Mechanical Engineering News

VOLUME 34

JULY 2003

New Features in CodeCalc 6.5 and PVElite 5.0

(by: Mandeep Singh)

CodeCalc version 6.5 (also included in PVElite version 5.0) was released in January 2003 and has many new features. Some of the new capabilities are listed below: Enhancements in the TEMA Tubesheet module in CodeCalc 6.5 In this version, the tubesheet module and the thick (flanged and flued) expansion joint modules were integrated. The tubesheet module contains the input for the expansion joint. This makes analyzing a fixed tubesheet with an expansion joint, quicker and easier. As the manual data transfer is eliminated, the chances of error are also reduced. See the article Integration of Tubesheet and Expansion Joint also in this newsletter.
> continued on p.2

IN THIS ISSUE:

FOR THE POWER, PROCESS AND RELATED INDUSTRIES

CADWorx Equipment

Whats New at COADE

CodeCalc and PVElite New Features .......... 1 COADE Releases CADWorx EQUIPMENT .. 3 COADE Announcement Procedures .............. 4 Graphics Speed Issues .................................. 5 Exporting Output to MS Word ........................ 7 CAESAR II Version 4.50 ................................ 8 CADWorx 2004 ............................................ 23

> see story page 3

Problems Exporting to Word

> see story page 7

CAESAR II 4.50

Technology You Can Use

> see story page 8

The COADE Mechanical Engineering News Bulletin is published twice a year from the COADE offices in Houston, Texas. The Bulletin is intended to provide information about software applications and development for Mechanical Engineers serving the power, process and related industries. Additionally, the Bulletin serves as the official notification vehicle for software errors discovered in those Mechanical Engineering programs offered by COADE. 2003 COADE, Inc. All rights reserved.

Integration of Tubesheet and Expansion Joint Analysis ............................................. 9 Satisfying Expansion Load Case Requirements .......................................... 11 Mass Spacing for Dynamic Analysis ............ 13 Assessing the Metal Loss Flaws using API Recommended Practice 579 ............ 14 PC Hardware/Software for the Engineering User (Part 34) ......................................... 18

API-579

> see story page 14

Dealing with Spam

> see story page 19

CADWorx 2004

Program Specifications
CAESAR II Notices ...................................... 20 TANK Notices .............................................. 22 CodeCalc Notices ........................................ 22 PVElite Notices ............................................ 23

> see story page 23

COADE Mechanical Engineering News CodeCalc version 6.5 also includes PD 5500 tubesheet rules (British code), for u-tube, fixed and floating tubesheets. Modifications were also made in the calculation of required fixed tubesheet thickness, to properly account for the non-linearity in its formula. In previous versions it may have been necessary to manually iterate on a required thickness. Now, CodeCalc performs this iteration automatically. Improvements were also made in U-tube and fixed tubesheet summaries. Fitness for Service using API 579 CodeCalc 6.5 implements API 579 Sections 4 and 5 covering Level 1 and Level 2 flaw assessments for metal loss on cylindrical shells, simple cones, and formed heads. Another article in this newsletter has an in-depth discussion of this capability with a solved example. Fitness for Service (FFS) assessment using API Recommended Practice 579 is performed to assess the operation, safety and reliability of the process plant equipment, such as pressure vessels, piping and tanks for some desired future period. The assessment procedure provides an estimate of the remaining strength of the equipment in its current state, which may have been degraded while in-service from its original condition. Color Syntax Highlighting in the Output Reports Color has been added to the output reports to highlight important values, thus increasing the readability of the reports. Important notes appear in blue while headings are in bold and failures/warnings/ errors are indicated in red color. This feature was added in the April 4th build of CodeCalc and PVElite. The color highlighting is also preserved when printing these reports and when exporting to MS Word. Enhancement in the Nozzle module For hillside or other non-central nozzles on elliptical or torispherical heads, the program now prints a warning if the nozzle is outside the spherical portion of the head and the user has indicated otherwise. A small nozzle is nozzle that does not require an area replacement calculation due to its smaller size. But, some small nozzles that are in close proximity to each other may require these calculations. In this version, the switch to force the program to perform area replacement calculation on small nozzle was moved from the file level to each individual nozzle input. Added WRC 107 Auto-calc on support lug

July 2003

An option was added in the support lug dialog to perform WRC-107 calculations without leaving the Leg Lug module. See the dialog in the figure below.

In the shell, nozzle and flange modules input errors (if any) are now displayed on the screen at the time of input. Here is an example:

The ASME Material database now displays the Class/Thickness of materials in the list view. This will help in finding the right material when more than one occurrence of a material is listed in the ASME database.

July 2003 A few of the new features in PVElite version 5.0 are: Enhancements listed in the CodeCalc part. Added ASME A-2002 updates, including revision to material databases Integrated 3D graphics into Input Added option to model Sump Elements Added Nozzle on Nozzle capability Implemented export to ODBC database (MS Access) Added OD basis support is for Division 1, vessels Added color syntax highlighting in the output, e.g. questionable results are shown in red color Added PD:5500 (British code) Annex C, fluctuating pressure fatigue analysis Added criterion of PD:5500 category C vessels Added bolting requirements for Horizontal vessels Added on screen calculations for Nozzle Weld sizes and a WRC 107 scratch pad Enhancements to the load case combinations for longitudinal stress Added option for 0.90*Yield for hydro-test allowable (Div 1) Added weld sizes for basering chair caps Improvements to the MS Word report generation Major update of the printed documentation.

COADE Mechanical Engineering News

COADE Releases CADWorx EQUIPMENT

(by: Scott Nickel)

COADE has announced the release of the latest module in its CADWorx 2004 Plant Design Suite, CADWorx EQUIPMENT. CADWorx EQUIPMENT adds the ability to quickly and easily create AutoCAD-based 3D models of equipment horizontal or vertical, single or multi-state vessels, heat exchangers and pumps to the piping, P&ID, structural steel, HVAC, and instrument loops capabilities already available with CADWorx 2004. CADWorx EQUIPMENT works on a hierarchal basis. The process begins by identifying the equipments component parts, such as a vessels heads, shells, nozzles and supports, via icons selected from the toolbar. Entering required dimensional data in the property editor instantly, parametrically updates the model. Equipment components, or details on the components, can be cut and pasted for easy modeling and modification of equipment. Selection lists containing flange ratings, flange facings, motor frames and more are available from the programs property editor. COADE has long been on the forefront of the industry as far as CAD-to-analysis integration is concerned, with its CADWorx PIPE to CAESAR II seamless bi-directional interface. The same strategy is brought to CADWorx EQUIPMENT, which offers a seamless, bi-directional interface between vessel models built in this module and PVElite, COADEs vessel analysis software. A vessel built in CADWorx EQUIPMENT may be exported to a native PVElite file, analyzed and modified, and then re-imported to ensure that all changes made during the analysis process are reflected back into the CAD model. Likewise, any vessel built in PVElite may be used to generate a 3D CAD model in CADWorx EQUIPMENT, ensuring accuracy and saving modeling time.

COADE Mechanical Engineering News

July 2003

COADE Announcement Procedures

(by: Richard Ay)

In an effort to keep all users informed of software releases, software updates, and other important occurrences, COADE directly e-mails registered users. This e-mail process is controlled by a web-based List Server. This procedure was introduced in the February 2002 issue of this newsletter. There are several important points to note about this announcement procedure. How do you register for this service? From the Help menu of COADEs engineering products, and the Pipe menu of the CADWorx application, is a link for On-Line Registration. Additionally, the first time you run the software following an installation, you are prompted to register. Either alternative takes you to a web page for you to fill in your contact information. Once completed, this information is stored on a List Server. (This List Server is maintained by a third party. COADE does not maintain this server.) If you dont register, you will not receive our announcements. When registering, if you have the capability to receive HTML email, you should check the appropriate checkbox indicating this. The HTML e-mails are laid out better, and are more informative than the plain text messages. How do you update your information? When your contact information (such as e-mail address) changes, you can follow the link at the bottom of any announcement to access your profile on the List Server. You should then make the necessary changes to your profile. What should you do if you no longer want COADE notices? If at some point you decide you no longer want to receive COADE announcements, you can follow the link at the bottom of any announcement to access your profile on the List Server. You can then opt out of future announcements. Same Vessel transferred into PVElite Please do not reply to the announcements: The e-mail announcements sent by the List Server contain a return address of info@coade.com. This is not a real e-mail address. We eventually do get the reply, but these responses are directed to an un-attended mailbox. The announcement itself includes the proper contact information if you need to communicate directly with COADE.

Vessel in CADWorx/Equipment

July 2003 What should you do if you registered, but don't receive announcements? Every time an announcement is distributed, we learn from the List Server that roughly 10% of the messages bounced back. This indicates bad or changed user e-mail addresses. When two consecutive messages bounce, the List Serve deletes the offending profile from its database. This means that you will no longer receive COADE announcements, even though you think you are registered. In effect, the List Server un-registered your profile. If you suspect this may have happened, please register again, as described above. We are hoping that the use of this List Server provides an additional service to our users. However, the success of this service depends on users maintaining their correct contact information.

COADE Mechanical Engineering News There are two groups of tests. The first set of tests used a Performance Test program from the HOOPS vendor. The tests performed here consisted of drawing simple shapes and text, over and over again, in various positions. These tests indicate that the better your graphics board, especially the more graphics memory available, the better the performance. The second set of tests consisted of plotting a number of CAESAR II jobs on the test machines. The results of these tests show the expected interaction between the CPU and graphics board. For example, the dual 700 Mhz with 8 Mbyte graphics card performed better than all but one of the machines with a single processor and 8 times the graphics memory! These tests indicate that in addition to a good graphics board, you also need either a very fast single CPU, or dual CPUs of medium speed. To illustrate the performance improvement made between different versions of HOOPS, in the CAESAR II tests, a number of the results are shown in blue, and are noted as Ver 8.12. These results were obtained with the graphics released for CAESAR II Version 4.40 Build 030403, using the 8.12 version of the HOOPS library. (Previous builds of CAESAR II Version 4.40 used the 8.00 version of the HOOPS library.) Depending on the job, the speed improved by a factor of from 4 to 30. This improvement can be attributed to 8.12 version of HOOPS, which now draws more primitives directly using the hardware, instead of COADE drawing them in software. In addition, most video cards now have OpenGL built-in, which allows HOOPS to push the rendering all the way down to the hardware, where before, most of the drawing had to be done with the CPU. Video cards with a lot of memory have big z-buffers, plus good optimization, which helps them avoid drawing things that will be obscured by objects on top. While the HOOPS library and COADE software can improve and optimize, the best performance can be obtained only by also utilizing fast hardware. This is one of the key concepts to grasp. By default, HOOPS utilizes the OPENGL capabilities of your system. Utilizing graphics cards with good OPENGL acceleration will improve the overall performance of the applications. (CAESAR II users note that Version 4.50 will provide even faster performance. The model may be manipulated while it is being drawn.)

Graphics Speed Issues

(by: Richard Ay)

COADEs engineering programs (CAESAR II, PVElite, and CodeCalc) utilize the HOOPS 3D graphics engine. This engine is 3rd party software, not developed by COADE. Using a 3rd party graphics engine provides a number of benefits to users: State of the art graphics technology (such as rendering and calculations in hardware) Faster implementation of the advances in new hardware Provides a more uniform handling of graphics across a variety of operating systems However, the draw back to using a 3rd party package is that the software is limited to the performance and capabilities of the 3rd party package. Recently, a number of CAESAR II users and dealers have expressed concern that these 3D graphics are slow. Slow is a relative term, slow compared to what? Each release of CAESAR II since Version 4.20 has offered a HOOPS speed improvement of at least 40%. These improvements are due to optimization of the CAESAR II code (to take better advantage of the capabilities of the HOOPS library), and improvements in the base HOOPS library. In an attempt to provide specific answers to this question, several performance tests using the HOOPS 3D graphics were performed on a number of COADE computers. The details of the hardware used can be found on the accompanying spreadsheet. Note that the test machines encompassed a wide range of CPUs (from dual 300 Mhz to single 2.8 Ghz) and a variety of graphics boards (from 8 Mbytes to 128 Mbytes).

COADE Mechanical Engineering News

July 2003

HOOPS Graphics System Speed Tests Setup A B C Dual 700 Mhz, 1Gbyte Ram Windows 2000 D E F 1Ghz, 512Mbyte Ram Windows 2000 Dual 700, Dual 300 Mhz, Machine Description 512Mbytes Ram 512Mbyte Ram Operating System Windows 2000 Diamond Multimedia Fire Gl 1000 Pro, Graphics Board Description 8Mbytes, AGP Techsoft Test 2) 3D Edges/sec 4) 3D markers/sec 7) 3D polygons/sec 8) 3D edgeless polygons/sec 9) 3D lit edgeless polygons/sec 11) 3D hello worlds/sec 14) unlit shells/sec 15) flat lit shells/sec 16) gouraud lit shells/sec Techsoft Test Notes: 1) Test results were obtained using Techsoft's performance test program "PT.EXE". 2) On all machines, the PT.EXE options of double buffering and culling were turned on. 3) Of the 16 tests performed, only those noted above were compared. 4) Tests results shown in red are those functions used most by COADE products. CAESAR II Test Job Description Plot Time fw-oper, 1910 elements, 593 restraints, 135 rigids. (12M allocated) Ver 8.00 - 37 sec rev-beattock28, 727 elements, 473 restraints, 80 rigids. (12M allocated) 100yrs1, 3935 elements, 3362 restraints, 0 rigids. (32M allocated) COADE Test Notes: 1) Tests were timed using the Task Manager. 2) Times are from start of plot request to active toolbar. 3) Ver 8.00 HOOPS was released with the initial CAESAR II Version 4.40. 4) Ver 8.12 HOOPS was released for CAESAR II Version 4.40, build 030403. Plot Time Plot Time Ver 8.12 - 33 sec Ver 8.00 - 612 sec Ver 8.12 - 18 sec Ver 8.00 -235 sec Ver 8.12 - 76 sec No Ver 8.00 Plot Time Plot Time Ver 8.12 - 8 sec Ver 8.00 -162 sec Ver 8.12 - 7 sec Ver 8.00 - 61 sec Ver 8.12 - 21 sec Ver 8.00 - 630 sec Plot Time 33,865 40,355 17,338 26,905 79,421 2,662 875 863 819 30,059 49,556 15,039 26,765 24,585 3,364 981 879 826 1,164,189 1,717,821 554,454 934,892 807,252 17,000 46,531 37,923 36,428 114,472 222,819 143,572 215,138 216,095 18,832 3,895 3,894 3,893 227,625 226,387 215,665 413,939 414,924 27,546 7,732 7,732 7,771 1,308,552 1,460,214 802,499 1,265,078 1,119,365 32,869 44,263 41,039 39,117 Windows 2000 1.9 Ghz, 512Mbyte 2.8 Ghz, Ram 1Gbyte Ram Windows XP Nvidia GeForce2 Mx400, 64Mbytes, AGP Windows XP

AccelStar II, 8Mbytes, AGP

Winfast A170, 64Mbytes, AGP

Nvidia GeForce4 Radeon 9000 IF MX 440, 64Mbytes, Pro, 128Mbytes, AGP AGP

Ver 8.00 - 237 sec

Ver 8.00 - 48 sec Ver 8.00 - 194 sec

Ver 8.00 - 82 sec

Ver 8.00 - 844 sec

July 2003

COADE Mechanical Engineering News

Exporting Output to MS Word

(by: Richard Ay)

All COADE engineering products include an option to export output data directly to MS WORD. Since introducing this capability a small but noticeable percentage of our users have been unable to utilize this option. Initial investigation revealed two reasons for this difficulty: 1) Failure to register the outword.dll DLL with the system. Manually registering the DLL using regsvr32.exe usually resolved this issue. 2) Norton Anti-Virus installations by default turn off scripting abilities. This prevents macros from running, which disabled the COADE interface into MS Word. Some versions of NAV (Norton Ant-Virus) allowed users to configure NAV to allow scripting. This allowed some users to then send COADE output to MS WORD. However, enough problems persisted so that we rewrote our MS WORD interface. We abandoned the macros and wrote the necessary procedures into the outword.dll DLL. This resolved more issues, but not all. A number of users were still facing the Unable to launch MS WORD message. This message is a COADE message that indicates the COADE product is installed and functioning as designed, but that WORD failed to start. The problem was finally traced, again, to Norton Anti-Virus. NAV inserts a key in the System Registry that forces MS WORD to ask permission before starting. NAV denies permission to all applications except Internet Explorer. (This is not good because this registry key is buried beneath a Microsoft Word key, in a part of the registry users should not really be adjusting.) However, the solution is to remove this key and contact Norton. Unfortunately, this isnt a permanent solution because NAV will re-insert the key, on some machines everytime it reboots. Even though some versions of NAV have an enable Word Automation option, it doesnt correct this problem with the registry. Some versions of NAV insert this key and provide NO enable Word Automation switch. Uninstalling NAV does not remove this key from the registry! Norton says they are working on the problem see the information in the figure below, from Nortons web site. For those who want to edit their registry and remove this key, perform these steps. 1. 2. 3. Click the Start|Run menu, then type REGEDIT in the Run dialog box. Expand the HKEY_CLASSES_ROOT\CLSID key. Scroll down the list until you see {00020906-0000-0000C000-000000000046} and expand it. There are several keys that look alike, or differ by only one number, so make sure that the one you choose matches exactly. This is the secret code used by Microsoft to determine when and how Windows will start MSWord. Under this key will be an entry named InProcServer32. Right-click it and select Delete. The value that is stored here points to a Norton file named Symantec Shared/Script Blocking/scrblock.dll. This part of the registry should look like the figure below.

4.

COADE Mechanical Engineering News

July 2003

CAESAR II Version 4.50

(by: Richard Ay)

CAESAR II Version 4.50 will likely be in Beta Testing by the time you read this. In addition to piping code revisions, some of the other enhancements for this release include: Revised material database for B31.1 A2002 changes Load Case Template implemented for recommending static load cases. Reducer element added. Major graphics improvements, including: 5. 6. Close REGEDIT. CAESAR II should be able to invoke MSWord now, at least until NAV modifies the registry again. A walk-through option is available. The static output processor can now produce colored stress plots of the piping system. A graphical find (zoom to) option has been added. Model drawing during CPU idle time.

The static output processor remembers all user settings (filters, labels, and report size). New dynamic (HTML) help system for piping input and configuration. Automatic acquisition of website software updates. Combined WRC-107/297 module for local stress calculations The structural steel interface has been redesigned for easier operation. Spectrum generation wizard

If modifying the registry isn't an option, the only solution we are aware of is to remove NAV from the machine. Users who continue to have problems exporting data to MS WORD, and have had NAV installed at anytime on their machine, should contact Norton directly.

July 2003

COADE Mechanical Engineering News Tubes tab: Input on the tubes tab was reorganized for consistency. Check the box Tube to Tubesheet Joint information, to enter information about the Tube-Tubesheet joint. CodeCalc will use this information to check tube to tubesheet welds, and in the case of fixed tubesheets, compute the allowable tube to tubesheet joint load.

Integration of Tubesheet and Expansion Joint Analysis

(by: Mandeep Singh)

In previous versions of CodeCalc, the fixed tubesheet and the thick Expansion Joint (flanged and flued) modules were not integrated. If the exchanger design required a thick expansion joint, a manual transfer of some information (spring rates and prime pressures) between the tubesheet results and the expansion joint input must have been made. In version 6.50 these two modules have been integrated. Hopefully this automation reduces the design time of these elements and reduces transcription errors. In this version, when multiple fixed tubesheet load cases are analyzed, the corresponding expansion joint cases are automatically run. A summary of results is provided at the end of the report. A single execution of the TEMA tubesheet module can accomplish what required many different runs and manual data transfer between modules, in the previous version. Discussion of the New Input On the Shell tab the design code can be selected from the drop down box. The British code (PD 5500) is available in conjunction with the TEMA code. This allows the program to customize the input per the appropriate code selected. Shell tab:

Using the Tube joint type and the tested check box, program automatically puts in the fr joint factor.

Option to specify design code

COADE Mechanical Engineering News Tubesheet Tab:

July 2003

This button sets default expansion joint dimensions, based on the shell thickness and material

Procedure for analyzing a Fixed Tubesheet with a Thick Expansion Joint:

This button merges the flange, gasket and bolting input from an existing flange, into this tubesheet.

Typically in the first pass a fixed tubesheet is analyzed without an expansion joint. If the configuration (tubes, shell or tubesheet) does not pass, and if the cause of the failure is due to differential thermal expansion, a thick or thin expansion joint can be added. If a thick (flanged and flued) expansion joint is selected, CodeCalc follows these steps: 1. The axial spring rate of the expansion joint is computed in both the corroded and new conditions. The expansion joint spring rate is used to compute equivalent differential pressure. Next the program extracts the prime pressures (Ps, Pt, Pd) from the output of the tubesheet calculation and uses those values to compute the expansion joint stresses. If multiple tubesheet load cases are selected, a corresponding expansion joint analysis is automatically performed.

This button merges the flange, gasket and bolting input from an existing flange, into this tubesheet input. Expansion Joint Tab:

2. 3.

4.

Results of all the runs are summarized in tabular format like the one below:
Fixed Tubesheet Required Thickness per TEMA 8th Edition: Reqd. Thk. + CA P r e s s u r e s Case Pass/ Case# Tbsht Extnsn Pt Ps PDif Type Fail 1uc 1.471 0.000 71.07 0.00 0.00 Fvs+Pt-Th-Ca Ok 2uc 0.750 0.000 0.00 2.39 0.00 Ps+Fvt-Th-Ca Ok 3uc 1.471 0.000 71.07 2.74 0.00 Ps+Pt-Th-Ca Ok 4uc 0.757 0.000 0.00 0.00 -37.66 Fvs+Fvt+Th-Ca Ok 5uc 1.471 0.000 71.04 0.00 -39.07 Fvs+Pt+Th-Ca Ok 6uc 0.784 0.000 0.00 2.65 -37.75 Ps+Fvt+Th-Ca Ok 7uc 1.471 0.000 71.04 2.74 -39.07 Ps+Pt+Th-Ca Ok 8uc 0.750 0.000 0.00 0.00 0.00 Fvs+Fvt-Th-Ca Ok 1c 1.491 0.000 70.45 0.00 0.00 Fvs+Pt-Th+Ca Ok 2c 0.775 0.000 0.00 2.25 0.00 Ps+Fvt-Th+Ca Ok 3c 1.491 0.000 70.45 2.64 0.00 Ps+Pt-Th+Ca Ok 4c 0.839 0.000 0.00 0.00 -43.40 Fvs+Fvt+Th+Ca Ok 5c 1.490 0.000 70.42 0.00 -45.02 Fvs+Pt+Th+Ca Ok 6c 0.863 0.000 0.00 2.55 -43.50 Ps+Fvt+Th+Ca Ok 7c 1.490 0.000 70.42 2.63 -45.02 Ps+Pt+Th+Ca Ok 8c 0.775 0.000 0.00 0.00 0.00 Fvs+Fvt-Th+Ca Ok Max: 1.491 0.000 in. Given Tubesheet Thickness: 2.0000 in. Note: Fvt,Fvs 0.0. Ps, Pt Th Ca

The above screen becomes active only in the case of fixed tubesheet exchangers. The expansion joint can be either a thin (bellows type) or thick (flanged/flued type) or there can be no joint at all. For a thin expansion joint, only the axial spring rate needs to be specified. For a thick expansion joint, either the spring rate needs to be specified (Design option set to Existing) or analyze the expansion joint geometry and allow CodeCalc to compute the spring rate and expansion joint stresses (Design option set to Analyze).

- User-defined Shell-side and Tube-side vacuum pressures or - Shell-side and Tube-side Design Pressures. - With or Without Thermal Expansion. - With or Without Corrosion Allowance.

10

July 2003
Tube and Shell Stress Summary: Shell Stresses Tube Stresses Tube Loads Pass Case# Ten Allwd Cmp Allwd Ten Allwd Cmp Allwd Ld Allwd Fail 1uc 75 20000 0 -13887 2163 19300 0 -5740 303 2160 Ok 2uc 67 20000 0 -13887 0 19300 -138 -5740 19 2160 Ok 3uc 142 20000 0 -13887 2163 19300 -138 -5740 303 2160 Ok 4uc 479 20000 0 -13887 0 19300 -1346 -5395 0 2160 Ok 5uc 517 20000 0 0 2158 19300 -1346 -5395 302 2160 Ok 6uc 479 20000 0 -13887 0 19300 -1483 -5395 19 2160 Ok 7uc 517 20000 0 0 2158 19300 -1483 -5395 302 2160 Ok 8uc 0 20000 0 -13887 0 19300 0 -5740 0 2160 Ok 1c 89 20000 0 -13824 2170 19300 0 -5740 304 2160 Ok 2c 67 20000 0 -13824 0 19300 -136 -5740 19 2160 Ok 3c 156 20000 0 -13824 2170 19300 -135 -5740 304 2160 Ok 4c 568 20000 0 -13824 0 19300 -1570 -5395 0 2160 Ok 5c 613 20000 0 0 2166 19300 -1570 -5395 303 2160 Ok 6c 568 20000 0 -13824 0 19300 -1705 -5395 19 2160 Ok 7c 613 20000 0 0 2166 19300 -1705 -5395 303 2160 Ok 8c 0 20000 0 -13824 0 19300 0 -5740 0 2160 Ok MAX RATIO 0.031 0.000 0.112 0.316 0.141

COADE Mechanical Engineering News 1) 2) 3) 4) 5) W+T1+P1 (OPE) W+T2+P1 (OPE) W+P1 (SUS) L1-L3 (EXP) L2-L3 (EXP)

What the article explains is that one more load case is required to completely satisfy the intent of the code, to address the phrase or any anticipated condition with a greater differential effect. This greater differential is created by cycling between T1 and T2. Therefore, to completely satisfy the intent of the code, another load case must be setup as follows: 6) L1-L2 (EXP) CAESAR II doesnt setup this last load case, since the program doesnt know what the loads (T1 and T2) represent. The construction of load case 6 above is the users responsibility. Other situations exist where the user must review the load cases recommended and consider whether or not they completely satisfy code requirements. For example, consider the system shown in the figure below, having a single operating temperature, but where at any given time, one of the pump branch legs could be spared.

Summary of Expansion Jt Results, (displays the worst case): Category Max. Stress Allowable Location Ld Case Pass (psi) (junction) Tsht ExpJt Fail Annul. Elm. -38456. 65000. Inside 5c Pt+Pd Pass Cyl. at Y -989. 65000. Inside 7uc Ps+Pt+Pd Pass Cyl. at L 850. 65000. Inside 5c Pt+Pd Pass Cycle Life 82089 10000 Inside 7c Pt+Pd Pass

In conclusion, the integration of thick expansion joint and fixed tubesheet analysis provides a solution technique that is easier and less prone to input errors. Also, the overall length of the printed reports is reduced.

Satisfying Expansion Load Case Requirements

(by: Richard Ay)

Typical Power and Process piping codes evaluate the stresses of a piping system under in three different states; sustained (or primary), expansion (or secondary), and occasional. The focus of this article is on the proper evaluation of the expansion stresses, and the corresponding load case setup. The B31.3 code in Paragraph 319.2.3.b states While stresses resulting from displacement strains diminish with time due to yielding or creep, the algebraic difference between strains in the extreme displacement condition and the original (as-installed) condition (or any anticipated condition with a greater differential effect) remains substantially constant during any one cycle of operation. This difference in strains produces a corresponding stress differential, the displacement stress range, which is used as the criterion in the design of piping for flexibility. A previous article in this newsletter (Expansion Case for Temperatures Below Ambient, May 1993, p32) discusses this requirement, using an example situation where the system has two temperatures, one above ambient and one below ambient. To summarize, in this situation, CAESAR II would recommend the following load cases:

Example System from WRC-449

11

COADE Mechanical Engineering News To analyze this system, the following temperature vectors could be defined. T1 vector entire system at operating temperature to design hangers T2 vector entire system at operating temperature except leg 1 at ambient T3 vector entire system at operating temperature except leg 2 at ambient T4 vector entire system at operating temperature except leg 3 at ambient 12) L3-L4 (EXP) 13) L3-L5 (EXP) 14) L3-L6 (EXP) 15) L4-L5 (EXP) 16) L4-L6 (EXP) 17) L5-L6 (EXP)

July 2003 OPE case A minus OPE case B OPE case A minus OPE case C OPE case A minus OPE case D OPE case B minus OPE case C OPE case B minus OPE case D OPE case C minus OPE case D

This situation is different than the one discussed in the previous newsletter article. There is only one operating temperature. However, to satisfy the intent of the code (the extreme displacement stress range), what are the necessary load cases? CAESAR II will recommend the following set of load cases: 1) W (HGR) 2) W+D1+T1+P1 (HGR) 3) W+D1+T1+P1+H (OPE) 4) W+D2+T2+P1+H (OPE) 5) W+D3+T3+P1+H (OPE) 6) W+D4+T4+P1+H (OPE) 7) W+P1+H (SUS) 8) L3-L7 (EXP) 9) L4-L7 (EXP) 10) L5-L7 (EXP) 11) L6-L7 (EXP) Are these cases sufficient? The answer is no, they are not sufficient. The system could cycle between OPE case B and OPE case C, or between OPE case B and OPE case D, or between OPE case C and OPE case D. So to completely satisfy the intent of the code, the following additional load cases must be setup. restrained weight case for hanger design hot case for hanger design OPE case A with all legs hot OPE case B with leg 1 spared OPE case C with leg 2 spared OPE case D with leg 3 spared Sustained (ambient) case OPE case A minus Sustained OPE case B minus Sustained OPE case C minus Sustained OPE case D minus Sustained

These six additional cases consider the effects of the system cycling between the different possible operating states. This cycling can cause the extreme displacement range the code requires. CAESAR II has no knowledge of what OPE cases 3, 4, 5, and 6 represent, therefore the program is unable (at the present time) to include cases 12 through 17 when it performs its recommendations. These additional load cases are the responsibility of the user. According to the code, the expansion stress range SE is the largest computed displacement stress range. However, SE could come from different load combinations, which is a point many analysts miss. For example, consider the metering station shown in the figure below.

Metering Station Either leg could be hot, with the other leg cold. The greatest stress on the tees occurs when switching from one leg to the other. The difference between these two operating conditions will produce the extreme condition for the proper stress evaluation of the tees. Understanding the requirements of the applied piping code, as well as what the recommended load cases represent, is necessary in determining if the intent of the code is completely satisfied, or if additional load cases are necessary.

12

July 2003

COADE Mechanical Engineering News Setting (n) to 1 (indicating the first mode of vibration) and setting (f) to the cutoff frequency (for the eigen extraction) yields the minimum wavelength of interest. (This is an important point, which indicates that the minimum wavelength depends on the type of dynamic analysis being performed.) Now that we have the minimum wavelength of interest in the model, idea is to provide sufficient mass points along this span to adequately model this mode (frequency). This can be accomplished by introducing a constant into the above equation, resulting in:
EIg (n ) 2 * 2f w

Mass Spacing for Dynamic Analysis

(by: Richard Ay)

Many times when constructing a model for static analysis in CAESAR II, node points are defined only when data changes. Examples of this are: pipe property changes, load changes, geometry changes, and boundary conditions. In most instances, this nodal layout is sufficient for a static analysis. However, if it is necessary to evaluate the system for dynamic response, the typical nodal layout for a static analysis may be insufficient. This is because in a dynamic analysis of a piping system, the mass is lumped at the node points. Insufficient nodal spacing causes insufficient mass lumping, leading to inaccuracies in the dynamic solution. Many codes, standards, and technical papers provide similar equations and guidelines for determining the maximum nodal spacing for dynamic analysis. What is the origin of these equations, and how can they be applied to piping systems? The basis for the maximum nodal (mass) spacing is founded on the Euler beam equation. Assuming a simply supported beam, the Euler beam equation relates the circular frequency of harmonic motion ( ) to the length of the beam (l), its flexural rigidity (EI), and its mass per unit length (w/g). The mass per unit length (w/g) should include the contribution from the pipe, the fluid contents, and any insulation if applicable.
EIg ( n ) 2 * 2 l w

l2 = k2 *

This equation therefore yields the maximum recommended distance between the mass points. How does one determine the constant k? Work has been done that shows when 3 intermediate mass points are used along the span, an accuracy of 99.7% is achieved for the first mode of vibration (of the span). When 2 intermediate mass points are used, an accuracy of 99% is achieved. (Of course, all frequencies below the cut-off frequency will be even more accurately modeled.) Relating the number of mass points to the constant k means that for 3 mass points the span is broken into fourths, thus k is . Similarly, for two mass points the span is broken into thirds, thus k is 1/3. Therefore the value of (k) is chosen based upon the accuracy desired. To aid COADE users in determining the suggested maximum nodal spacing (the distance between mass points), a small utility program has been developed. The first step in utilizing this utility is to select the desired units system, as shown in the figure below.

n =

The term (n )2 is valid for simply supported beams only, where (n) is the mode of vibration. This equation can be easily rearranged to solve for the length (l), which will correspond to a specified frequency. Substituting = 2 f, the equation used to determine the span length corresponding to frequency (f in Hz) is:
EIg ( n ) 2 * 2f w

l2 =

How do we use this? The vibrating wave in a pipeline can be approximated as the vibration of a simply supported pipe (beam). Therefore this equation can be used to calculate the distance between nodes (points of no movement) in the vibrating wave (this is the half wavelength).

13

COADE Mechanical Engineering News Once this selection has been made, the Pipe Data tab can be presented. Filling in the necessary data and clicking on the [Calculate] button yields the maximum suggested nodal spacing, as shown in the figure below.

July 2003 Typical approaches for FFS (Fitness For Service) as indicated in API 579 are as follows: Identifying the Flaw and Damage Mechanism. Reviewing the Applicability and Limitations of the FFS Assessment Procedures. Gathering data. Applying the assessment techniques and comparing the result to the acceptance criteria. Estimating the remaining life for the inspection interval. Applying remediation as appropriate. Applying in-service monitoring as appropriate. Documenting the results

Common degradation mechanisms include general corrosion, localized corrosion, pitting corrosion, blister, mechanical distortion etc. The procedures on how to assess these common flaws are discussed in the sections described in the Table of Contents of the API 579 document. This utility program can be acquired from the CAESAR II download area of the COADE web site. This program can be used to check the maximum nodal spacing of models before running dynamic analysis. Section 1 Introduction Section 2 Fitness-For-Service Engineering Assessment Procedure Section 3 Assessment of Equipment for Brittle Fracture Section 4 Assessment of General Metal Loss Section 5 Assessment of Local Metal Loss Section 6 Assessment of Pitting Corrosion Section 7 Assessment of Blisters and Laminations Section 8 Assessment of Weld Misalignment and Shell Distortions Section 9 Assessment of Crack-Like Flaws Section 10 Assessment of Component Operating in the Creep Regimes Section 11 Assessment of Fire Damage

Assessing the Metal Loss Flaws using API Recommended Practice 579
(by: Kevin Kang)

Corrosion and groove-like flaws are common problems that are experienced by vessels in service. Loss of metal through the vessel wall thickness reduces the strength of the component. At some localized points, these flaws may reduce the vessel wall thickness below a minimum Code requirement. Rather than replace the vessel with a new one, the integrity of the vessel may be checked to determine its remaining life and whether it can continue to operate at some specified conditions. API document Recommended Practice 579 can be used to evaluate the integrity and operational safety of process plant equipment, such as pressure vessels, piping and storage tanks. The results of the assessment procedure will provide an estimate for the strength and the remaining life of the equipment.

The recently released PVElite 5.0 and CodeCalc 6.5 programs have included metal loss assessments according to API 579 Section 4 and Section 5 for vessel elements such as cylindrical shells, simple cones and formed heads. The analysis can be performed using the Shell and Head Module as depicted in Figure 1 below.

14

July 2003

COADE Mechanical Engineering News Section 4 covers FFS assessment procedures for components subject to general metal loss resulting from corrosion and/or erosion. Section 5, on the other hand, is a method for analyzing local metal loss or Local Thin Areas (LTAs) that include groove-like flaws or gouges. In general, flaw assessment using Section 4 provides conservative results. The differences between Section 4 and 5 when applied to LTAs are as follows: Section 4: rules for all Level 1 and 2 assessments are based on average thickness averaging approach in which is used with Code rules to determine acceptability for continued operation. Section 5: rules for Level 1 and Level 2 assessments are based on establishing a Remaining Strength Factor (RSF) in which is used to determine acceptability for continued operation.

Figure 1. API 579 Analysis Selection The assessment type either using Section 4 or Section 5 can be selected from the API 579(FFS) tab, as shown in Figure 2.

The Assessment of General Metal Loss described in Section 4 can be performed using either point thickness (random type readings) or profile thickness (grid type readings) measurement data. The selection of the data type readings can be made in the Data Measurement tab as shown in Figure 3a. The API Recommended Practice 579 requires a minimum of 15 measurement data. CodeCalc can accommodate up to 99 points.

Figure 2. General and Local Metal Loss Selection

Figure 3a. Data Measurement Type Selection

15

COADE Mechanical Engineering News The localized metal loss assessment described in Section 5, however, can only be performed using profile thickness data. The data matrix can be set up by providing the number of points in both circumferential and longitudinal directions. The matrix size in this case is limited to maximum 9x9. For convenience, Critical Thickness Profile (CTP) data entry is also provided. The measurement data grid dialog is pictured below in Figure 3b.

July 2003 API 579 Section 4 limitations for Level 1 and Level 2 assessments are as follows: Original design in accordance with a recognized code or standard. The component is not operating in the creep range. The region of metal loss has relatively smooth contours without notches. The component is not in cyclic service (less than 150 total cycles). The component under evaluation does not contain crack-like flaws. The component under evaluation has a design equation in which specifically relates pressure and/or other loads, as applicable, to a required wall thickness. With some exception, the following specific components not having equation relating pressure and/or other loads to a required wall thickness may be evaluated using Level 2: Pressure vessel nozzles and piping branch connections Cylinder to flat head junctions Integral tubesheet connections Flanges Piping systems

Figure 3b. Profile Thickness Data Grid For most evaluations, it is recommended to first perform the assessment using Section 4, then move on using Section 5 if necessary. The rules in Section 4 have been structured to provide consistent results with Section 5. However, it is the engineers responsibility to review the Assessment Applicability and Limitation whenever the assessment is changed. When the acceptance criteria either passes or fails, a respective remaining life using a thickness approach or a de-rated value of the MAWP of the vessel will be calculated automatically. There are three (3) levels of evaluations available for each flaw type described in general as follows: Level 1 - typically involving a simplified method using charts, simple formulae, and conservative assumptions. Level 2 - generally requires more detailed evaluation and produces more accurate results Level 3 - allows flaw assessment using more sophisticated methods such as FEA.

Note: Currently CodeCalc does not support API 579 analysis on nozzle, flange, tubesheet, flathead, and piping system components. The following limitations on applied loads are satisfied: Level 1 assessment: components are subject to internal and/or external pressure (negligible supplemental loads). Level 2 assessment: components are subject to internal and/or external pressure and/or supplemental loads such as weight, wind and earthquake.

Limitations for the API 579 Section 5 Level 1 and Level 2 assessments are similar to the limitations for Section 4 above with the following additions: The components cannot be subjected to external pressure, or if the flaw is located in the knuckle region of elliptical head (outside of the 0.8D region), torispherical/toriconical head, or conical transition.

16

July 2003 The material component is considered to have sufficient material toughness. Special provisions provided for groove-like flaws such as: Groove (no mechanical cold work). Gouge (mechanical cold work).
Input Echo, Component

COADE Mechanical Engineering News

1, Description: E5111 P 300.00 650.00 NO SA-516 70 K02700 17500.00 20000.00 B 1.00 120.0000 20.0000 96.0000 1.2500 1.2500 0.1250 Inside 0.1000 0.9000 60.0000 300.00 0.0010 No Profile 5 9 1.0000 0.5000 NO Cylindrical Shell 1, Desc.: E5111 96.2000 1.1500 in. in. psig F

Design Internal Pressure Temperature for Internal Pressure Include Hydrostatic Head Components Material Specification (Not Normalized) Material UNS Number Allowable Stress At Temperature Allowable Stress At Ambient Curve Name for Chart UCS 66 Joint efficiency for Shell Joint Design Length of Section Length of Cylinder for Volume Calcs. Inside Diameter of Cylindrical Shell Minimum Thickness of Pipe or Plate Nominal Thickness of Pipe or Plate Future Corrosion Allowance Flaw Location Uniform Thickness Loss Allowable Remaining Strength Factor Minimum Dist. to a Major Struct. Disc. User Input MAWP Annual Corrosion Rate Near Axisymmetry Discontinuity

S SA E L CYLLEN D T T FCA XLOSS RSFA Lmsd MAWP Crate

psi psi

It is important that the user fully understand the scope limitations on each level of the assessments. Please refer to API Recommended Practice 579 for more details. The following is the assessment results of the example problem 5.11.1 described in the API Recommended Practice 579 book analyzed using PVElite 5.0 or CodeCalc 6.5. The metal loss flaw is categorized as localized corrosion with CTP measurements along longitudinal and circumferential directions as outlined in Figure 4.

in. in. in. in. in. in. in. in. psig in.

Thickness Measurement Type Number of Points in Circumferential Dir. NROW Number of Points in Longitudinal Dir. NCOL Circumferential Grid Size GRIDSC Longitudinal Grid Size GRIDSL Skip UG-16(b) Min. thickness calculation Type of Element: API579 ANALYSIS RESULTS, SHELL NUMBER Inside Diameter (D) with XLOSS: Thickness (T) with XLOSS:

Circumferential Minimum Required Thickness (TMINC): = (P*(D/2+FCA))/(S*E-0.6*P) per UG-27 (c)(1) = (300.00*(96.2000/2+0.1250))/(17500.00*1.00-0.6*300.00) = 0.8353 in. Longitudial Minimum Required Thickness (TMINL): = (P*(D/2+FCA))/(2*S*E+0.4*P) + tsl = (300.00*(96.2000/2+0.1250))/(2*17500.00*1.00+0.4*300.00)+0.000 = 0.4119 in. Max. All. Working Pressure at Given Thickness (MAWP): = (S*E*(T-FCA))/((D/2+FCA)+0.6*(T-FCA)) per UG-27 (c)(1) = (17500.00*1.00*(1.0250))/((96.2000/2+0.1250)+0.6*1.0250) = 367.27 psig Min. Metal Temp. w/o impact per Fig. UCS-66 Min. Metal Temp. at Reqd thk. (per UCS 66.1) Minimum Required Thickness (TMIN): = MAX[ TMINC, TMINL, Tca ] = MAX[ 0.835, 0.412, 0.000 ] = 0.835 in. Thickness Profile Analysis Results: Critical Thickness Profile in Longitudinal Dir., CTPL (in.): 1.150 0.810 0.750 0.700 0.620 0.450 0.650 0.900 1.150 Critical Thickness Profile in Circumferential Dir., CTPC (in.): 1.150 0.700 0.450 0.810 1.150 Minimum Measured Thickness Remaining Thickness Ratio ((TMM-FCA)/TMIN) Factor from Table 4.4 Length for Thickness Averaging (XL): = Q * SQRT(D * TMIN) = 0.447 * SQRT( 96.000 * 0.835) = 3.999 in. Using Para. 4.4.2.1.e.2.b Circ.(C) | Long.(S) (TMM at midpoint of XL) in. | in. Flaw Dimension 3.021 | 3.342 TMM = Rt = Q = 0.450 0.389 0.447 in. 38 -72 F F

Figure 4. Flaw CTPs Along Longitudinal and Circumferential Directions

17

COADE Mechanical Engineering News

SECTION 5, Local Metal Loss Analysis Limiting Flaw Size Check: Rt >= 0.20 0.389 >= 0.20

July 2003

TRUE

(TMM - FCA) >= 0.10 ( 0.450 - 0.125) >= 0.10 0.325 >= 0.10 TRUE Lmsd >= 1.8 * SQRT(D * TMIN) 60.0000 >= 1.8 * SQRT( 96.000 * 0.835) 60.000 >= 16.119 TRUE SECTION 5 LEVEL 1 ANALYSIS: Shell Parameter (LAMDA): = 1.285 * S / SQRT(D * TMIN) = 1.285 * 3.342 / SQRT( 96.000 * 0.835) = 0.480 Longitudinal Check: Figure 5.6 check is ACCEPTABLE with: LAMDA = Rt = Folias Factor (Mt): = SQRT(1 + 0.48 * LAMDA) = SQRT(1 + 0.48 * 0.480) = 1.054 Remaining Strength Factor (RSF): = Rt / (1 - 1 / Mt * (1 - Rt)) = 0.389 / (1 - 1 / 1.054 * (1 - 0.389)) = 0.926 >= RSFA ( 0.900 ) Acceptable Circumferential Check: * Figure 5.7 check is ACCEPTABLE with: C/D = 0.031 Rt = 0.389 SUMMARY SECTION 5 LEVEL 1 ANALYSIS: Calculated Allowable Long./Merid. - RSF 0.926 0.900 Passed Circ. - Figure 57 Check Passed ** Section 5 Level 1 Assessment criteria are satisfied ** RLife (Iteration Method) |RSF>RSFA years | : 168.90

PC Hardware/Software for the Engineering User (Part 34)

(by: Richard Ay)

Windows XP Windows Messenger If you dont use Windows Messenger, you probably want to turn off the auto load of this application. This will save some system resources, and simply avoid the nuisance of seeing its icon in your task bar tray. Here is how to turn this off, for good.

0.480 0.389

1. 2. 3.

Click [Start], then [Run] Type in gpedit.msc and press [Enter]. This will start the Group Policy Editor. Double click these items to expand them: local computer policy, computer configurations, administrative templates, Windows components, Windows Messenger. Now double-click Do not allow Windows Messenger to run, then click [Enabled]. Click [OK] and then quit the Group Policy Editor.

4. 5.

For Windows Messenger Version 4.5 or later, please refer to http:// support.microsoft.com/default.aspx?scid=KB;EN-US;q302089&. (Note the semicolons in this link!) Advanced Searching on the Microsoft (and other) Web Sites Sooner or later you will have a problem where you need to search the knowledge base on Microsofts web site. This is such a large web site, navigation can be difficult. However, the link below is designed to bring up a Google search page, that only searches Microsofts support database. The link is: http://www.google.com/advanced_search?q=+site: support.microsoft.com Once this search page is displayed, fill in your search criteria, then click [Google Search]. This same idea can be extended to any web site. For example, to search the COADE web site for any articles or documents on friction, this link can be used: http://www.google.com/advanced_search?q=+site:www.coade.com Simply change the URL of the website, following the +site: qualifier.

|Rt>(TMM-(Crate * Time))/TMIN| SECTION 5 LEVEL 2 ANALYSIS: Longitudinal Check: Using Slicing Method: With * S = * Length Inc.= Acceptable

3.342 0.025

in. | Calculated RSFmin: 0.977 in. | RSFmin > RSFA (0.900)

Circumferential Check: * Figure 5.7 check is ACCEPTABLE with: C/D = 0.031 Rt = 0.389 SUMMARY SECTION 5 LEVEL 2 ANALYSIS WITH NO LOAD: Calculated Allowable Long./Merid. - RSF 0.977 0.900 Passed Circ. - Figure 57 Check Passed ** Section 5 Level 2 Assessment criteria are satisfied **

CodeCalc 6.5, 2003 1989-2003 by COADE Engineering Software

As shown in both the Level 1 and Level 2 summaries of the report, the flaw located on the inside of a vessel does pass Level 1 and Level 2 assessment criteria. In general, the Level 2 assessment is not needed when the flaw has passed the Level 1 assessment. However, for checking purposes, both levels of assessment are performed unless restricted by the scope of limitations.

18

July 2003 Is Spam the Majority of Your E-mail? Many articles have recently been published concerning the e-mail spam problem. Spam is the current hot topic in the press. Some views even go so far as to claim that spam could render e-mail useless in the near future if a solution to this problem is not found. The discussion of one possible solution appeared in the April 2003 issue of Network Magazine, in an article titled Fighting the Spam Monster and Winning. The article discussed the various methods used to fight spam, with particular attention on Bayesian Filtering. Bayesian filtering is an attempt to classify e-mail based on snippets of text from the e-mail, and a mathematical algorithm to determine the probability that the message is good, bad, or unsure. The advantages and disadvantages of Bayesian filtering are summarized in the following table. Advantage Very effective filtering, over 95% of spam caught Generates few false positives Automatically learns Hard for spammers to trick Allows user fine tuning Disadvantage Computationally intensive Not well suited for upstream server installation

COADE Mechanical Engineering News 2) Shut down Outlook for the installation procedure. 3) Install the Python package, then the Win32 extensions, then the SpamBayes package. 4) After installation, run addin.py, then view about.html. 5) If the installation succeeded, you should see three new controls on the Outlook toolbar, as shown in the figure below.

6) Create two new folders in Outlook, the names are irrelevant, but spam and possible-spam are good choices. 7) Gather as many spam e-mails as possible, and move them to your spam folder. 8) Use the Anti-Spam control to specify all of your folders with good e-mails, and your spam folder. 9) Then use the train now option to initialize and train the Bayesian filter. As new e-mails arrive, they are evaluated. If the filter decides the email is good, it is left in your in-box. If the filter decides the e-mail is spam, it is moved to your spam folder. If the filter is unsure about a particular e-mail, it is moved to the possible-spam folder, at which point you can use either the Delete As Spam or Recover from Spam controls. These controls also enhance the training of the filter, so future, similar, e-mails are handled automatically. It is also recommended that you occasionally re-train the filter, since spam is continually changing. Initial use at COADE has shown that 99% of all spam no longer resides in the in-box. After a few days of use, virtually all spam goes to the spam folder, with the remainder going to the possiblespam folder.

Additional details are discussed in the referenced article. So, what is necessary, how is it setup, how does it work. (The following details are provided for information only. While COADE is currently testing the implementation described here, this is not COADE software. Therefore, COADE can not provide assistance or support for this anti-spam tool.) The starting point should be http://spambayes.sourceforge.net, which contains explanations and other necessary links. To summarize, the following steps should be implemented: 1) Download:spambayes-1.0a2.zip which is the Bayesian filtering package (available from https://sourceforge.net/ project/showfiles.php?group_id=61702) win32all-153.exe, which is a set of Windows extensions for Python (available from http://starship.python.net/crew/ mhammond/win32/Downloads.html) python-2.3a2.exe, which is the Python compiler (available from http://www.python.org/download/)

19

COADE Mechanical Engineering News

July 2003 Configuration Module Corrected a version identification problem which prevented the stress color range settings from being read from existing configuration files. Added B31.11 as an option for the default piping code.

CAESAR II Notices
Listed below are those errors & omissions in the CAESAR II program that have been identified since the last newsletter. These corrections are available for download from our WEB site, for Version 4.40. Static Load Case Setup Module / Dynamic Input Corrected the friction multiplier application when static load cases were deleted on the edit dialog. Corrected initialization of load case options when changing piping codes. Corrected the storage of wind topographical data for wind vectors 2 through 4. This also affects wave data.

Buried Pipe Module Modified to address new B31.1 materials Corrected the element data space initialization.

Dynamic Output Module Corrected the input echo of configuration data for:

translational stiffness units labels B31.3_SUS_SIF_FACTOR display OCCASIONAL_LOAD_FACTOR display

Large Job Printing Module Modified to handle correction for spring hanger load variation calculation when cold load design is activated.

Corrected an instance of pathname allocation being too short. Corrected the access of nozzle data for input echos.

Material Database Editor Corrected the identification of the piping code when adding materials to the data base for codes listed after B31.11.

Animation Module Corrected an instance of pathname allocation being too short.

Material Database Corrected allowables for B31.3 A312 TP347H over 1000 degF.

Element Generator Corrected an error in generating element loads when uniform loads are in Gs, and W (weight), WNC (weight no contents), or WW (water weight) are not present in the load case. Changed to include buoyancy effects in load cases based on WW (weight with water).

Low Level Graphics DLL Corrected a problem on Win98 with critical code section that caused module linked to this DLL to crash on exit.

PipePlus Interface Corrected acquisition of bend data in Pipeplus interface. Corrected units translation for densities in Pipeplus interface. Corrected testing for material and allowable specifications. Corrected restraint processing for multiple restraints at the same node point.

Static Solver Corrected the stiffness used for designed constant effort supports when changing hanger status to a setting other than as designed. Corrected the friction loads when the new friction mulitplier was set to zero in the load case setup details. Tweaked the friction algorithm for stiffness reset when convergence problems occur. Corrected to properly lock predefined spring hangers for the hydrotest load case.

Structural Modeler Corrected so that the interface wont eat trailing zeros on exponential notation, i.e. 1e10.

20

July 2003 Intergraph Interface Updated the splitting of bend elements to address both tees and dummy legs.

COADE Mechanical Engineering News Corrected re-initialization of graphics data space when switching jobs from within the output processor. Corrected the determination of the data directory path when switching jobs from within the output processor. Corrected the formatting of node numbers in the restraint summary report for nodes with multiple restraints. Corrected the input echo of configuration data for:

Miscellaneous Computation Processor Modified to address new B31.1 materials Corrected imposed limits on the crotch radius for B31.8 extruded welding tees. Corrected the operation of the spinner control on the bend SIF dialog so that the number of miter cuts is properly obtained. Corrected the SIFib division by 2 for use in WRC329 EQ 46 when SIFob was previously divided by 2. Corrected message handler from eating the 3 key on the numeric keypad.

translational stiffness units labels B31.3_SUS_SIF_FACTOR display OCCASIONAL_LOAD_FACTOR display

Initialized printer device context flags before getting the device defaults. Corrected the graphical display of displacement values when Z axis vertical is activated. Corrected the access of nozzle data for input echos. Corrected the printing of stress titles for TD/12 code when the yield stress criterion is set to Von Mises. Corrected the computation of spring hanger load variation when cold load design is activated. Corrected shutdown of program using File\Exit when printing is in progress.

Offshore DLL Corrected a dimensionless parameter used in the Stream Function wave theory, which was actually units dependent.

Nozzle Input Echo Format File Corrected template for nozzle input echo.

ODBC Export DLL Corrected the output of the Lisega spring size when sending data through the Data Export Wizard (ODBC). Modified to handle correction for spring hanger load variation calculation when cold load design is activated.

Piping Error Check Module Corrected the over-ride of the thickness used in the B31.1 effective section modulus calculation for SUS and OCC load cases when the B31.1 Reduced Z Fix configuration directive was activated. This change only affects those tees where the branch connection equation is used. Modified to address new B31.1 materials Corrected imposed limits on the crotch radius for B31.8 extruded welding tees. For B31.3 Welding Tees and Sweepolets, changed the test for Note 11 to correct a code error. Corrected table pointers for B31.1/B31.3 y parameter used in minimum wall thickness calculations. Corrected the SIFib division by 2 for use in WRC329 EQ 46 when SIFob was previously divided by 2. Corrected handling/storage of material data for included job files.

Static Stress Computation Module Corrected a material input/output procedure to address combined piping files. Corrected the usage of Sh values for B31.8 Ch VIII, affecting multipe OPE cases. Modified the bending term in 3D Max Stress Intensity calculation for hydrotest load cases.

Static Output Module Modified all output filters to use a logical and instead of a logical or when a node number range is specified. Added the conversion of the hydrotest pressure field to the routine for input echo display.

21

COADE Mechanical Engineering News Interfacing DLL Corrected handling of temporary material file. 2) Output Module:

July 2003

Corrected the output of the annular base plate weight. Corrected the use of the units conversion constant for threads per unit length. Corrected the use of the units conversion constant for nozzle expansion coefficient. Corrected the output of two configuration directives for the input echo.

Piping Input Module Corrected the activation of the Eff field for new jobs when the code is switched to B31.8. Modified so the seam weld setting is assumed for new bends when using the TD/12 code. Corrected the operation of the Element LIST dialog to allow proper editing of the fields following the hydrotest pressure field. Corrected the input echo of configuration data for:

3) Units Generation Utility: Corrected the conversion factor for rotational stiffness for N-m/deg.

translational stiffness units labels B31.3_SUS_SIF_FACTOR display OCCASIONAL_LOAD_FACTOR display

4) Error Check Module: Corrected the check of seismic data to allow -1 as valid input for the seismic zone.

Modified to address new B31.1 materials Corrected the usage of the block rotate setup options. Corrected SaveAs function to handle .SOI and .XML files. Corrected the display of the count of node names in the model status auxiliary display. Corrected the UNDO operation when invoked from the LIST. Corrected the access of nozzle data for input echoes. Listed below are those errors & omissions in the CODECALC program that have been identified since the last newsletter. 1) TEMA Tubesheet module: Properly corroding the outer cylinder of the Expansion Jt. Added warning for the tube pitch. Corrections in the Tube-Tubesheet full-strength weld calcs. Correctly interpreting the flange load transferred to the tubesheet that are entended but the bolt load is not transferred to them. For floating tubesheet, added input for G of the stationary tubesheet.

CODECALC Notices

WRC107 Module Corrected the use of the Z-up flag. Corrected to initialize all graphics variables between different load cases

TANK Notices
Listed below are those errors & omissions in the TANK program that have been identified since the last newsletter. These corrections are available for download from our WEB site, for Version 2.40. 1) Solution Module: Corrected the reset of minimum shell thickness (6mm) when working in metric units. Corrected the use of the FULL_SHELL_ WEIGHT_IN_APP_F directive in the computations.

2) Program Interface: Addressed issues relating to switching between input and graphics views. Fixed some dialogs that were closing on hitting Enter key, instead of tabbing to the next input field. Modified the reading of title page data for older files. Added some missing materials in the material database.

22

July 2003 3) Shell: Corrections to API-579, calculation of c value. Fixed MAWP on the status bar for jobs with static head specified.

COADE Mechanical Engineering News 11) Some 3D graphics features were fixed, such as nozzle on nozzle plotting. 12) For leg baseplates when there were 0 bolts in tension the program could abort. 13) When user defined wind pressure was specified and there was a top head platform, the wind load on the platform may not have been calculated in some cases. 14) When user specified axial loads were entered at cone/cylinder junction, they may not have been consider in the Q calculation.

4) Nozzle: Fixed some issues relating to results on the status bar.

5) Flange: Corrections to the blind flange calculations. Added check for Lap Joint.

PVElite Notices
Listed below are those errors & omissions in the PVElite program that have been identified since the last newsletter. 1) Algebraic force/moment summation for base skirt supported vessels was corrected. 2) On screen nozzle calcs for external pressure were not considering the shell CA. This has been corrected. 3) Changed Nozzle diameter limit and added pad area for 5500 closely spaced nozzle check. 4) The F factor was being used on offset hillside nozzles in cylinders unintentionally for the external tr case. 5) Fixed on screen calc of the pad diameter when the pad width was entered on actual thickness basis for larger nozzles. 6) Fixed the on screen weld calc for required thickness of the inside weld. 7) Sorted out a memory issue with the output processor and color hightlighting. 8) Implemented new computation for partial volumes of nonstandard F&D Heads. 9) Sorted out a sign issue for cone/knuckle/ring/shell inertia calc. 10) Fixed the use of local shell thickness for the on screen nozzle calcs.

COADE Releases CADWorx 2004 Simultaneously with AUTOCAD 2004, Co-Promotes Products with AUTODESK
(by: John Brinlee)

On March 17, 2003, COADE released CADWorx Version 2004, the latest version of its AutoCAD-based Plant Design Suite. The release culminated a development period during which COADE worked closely with Autodesk as one of only four software developers worldwide selected to release an AutoCAD 2004-based product on the same day, coincident with the debut of Autodesks new release, AutoCAD 2004 (note that CADWorx is also compatible with AutoCAD versions 2000, 2000i, and 2002). CADWorx 2004 not only offers process industry designers the ability to immediately take advantage of the groundbreaking enhancements in AutoCAD 2004, but also provides many new features compared to COADEs previous version, CADWorx 2002. By leveraging Autodesks investment in the next release of AutoCAD, COADEs new version of CADWorx greatly increases the size of plant models that teams of designers can work on simultaneously, while making it much easier to learn and use the program. These dramatic improvements in both performance and functionality are made possible by significant file size reduction, faster load and save times, enhanced management of external references (XREFs), and user interface refinements such as the new tool palette system. In addition to offering compatibility with AutoCAD 2004, CADWorx 2004 offers many other new features. The CADWorx PIPE 2004 module provides improved integrated steel capabilities, integrated HVAC/cable tray components, automatic weld gaps, and layering by line number. New capabilities in CADWorx P&ID 2004 include enhanced copy procedures, an auto repeat feature, and a dropdown list for instant data entry.

23

COADE Mechanical Engineering News

July 2003

Thomas Van Laan, president of COADE, believes that the improvements offered in AutoCAD 2004 are exactly those for which the plant design industry has been hoping. Says Van Laan, Our customers are always concerned with three things - speed, size, and how to manage them - so we think our customers will love this new version of AutoCAD. Weve found that a 22-megabyte project created under CADWorx 2002 drops to less than 6 megabytes under CADWorx 2004, a dramatic 70%+ reduction in project file size. Van Laan continued, CADWorx has always taken maximum advantage of AutoCADs XREF capabilities to the hilt. The new XREF management features, including improved load speed and change notification, are perfect complements to the way that our customers manage large projects. John Sanders, Vice President Platform Technology Division for Autodesk, agrees that COADE has done a great job leveraging the best features of AutoCAD 2004. Sanders says, We are very pleased that COADE was able to develop a 2004-compatible version of CADWorx so quickly. Autodesk has been working very closely with COADE to determine what AutoCAD enhancements that would be most valued by the plant design community. Were impressed by how they have leveraged the strengths of AutoCAD 2004 - speed, teamwork and management - and translated these strengths into productivity improvements for process plant designers. CADWorx 2004 is a great tool for anyone involved in the design of process plants. Following the release, COADE and Autodesk collaborated on a multi-city tour to conclusively demonstrate the advantages of CADWorx 2004 operating in an AutoCAD 2004 environment. Presentations showing how CADWorx 2004 can offer a more economic plant design solution over a full range of project sizes were made to receptive audiences in Houston, New Orleans, Philadelphia, New York, Boston, Chicago, Calgary, Singapore, Antwerp, and Moscow with more of the same touted for Birmingham, Atlanta, Seattle and Kuala Lumpur in mid to late July.

CADWorx 2004 Takes Advantage of AutoCAD 2004's Finest Features

Gates Barman of Hanover Corporation Receives the First Copy of CADWorx 2004

COADE Engineering Software

12777 Jones Rd. Suite 480 Houston, Texas 77070 24 Tel: 281-890-4566 Fax: 281-890-3301 Web: www.coade.com E-Mail: techsupport@coade.com