MANUFACTURING PROCESS PLANNING AND DESIGN:TOOLS FOR PROCESS PLANNING

TOOLS FOR PROCESS PLANNING

Process-planning steps were introduced in the previous section. This section discusses the tools used to carry out these steps. Manual process planning, which relies on human experience, will not be discussed. The tools discussed in this section are those used in computer-aided process-planning systems. They are used to assist the human planner in developing process plans. Most of these tools have been used in practical process-planning systems. Methodologies or algorithms used only in advanced research will not be introduced here.

Group Technology

Group technology is a methodology of using the similarity among parts to simplify the production. The key is to identify the similarity. In a machine shop, the thousands of different parts produced may share a small number of geometric characteristics. Parts of similar shape may be produced on the same set of machine tools. The material-handling requirements may also be the same. Manufac- turing cells can be designed to produce similar parts. This simpliﬁes the material handling, reduces the complexity of scheduling, and increases the slope of the learning curve. A complex problem can be decomposed into smaller and simpler problems. This set of similar parts is called a part family. Identifying part families is an important step in applying group technology to different manufacturing problems. Group technology has been used in manufacturing system design, scheduling, product retrieval, ﬁxture design, and process planning.

How GT Is Used in Process Planning

As mentioned in the introduction, similar parts can be produced on the same set of machine tools. Most likely they will follow the same process sequence. Therefore, if one collects all the existing process plans used in the shop and groups them based on the part family, one will ﬁnd that the process plans for the family members are similar. Summarizing the process plans for the family allows a standard process plan to be deﬁned. All parts in the family may share this standard process plan. When a new part is to be planned, one can ﬁnd the part family of this part, based on the geometric characteristics. The standard process plan for the family is then modiﬁed for this new part. Using group technology for manufacturing is like using a library to ﬁnd references for writing a paper. Without the library database, locating appropriate references will be much more difﬁcult and thus so will the writing.

Coding and Classiﬁcation

Group technology is based on the concept of similarity among parts. Classiﬁcation or taxonomy is used for this purpose. The dictionary deﬁnition of taxonomy is ‘‘orderly classiﬁcation of plants and animals according to their presumed natural relationships.’’ Here, taxonomy is used to classify parts in manufacturing. There are many methods of part classiﬁcation. To name just a few: visual obser- vation, manual sorting of the parts, sorting photographs of the parts, and sorting engineering drawings. Because keeping the physical parts or drawing in the sorted order is tedious or sometimes impossible, it is necessary to create a convenient representation, called a coding system.

A coding system uses a few digits or alphanumeric codes to represent a group (family) of similar parts. The classiﬁcation system is embedded into the coding system. For example, one can easily see that parts can be classiﬁed as rotational and nonrotational. A crude coding system can have ‘‘0’’

representing any rotational parts and ‘‘1’’ representing any nonrotational parts. Rotational parts can further be classiﬁed as rods, cylinders, and disks, based on the length-to-diameter ratio. The code can be reﬁned to have ‘‘0’’ represent rod, ‘‘1’’ cylinder, ‘‘2’’ disk, and ‘‘3’’ nonrotational parts. Further, the external and internal shapes of the part can be classiﬁed as smooth, step, screw thread, and so on. Additional digits may be used to represent the external and internal shapes. Each digit reﬁnes the classiﬁcation or adds additional characteristics.

There are many public domain and proprietary coding systems. Opitz (Opitz 1970), Dclass (Allen 1994), MICLASS (OIR 1983), KK3 (Chang et al. 1998) are but a few popular ones. Opitz code (Figure 12), developed by Professor Opitz of Aachen University in the 1960s, uses ﬁve digits to represent the geometry of the part and four supplemental digits to represent part dimension, material, raw material shape, and accuracy. If only the geometry is of concern, the supplemental digits need not be coded. Extensive code tables and illustrations are given to guide the user in coding parts. Given a ﬁve-digit code, one can have a rough idea of the shape of the part. The code can also be used for searching the part database to ﬁnd similar parts. Other coding systems may be more detailed or cover a different part domain, but they all serve the same purpose.

Family Formation

To take advantage of the similarity among parts, one must group parts into families. If the geometry is used in deﬁning the part family, the coding and classiﬁcation system discussed in the previous subsection can be used. Such families are called design families because they are design geometry based. However, often one would like to form part families based on the production methods used. In this case, the part family is called a production family. Because the manufacturing methods are geometry related, members of a production family share many similar feature geometries as well.

Families may be formed using the visual observation method, as mentioned above, or using the sorting approach. In forming design families the coding system can be used as well. Parts with the same codes always belong to the same family. To enlarge the family, several codes may be included in the same family. The determination as which codes should be included is based purely on the applications considered.

To form a production family, one has to consider the manufacturing method. The ﬁrst well-known production family formation method was called production ﬂow analysis (Burbidge 1975). First, production ﬂows, or process sequences, for all parts are collected. An incidence matrix with columns representing each part and rows representing each process, machine, or operation code (a certain process performed on a machine) is prepared based on the production ﬂows. By sorting the rows and columns of the matrix, one may move the entries along the diagonal of the matrix (see Figure 13). In the matrix, one may conclude that parts 8, 5, 7, 2 belong to one family and 4, 1, 3, 6, 9, 10 belong to another family. Family one needs processes P1, P5, P6, and P3. Family two needs processes P3, P2, and P4. This approach can be tedious and requires human judgment on separating families. As can be seen, P3 is needed for both families. It is not uncommon to have overlapping entries.

Over the years, many mathematics-based sorting methods, called clustering analysis, have been developed. For more details see Kusiak (1990) and Chang et al. (1998).

Composite Component Concept

Given a family of parts, one can identify several features. When one takes all the features and merges them into an imaginary part, this imaginary part, called a composite component, is a superset con- taining the features of the family members. The composite component concept was developed before World War II in Russia, where it was used to design ﬂexible ﬁxtures for a part family. By adjustment of the ﬁxture, it can accommodate all the parts belonging to a family. This concept can also be used in design. For example, parametric design uses a parameterized design model for a narrowly deﬁned

family. One is the model for spur gears. Assigning parameters, such as pitch, number of teeth, outer diameter, pressure angle, teeth face width, hub diameter and length, hole diameter, and keyway and set screw, allows a drawing of the gear to be generated.

The same concept can be applied to process planning. When a process model (processes, tools, and cutting parameters) is developed for each of the features belonging to the composite component, a process plan can be generated using the parameters speciﬁed by the user. This concept has been used in developing several experimental and commercial process planners, such as CPPP (Dunn and Mann 1978). Because the same set of parameter data can be used to generate a process plan or a design drawing, CAD / CAM integration can be done. The limitation of this approach is that the family members must have very similar geometry. Not only do features have to be shared, but the relative position of features on family members have to be maintained. Otherwise, not only can the drawing not be done correctly, but the process plan generated will not be usable. The com- posite component concept is a tool for part families with minor differences among family members.

Process Mapping

Why a process can generate certain shapes depends on the geometry generation process of the tool. For example, a drill bit has two cutting edges (lips) (Figure 14). The drilling process requires the drill bit to rotate along its axis, then move the cutting edges downward. The rotating cutting edge creates a cone. Sweeping down the cone will remove a cylindrical volume of materials. Thus, the holes created always have a cone-shaped bottom. The turn tool has a single cutting edge. A layer of the material on a rotating workpiece is shaved off by the cutting edge. This layer is actually a tube- like volume. Therefore, the turn tool can reduce the diameter of a rotational workpiece.

A human process planner can use his or her experience and imagination to envision the shape a process / tool can create. However, in trying to automate process planning, it is essential to deﬁne the geometric capabilities of manufacturing processes explicitly. During process planning, for a given feature an inverse search is conducted to ﬁnd the candidate process(es) for the feature. The relation- ship between features and processes is deﬁned in a mapping between the two. In an automated process planning system, rules or algorithms are written based on this mapping.

Process for Features Mapping

The geometric capability of a process is summarized in Table 3. As can be seen, milling can create many different features (Volume Capabilities column). Drilling, reaming, and boring primarily create holes. Turning can create different axial symmetric parts. Other processes are also listed in the table.

One can easily ﬁnd the entries for each process and determine the geometric capabilities of the process. Process planning is an inverse mapping. Given a feature, a process planner tries to ﬁnd all processes that can create that feature.

The table alone is not sufﬁcient. To select the correct process based on the geometry, one needs to look into the geometric constraints as well. Figure 15 provides a small sample of process con- straints based on geometry. For example, on the upper-right corner is the ‘‘large hole through a small slot’’ constraint. One should not try to drill such a hole after the slot has been cut. The drill center will be in the air and not cutting any material. It will tend to slip and thus produce an inaccurate hole.

Relative-Cost Table for Manufacturing Processes

When conducting a feature-to-process mapping, one may ﬁnd several candidate processes for the feature. Which process to choose also depends on the cost of the process. The process cost equation consists of a few terms: the tool and machine costs, the material removal rate, and the energy consumption. The relative cost of a process is the cost of removing a unit volume of material. Since the machining time is the inverse of the material removal rate (for a given machining volume), the cost is:

where tool and machine rates are overhead cost of using the tool and the machine and energy cost is the energy cost per unit time.

Processes such as drilling, milling, and turning have higher material-removal rates and thus can ﬁnish a job faster at a lower cost. Finishing processes such as grinding, boring, and polishing have very low material-removal rates, and also consume more energy for the same amount of material removed. The relative cost is higher. Table 4 gives the price of machine costs, which are one of the factors in the relative cost equation. The energy consumption, for example, for cutting cast iron is 0.5–1.2 hp · min/ in3. When grinding is used, the energy consumption is 4.5–22 hp · min/ in3. Non- traditional processes such as a laser process consume much more energy.

Process Capability Analysis

Table 5 shows the technological capabilities of 13 processes. Because each shop may use machines and tools of different precision, the data are for reference only. Please note that all dimensions and tolerances are in inches and all surface ﬁnish values are in microinches. The process capability values can be used to decide whether a process can satisfy the design speciﬁcations of a feature. They can also be used to determine the need of a secondary process (ﬁnishing process). For example, a ﬂat surface has a speciﬁed surface ﬁnish of 20 11in. Using Table 3, we chose ﬂat end mill to cut the surface. From Table 5, we ﬁnd that ﬁnish cut of end mill can create a surface ﬁnish of 50 11in. This is deﬁnitely not sufﬁcient. Yet ﬁnish grinding can create a surface ﬁnish of 2 11in. Therefore, ﬁnish grinding will be used for ﬁnishing and milling for roughing. Milling is chosen for roughing because grinding has a very low material-removal rate. It is not economical to remove all the feature volume using grinding.

Process capability is shop speciﬁc. Each shop needs its own process capability database of its own before process planning can be done automatically. Collecting and analyzing capability data can be tedious. Some of these data can be collected through inspection, such as from the control charts. Others require experiments on the machines. Most of the processes that remove material quickly, such as milling, drilling, and turning, create poorer surface ﬁnish and accuracy.

Cost Model

Another extremely important factor is process economics. We are always interested in ﬁnding the most economical solution. Often it means the survival of the company. Process economics means the cost efﬁciency of the processes. For mass production, a very detailed economic analysis is necessary before a speciﬁc processing method can be selected. However, for the usual small to medium batch production, it is not practical to conduct a very detailed study. The savings cannot justify the amount of effort spent. Some rough estimation or just common sense should be used to select a better process. Whenever there are more than two candidate processes, both technologically suitable for the task, it is time to compare their relative costs. A process cost model can be stated as:

Unfortunately, several difﬁculties prohibit us from using this model to predict the operation cost. First, the cutter path is not known at the process-selection time. Generating a cutter path for each possible process to be machined would be very time consuming. The second problem is the availa- bility of coefﬁcients for each combination of tool-material type and workpiece-material type. There is little published data for tool life equations. Most of the tool life and machinability data are pub- lished in terms of recommended feed and speed. With these two major problems, this approach will probably not work for real-world problems. A quick and dirty way must be found to estimate the cost.

Since we are dealing with the machining of a single feature, it is reasonable to assume that the material-handling time is negligible. The chance of changing a tool during the operation is also minimal. Also, the feed and speed recommended by the Machining Data Handbook (Metcut 1980) usually make the tool life to be about 60 minutes. Since the recommended feed and speed are what are used in most machining operations, it is reasonable to assume that Tl = 1 hr. Therefore, the cost function can be simpliﬁed to:

The above model does not consider the ﬁxed cost of tooling. The tool cost used in the model is the incremental tool cost. If special tools are needed, a ﬁxed cost may be incurred. In that case, the ﬁxed cost must be evenly distributed to the entire batch of parts made.

Tolerance Charting

Tolerance charting is a method for checking the proper in-process dimensions and tolerances from a process plan. It is used to verify whether the process sequence will yield the designed dimensions and tolerances. In most of the literature, process is replaced by operation. In this section we will use the term process. Tolerance charting begins with a part drawing and the process plan. On the process plan are processes and machines. Consequences of processes in terms of resultant dimensions and tolerances are marked on the chart. The processes that were used to produce the dimension and the tolerance are labeled for trace. This is done step by step following the sequence of processes. Finally, the speciﬁed dimensions and tolerances are compared to the resultant dimensions and tolerances. If any of the dimensions and tolerances are not satisﬁed, one can trace back to the sources. Then either a different process / machine is used to reduce the process tolerance or the process sequence is changed.

Figure 16 illustrates how a simpliﬁed tolerance chart works. The example part is a cylindrical part to be turned. The calculation section of the chart is omitted. On the top of the chart is the design. Note that the tolerance chart can handle one dimension at a time. The drawing is 2D and features are vertical lines (representing surfaces). The solid line shows the boundary of a cylinder with greater diameter at the center. The dashed line shows the stock boundary that encloses the part boundary. The designer has speciﬁed dimensions and tolerances between three feature sets. The dimension values are omitted in this example. The next section is the dimension and tolerance section. The thick horizontal lines show where the dimension and tolerance are speciﬁed. For example, the overall dimension is 3 and tolerance is 0.01. The following section is the process plan section. Four cuts are shown in the process plan. The ﬁrst two cuts (10 and 12) use the left-hand side of the stock as the reference. They create two surfaces: surface C and surface D (at the same time the diameters are turned). The next two cuts (20 and 22) create the dimensions between BD and AD. Dimension AB is the result of cuts 20 and 22. Therefore, the tolerance for AB equals the sum of process tolerances for 20 and 22. In this case both are the same. To achieve the designed tolerance of 0.01, the process

tolerance must be less than or equal to 0.01 / 2 = 0.005. The same can be found for CD, which is the result of processes 10 and 12. However, AD is the result of only one cut and thus the tolerance is better.

More details on this subject can be found in Curtis (1988).

Search This Blog

Operation Research course