Browse Prior Art Database

Self-Learning Manufacturing Workload Planning Tool Based on Production Data

IP.com Disclosure Number: IPCOM000184064D
Original Publication Date: 2009-Jun-09
Included in the Prior Art Database: 2009-Jun-09
Document File: 3 page(s) / 316K

Publishing Venue

IBM

Abstract

Presented in this paper is a unique approach to calculating unit hours and efficiencies for workload planning purposes. The method described entails using floor control system captured data during work in process resulting in a non-static unit hour that may be used to plan for different periods within the product lifecycle.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 3

Ȉ ˇ

Existing methods of workload planning are usually inaccurate, too general, and the data used to come up with the plans is usually a snapshot in time and therefore static. Proposed in this paper is a method and process to come up with a more accurate, specific and dynamic way of workload planning. The elements that are required for workload planning include process cycle times, shift hours and other attributes of the work to be done. The core idea is to take historical time stamp data and use statistical analysis to calculate real unit hours and then efficiencies in a "configure to order" manufacturing environment. The unit hours that are calculated are for the activities or process steps that it takes to manufacture a product based on individual workers actions and configurations of the customer order. These unit hours are then fed into production planning tools to accurately plan staffing levels. Key advantages to this approach are:
- Ability to automatically update cycle times based on historical data
- Self-learning and automatic updates to the data modeling
- Ability to identify efficiency variances between processes, workers, product lines, sites, and this can ultimately help with targeted training
- Ability to calculate and apply worker learning curve to the workload calculation
- Quickly determine effects of a new product introduction or process change
- Actual efficiency embedded in unit hour vs. using a % based approach
- Automatically adjusts as efficiency changes over time
- Eliminates need to manually estimate efficiency factors

Process Description:

- Efficiencies are usually calculated by:
- Determining how many hours of work are available per day (# of machines/people x hours in a shift/period).
- Determining the number of unavailable (meals, breaks etc…) hours is then calculated.
- Unavailable hours are then subtracted from available hours to yield remaining available hours.
- Determining the number of hours that are not directly related to production such as signing on to systems, putting on protective clothing, and shift meetings etc…

    Hours not directly related to production are then subtracted from the remaining available hours yielding the truly available hours for the day that can be multiplied by 5 for a normal working week. The projected work hours for a particular activity can be divided by the truly available hours to give an efficiency .

    The efficiency calculated by the above method is then sometimes followed by some time studies to validate the projected time estimates and this is then typically adopted as the manufacturing efficiency. Under the proposed process the unit hours for a given operation are not based on time studies but utilize a series of queries that interrogate the floor control systems based off a set of pre-defined parameters to pull the time stamps off var...