Albert Ackermann GmbH & Co. KG

Background

The production of electrical installation systems is organised into 12 different manufacturing cells, of which 3 are to be examined in-depth for the project. One manufacturing cell mainly produces wall trunking, while another mainly produces underfloor trunking. The third manufacturing cell is responsible for painting by means of powder coating and therefore represents a potential bottleneck. The range of products is usually produced to customer-specific requirements on an order basis, whereby the trend is towards an increasing number of smaller orders and this results in greater complexity of the production organisation. The products are produced product-specifically; therefore there is not fixed production sequence via machines or workstations. The manufacturing cell employees are mostly qualified for several activities within the cell.

Coordination problems between the cell managers, especially in fault management, were the reason for selecting these three manufacturing cells. The brief was to remove employee load peaks, remove, as far as possible, scheduling difficulties and difficulties in meeting deadlines and utilise the available resources as optimally as possible.

Task

The brief was to provide an aid for creating the optimum production order for the orders on hand for a week against the background of improving schedule effectiveness and meeting the deadlines for the individual orders. Further, it should be possible to reliably estimate the consequences of faults or changes in production plans in advance.

Implementation

The production was theoretically reproduced with the help of the simulation; it was possible to evaluate the order of the jobs in production with respect to costs and deadlines, and automatic optimisation of the order of the jobs was made possible. The model was then validated, a human-machine interface was developed for operation of the simulator and a trial run was performed. The workstations are displayed in the simulation model, while the order data and order products determine the respective sequence and the times required. The job data comes from Ackermann's PPS system and contains information on the quantity, deadline, products, parts list status, priority and the work plan. The work plan of a product determines the machining stations, including setting up and machining times, for a job. In addition, the model contains data on the employees' qualifications and their working hours. The model is therefore capable of requesting a qualified, present employee for the work due at a specific machine.

acker1
Figure 1: The Ackermann simulation model

The manufacturing cell organisation is reproduced in the model accordingly (Fig. 1). The hierarchical structure of the model enables a cell to be opened to achieve a greater detailing (Fig. 2). Both figures show a situation during a simulation. Production jobs are shown as simple trunking symbols. The employees are also shown in the model.

acker2
Figure 2: Manufacturing cell 01 - workstations and general cell function

Individual workstations can in turn be opened to enable specific properties to be realised. The workstation-relevant part of the employee management is also located there. Two workstations are shown as an example. The first contains "only" one standard function (Fig. 3), which is used to take into account the specific data from the work plan. The second example shows the powder coating (Fig. 4), in which a special logic for collating jobs ensures the same colouring.

acker3
Figure 3: Standard workstation

A standard workstation ensures correct working through a work sequence of the order based work plan. The setting up and machining time are taken into account. Jobs which arrive at a workstation wait in the buffer until the previous job has been completed. Depending on the setting, the next job is then selected from the buffer for machining. Possible selection strategies are FIFO (First-In-First-Out), the most urgent deadline or specific criteria. The latter can be programmed in the "criterion" method. Each job passing through the workstation carries a log in which each workstation enters the times required for the work or machining operation carried out there. A workstation also logs the work carried out by it and the jobs completed in their workstation plan (APLZ plan). In addition, there is a sequence function with which the order in which the work is to be carried out at the workstation can be explicitly defined.

acker4
Figure 4: Specialised workstation - powder coating

Ackermann provided production data on several occasions for the model validation. This means, after a selected week, all the start reports and feedback from the period under consideration was selected and made available to tms together with the corresponding order, job and work plan data. The data supplied was fed into the simulation model. As a result, the real data was compared with its simulated counterpart. This examination was carried out for several, very different production weeks. Initial checks led to subsequent to the model until the correspondence between the model and reality was found to be sufficient. A user interface was implemented for a trial run on site, which should also enable the cell managers on site to use the simulation and optimisation. To this end, an interface should provide the following options:

  • Selection of the week to be examined
  • Reading in current PPS data
  • Changing the PPS data:

Changing target quantity, deadline, parts list status, work plan, job order, addition of orders

  • Managing employee attendance times for a week
  • Managing employee qualifications
  • Managing faults and workstation maintenance
  • Managing the order in which the jobs are to be completed at the individual workstations
  • Start simulation or optimisation
  • Evaluation of the simulation results: Statistics, details, comparison of two results

From this, a dialog was set up which divides the functions into three areas: Input, processing and output (Fig. 5).

acker5
Figure 5: SIGNAL main dialog for Ackermann

The processing area is kept especially simple, as the simulation is solely controlled beforehand through the data input. This was prerequisite for a spontaneous simulation, which often has to be performed with an automatically varied job order for optimisation purposes. There is only one small progress indicator. Following a simulation, the evaluation is displayed immediately. It also contains a small statistics feature for evaluating the production schedule and the other input factors. At the end of the optimisation, the best solution found is simulated and its result is displayed.

Results

The manufacturing cell managers can quickly and conveniently perform simulations, the consequences of their decisions can be evaluated in quantified terms at an early date. The exchange of information between the cells was initially handled by the company's usual means of communication, the necessary recording of data not yet available in the PPS system had to be entered manually, which reduced the up-to-dateness of the data. In a further development stage, an integrated system solution is to be established, in which a planned production sequence is to be switched to a cell as currently valid and is to be synchronised with the other cell schedules. Any problems which occur should then be automatically displayed to the responsible cell managers.

Conclusion

The simulation was able to show that optimisation potential still exists in the production and can be fully utilised using the information from the model.