1,051 472 580 Classifier feed pump (S) P024451 Classifier feed pump (S) P024452 Screen feed pot 1 V024161 1,152 2,142 982 1,159 Screen feed pot 2 V024261 1,152 2,142 982 1,159 Screen feed pot 3 V024361 1,152 2,142 982 1,159 Screen feed pot 4 V024461 Ball mill 1 X024141 515 1,056 365 690 Ball mill 2 X024241 515 1,056 365 690 Ball mill 3 X024341 515 1,056 365 690 Ball mill 4 X024441 515 1,056 365 690 510 4 Availability and Maintainability in Engineering Design Fig. 4.56 Process flow diagram for simulation model sector 2 pumps), as depicted in the preliminary design data given in Table 4.17. The PFD is systematically examined to analyse deviations in process flow and system perfor- mance and, in this case, to determine solids to fluids mass-flow balance through the integrated assemblies. Each assembly is graphically represented in the simulation model b y a virtual prototype process equipment model (PEM). Each of the assemblies o f the PFD de- picted in Fig. 4.56 (i. e. the four mill feeder chutes, eight mills, eight pumps, four mixer chutes, four multi-bin feeders) is a process equipment model. Process design specifications Each PEM contains model componentsthat are con- figured in such a way that the design specifications of each system or assembly are met through the component’s attributes. The model component’s attributes for the mill input feeder chute and the mill output mixer chute connect three input values to a single output, and two input values to a single output respectively. The attributes for the multi-bin feeder convert the output by modifying the component’s multi- ple inputs through a selection of statistical functions. The attributes for the mill pump also convert the pump’s output by modifying the component’s inputs through a selection of statistical functions representing typical pump delivery character- istics. 4.4 Application Modelling of Availability and Maintainability in Engineering Design 511 Fig. 4.57 Design details for simulation model sector 2: holding tank process design specifications Figure 4.57 illustrates the model component’s attributes of the rod mill, specifi- cally the holding tank process characteristics such as operating contents, maximum and minimum contents, initial flow, final flow, initial contents,final contents,as well as initial and final flow surge. Output performance results Performance variables relate to system or assembly contents, input and output flow quantities, as well as flow surges. The flow surge gives an indication of mass-flow balancing in the process. The output document is particular to each PEM and can be opened at any time, anywhere, in the dynamic systems simulation to determine the status of the process flow. The Extend c  Perfor- mance Modelling prog ram plotters can be placed anywhere in the modelled system configuration, and connected between any of the PEM input/output interface con- nectors, or within each PEM between model component connectors. The different process equipment models illustrated in Fig. 4.58 have plotters connected to the model components of each PEM’s model configuration. Figure 4.5 8 illustr ates a typical output document showing performance results of the second-stage mill assembly. The plotted graph shows the trend of flow through the mill from start-up to steady state. 512 4 Availability and Maintainability in Engineering Design Fig. 4.58 Design details for simulation model sector 2: output performance results d) Conclusion of Simulation Model Sector 2 Evaluation Table 4 .18 gives the values of a comparative analysis of prelimin arydesign data and simulation output data for simulation model sector 2. Figure 4.59 shows the simulation model’s output for simulation model sector 2. As with simulation model sector 1, the range or variance of the model’s output data is compared to acceptable lower and upper confidence limits within a specified exact probability. The design specification is ag ain used as the mean, and the allowable design tolerance of ±2.5% of the mean is used as the standard deviation in the t-distribution, to determine a confidence range or interval with lower tolerance limit (LL) and an upper tolerance limit (UL) at a 99% level of confidence for ten simula- tion runs. The minimum and maximum values of the simulation model’s output data are similarly compared against this confidence range or interval. The last column of Table 4.19 indicates whether the model’s output is acceptable in meeting the design criteria within a 99% level of confidence. As can be seen, the mills and classifier feed pumps have a flow volume variance that is not acceptable within the 99% con- fidence interval as set by the de sign criteria, whereas the ball mills partially comply with the design criteria in that the simulated m inimum flow is within the acceptable lower limit (LL). . preliminaries for preliminary engineering designs of large integrated process systems. 4.4 Application Modelling of Availability and Maintainability in Engineering Design 505 Fig. 4.52 Design. its own specific model con- figuration, in contrast to that of the storage bins, as depicted in the design details of 504 4 Availability and Maintainability in Engineering Design Fig. 4.51 Process. selection of statistical functions representing typical pump delivery character- istics. 4.4 Application Modelling of Availability and Maintainability in Engineering Design 511 Fig. 4.57 Design