|
 |
|
A manufacturer of MDF (Medium Density Fiberboard) hot-presses resin and wood fiber into 20, 6 ft by 18 ft
panels that range from .25-inch to 1.250-inch in thickness.
The hot-pressed panel passes between 5 lasers and receptacles, which give a measure of cross-panel thickness for
up to 30 slices down the length of each panel. The point data is stored (about 150 per panel) in Excel workbooks
and is used to draw 3-d profiles of the panel. The 3-d profiles determine the mechanical condition of the
hot-press.
The problem was the presence of corrupt data that distorted the thickness variation statitstics as shown in the
3-d profiles. This inhibited the ability of the company to determine if the press mechanism had improved after a
maintenance program. The company sought a programming solution for the data anomalies using an Excel VBA
application.
|
Approximately 1 percent of the data points gathered at the laser gauges were questionable data. A good number
of these were, intuitively, obvious anomalies. The questionable data, however, could be statistically compared
to data gathered near the same position on the same panel and to data at and near the same position from
previous hot-pressed panel.
For example, data point 1 at laser position 1 for the first reading could measure .752 on four press loads and
measure .790 on the fifth. The data point at the fifth load could be blown or thick panel for unknown reason.
The data point could also, however, be an anomaly. Now, there are between 2 and 8 data points next to the data
point in question on the same panel. There is also a significant amount of historical data from the same press
opening at and near the same position.
The programming innovation was to load each panel into two-dimensional arrays, and use SPC methods to compare
the questionable data with proximity and historical data. The questionable data point was replaced, if necessary,
with an approximate value based on the average of the slopes between it and nearby data points and the slopes
of historical data points at the same position.
TOP ^
|
A global corporation was converting their corporate database system from Oracle to SAP. The data and processes had
to be mapped between the two disparate software systems. The corporate management team hired an Oracle consultant
to be a project leader, and he sought a programming solution to convert the Oracle data according to his rules and
formulas.
|
The original unaltered data was transferred from the Oracle database to Excel workbooks. Excel VBA software
applications were developed to change the data as outlined by the project leader. Lastly, the company employees
transferred the reprocessed data from the Excel workbooks into the SAP database.
TOP ^
|
The Technical Director of a manufacturer of Medium Density Fiberboard (MDF) wanted improvements to the
Density Gradient Analyzer (DGA) nuclear gauge program. The DGA nuclear gauge gathers density data from a
2” X 2” X “Y” sample of MDF and inserts the data into a text file. The text file is loaded into a spreadsheet
by an Excel application. The application draws a chart of the density profile of the sample at thickness “Y” and
displays important process control data in a text box on the chart.
When new and more innovative statistical process control techniques became available, the Technical Director of
the company saw the need for additional data to be displayed by the Excel application. Moreover, some code in the
old program caused a runtime error on samples less than .1875” in thickness.
|
Peak values of density, distance between peak values of density, average values of face density, average sanded
density, and depth of precure are displayed in text sections along with the other data in the DGA chart
drawn by the macro as requested by the Technical Director. The code causing the runtime error on samples less than
.1875” in thickness is repaired. Since the original macro was written in Excel 5.0 and is not to be changed
anytime soon, care was taken not to introduce code that would only run on a newer version of Excel.
The Excel application was built to the Technical Director’s satisfaction.
TOP ^
|
|
|
The Lab Supervisor of a manufacturer of Medium Density Fiberboard (MDF) wanted an Access database program
to store hold wood data. Hold wood data comes from units of MDF panel that the Quality Control (QC) Department
flags and tests because a random sample from the batch had a property or properties that were not up to process
control specifications.
The Access program needed a user-friendly form for data entry, reports to display graphical data, and reports to
display numerical data and averages. Since it is unfeasable to test a panel from each unit of a large purchase
order, the Lab Supervisor requested a “Grade the wood” button on the form with programming behind it to grade
individual units that were held but not tested. The program would insert the new records into a table.
To see how the “Grade the wood” programming works, consider the following simple example: Suppose that from a
particular production order, samples tested from unit 2 were utility grade and unit 6 were industrial grade.
Units 1-5 are put on hold. Now, the QC technician tests samples from units 1, 2, and 5. Suppose the samples
tested from unit 1 are industrial grade and units 2 and 5 are utility grade. Instead of testing units 3 and 4,
they are considered to be utility grade because both units 2 and 5 are utility grade. The new records are
created by pushing the “Grade the wood” button. The programming behind the button, however, had to be intelligent
enough to identify if another unit should be tested.
|
The Access program was developed to perform the tasks requested by the Lab Supervisor. Tables, queries, reports,
and forms were designed to accommodate his requirements for data storage, management, retrieval, display, and
entry. The program was thoroughly tested and documented.
TOP ^
|
The Document Office Manager of a production plant of a global corporation requested an Access database application to
replace a VB program that indexed hundreds of plant operating procedures. The new program needs a friendly user
interface that helps any employee to search for and find the location of an operating procedure by keyword, title,
subject, department, author, and/or date. For security purposes, only the Document Office Manager will be permitted
to update the database. Additionally, the application and associated documents must meet stringent FDA approval.
|
A search engine behind a friendly graphical interface was designed to retrieve the relevant documents, and reports
were built to display the results. The application was carefully designed to accommodate all users, and they were
trained to use the software. Updates to the database are restricted to the Document Office Manager by
password-protected access. The final tested, validated application, operations manual, and software validation
documentation were inspected and approved by the FDA.
TOP ^
|
A software project manager of a production plant of a global corporation asked for an Access application that would
automatically input production data into an SAP database system. The manual entry of data into the SAP system was
taking too long. |
In collaboration with the project manager, Access programs were developed to gather production data from
four different departments. VBA functions and API calls were used to automatically input hundreds
of records of production data into the SAP database system. The company was saved thousands of man-hours.
TOP ^
|
An engineer overseeing process control of a production plant of a global corporation needed two Access programs to
display SPC charts of the production data input of two slitter machines. The charts were necessary to ensure that
the quality of their product was not exceeding critical limits. |
An Access database program was developed for each machine, and data input by the machine operators is restricted
by password logon. Past and current records of data are viewed on a read-only basis for all employees in the
production department. Both programs run 24/7.
TOP ^
|
