PLANMAN developed a generic software
system which collects, analyses and stores process data from the manufacturing
shop floor and makes it available to planners to improve the accuracy
of their time and cost estimates.
Targeting small and medium-sized enterprises, the project involved
developers and end-users in three countries. The system is well adapted
to the technical and financial constraints experienced by 'real world'
manufacturers - it has been designed to interface with existing process
planning software, and is effective in 'manual' mode when on-line
data is not available from current machine controls.
It was launched at the September 1996 CIM Show in the UK, and a prototype
has already proved itself in use in a variety of engineering contexts.
The profitability of any manufacturing
company depends crucially on the accuracy with which it is able
to predict production costs and times. If process costs or times
are over-estimated, the company will fail to win new business. If
they are under-estimated, it may take on contracts that result in
losses, or let down its customers.
Yet in many small and medium-sized engineering businesses, production
planning is a theoretical exercise which fails to take into account
the actual conditions on the shop floor. A particular combination
of operator, machine and materials may produce performance which
is better or worse than the 'ideal' which forms the basis of planners'
calculations. In either case, the company will lose out.
PLANMAN set out to develop software tools which would support progressive
improvement in the accuracy of process planning by giving planners
access to accurate data directly from the shop floor.
Informed throughout by feedback from industrial end-users, the project
has produced a flexible, generic solution which neatly adapts the
'self-learning' concept to the realities of the industrial world.
The system is suitable for use in applications as diverse as steel
milling and medical scanning, and is in each case able to improve
the tracking, analysis and prediction of critical processes.
Helping planners learn from experience
The PLANMAN project was suggested by the Institute for Production
Engineering and Machine Tools (IFW) at the University of Hanover.
Process planning systems were available, but none could incorporate
feedback from what actually occurred within an operation on a particular
Instead, planners were using values based on specifications supplied
by the machines' manufacturers. IFW found, however, that actual
performance deviated from estimates in at least 20% of cases. Process
parameters were changed on the shop floor, but planners often remained
unaware of these changes, so the accuracy of their predictions never
The project's partners wanted to close the gap between shop floor
operation and the planning department's estimates by linking them
both to a comprehensive Technological Information System (TIS).
Feedback from the shop floor would allow the nominal values used
for costing and work planning to be adjusted in the light of actual
experience. As the bank of data grew, the quality of the estimates
would get steadily better.
The partners foresaw substantial benefits. Not only would the accuracy
of process planning be improved, but PLANMAN made keener pricing
and better control of margins possible. Accurate estimates would
also provide a more reliable yardstick against which to monitor
manufacturing operations, allowing problems to be diagnosed and
corrected at an earlier stage. By providing an interface between
them, PLANMAN would lead to enhanced performance both in the planning
department and on the shop floor.
IFW's initial mapping of currently available theoretical models
and approaches led to the specification of a system with five distinct
The partners had recognised from the outset that the monitoring
of all manufacturing operations would generally not be practicable.
The Process Data Definition module therefore focuses on operations
which are most critical to overall productivity. Planners are asked
to define these in terms of parameters (for example, cutting speed
and depth) and resources, such as tool and workpiece materials).
The Process Tracking module handles the collection of actual data
for each specified parameter. Data can be keyed in by the operator
or, if the machine controller offers a suitable interface, collected
automatically from the machine itself. In each case, 'pop-up' screens
on a standard PC prompt the operator for manual input when it is
needed, for example, to explain an unscheduled stoppage.
The Process Analysis module enables the planner to interrogate process
data stored in the system's database to detect deviations from predicted
performance. Inquiries can be made in relation to particular machines
or tools, particular materials, or specific orders or workpieces.
Results may be visualised in chart form or summarised in printed
reports, which are configurable by the user.
The Process Prediction module uses data selected from the database
as the basis for predicting the outcomes of common operations under
different process conditions. Planners can calculate the costs of
a specific operation using past results, or predict the effects
of adjusting process parameters.
Lastly, the Information Manager provides tools for the integration
of PLANMAN with a company's existing process planning, tool management
or quality management systems.
The project involved six partners, with a technical developer and
an industrial end-user in each of three countries. The national
pairings proved extremely helpful, according to British software
developer CIMTEL, which led the project. Each partner had their
specialist role, but the close relationships between the two partners
in each country created a framework within which a high level of
focus and commitment was achieved.
While IFW developed the underlying data model, TEKNIKER was responsible
for researching the end-user requirements on which specification
of the system's functionality was based. Nevertheless, full end-user
participation in the testing phase was crucial to the system's thorough
adaptation to real-world manufacturing environments.
The original design concept was adjusted in a number of key respects
in order to satisfy the practical needs of the project's industrial
partners. As a result, the system as a whole is more flexible, and
will be attractive to a wider range of customers.
For example, the automatic collection of process data from machine
controls had originally been specified as a required feature. In
practice, however, many manufacturers are unwilling to invest in
more sophisticated controllers capable of outputting data directly
to a computer. The partners realised that the system must also be
able to operate in 'experimental' mode.
ADWEST Engineering, a British firm, uses the system in this way,
entering data manually in order to test alternative set-ups before
production is started or modified. As a result, for one product
they have achieved a better balance of throughput across a group
of machines, increasing overall productivity by 30%.
A further extension of the project's original technical aims arose
from the request of the Spanish machine-tool manufacturer ZAYER
for shop floor analysis and control functions. These enable machine
operators themselves to examine on-line process data in order to
investigate deviations from the workplan.
Flexibility gives lowest-cost solution
CIMTEL's approach to the design of database functions was also
driven by industrial cost-consciousness. PLANMAN gives users the
option of fully relational data storage and retrieval. However,
that is likely to require new hardware and software which may be
beyond the reach of some potential users, so a simpler file management
solution was built into the system, and offers 98% of the functionality
for no extra cost.
The completed package was launched at the 1996 CIM Show held in
the United Kingdom, and will form an important addition to CIMTEL's
established product range. ZAYER will also gain a significant competitive
edge from their ability, gained through the BRITE-EURAM project,
to offer purchasers of their machine-tools the option of controllers
which can feed results directly to an analytical system.