Search

Surface texture generation in machining processes is examined in view of establishing a model building procedure capable of linking surface finish parameters to machining conditions. In the first part of the paper, with reference to cutting processes with defined tool geometry, a model capable of describing a surface profile as an ordered sequence of tool marks is developed. The use of such a model for process control is also discussed. In the second part, the identification procedure of a model describing profile generation in face milling versus cutting parameters is presented.

A key problem in transfer line design is that of determining what size should be assigned to intermediate storage buffers in order to maintain a desired production rate despite the imbalances in supply and demand which may arise within the system. In this paper, a deterministic model of production and storage operations of a transfer line, subject to breakdowns and subsequent repair of stations, is presented. Then a simple method, based on the cumulative production curve of each station, for analyzing transfer lines operations is suggested. An extensive use of the method, both numerically and graphically, allows an accurate selection of intermediate storage capacities in the light of guaranteeing the system a stationary mean production rate.

A tool-life model based on the assumption that wear and fracture are the causes of tool death is re-examined from the theoretical standpoint, and extended to include the effect of cutting speed on the fracture-induced failure rate. A stochastic model for multi edge cutting tools is also proposed. This enables the overall statistics to be derived from the data for each individual cutting edge.

A simple machining operation is modeled according to some management choices by stochastic simulation, taking into account two main tool failure mechanisms. Under the assumption that premature tool failures account for either work interruption or production of defective pieces, the effects of tool change schedule, production block size and sudden failure on production rate and cost are assessed. Some minimum regions are found for expected values of dependent variables, which have no counterpart for upper bound values.

The joint distributions of economic parameters corresponding to “optimum” conditions (minimum cost, and maximum production rate), as affected by a single tool wear mechanism, are analyzed and the information process is examined using a Bayesian approach. Results obtained for an example show that even scanty experimental evidence may be profitably exploited. However, the size of confidence regions for parameters is quite large unless a very large amount of information is available.

The paper deals with the main design problem for series production systems, namely the choice of efficiency-optimal values of interstage buffer capacities. It is shown that the design tool must be an accurate dynamic model, able to describe both work station reliability and the couplings among stations (i.e., blocking and starving situations) due to finite storage spaces. By use of this model, the buffer size planning problem can be solved by developing a sensitivity analysis of the model operation modes with respect to buffer capacity values. According to this approach, an analysis of the transfer line efficiency reduction, due to either nonoptimal capacity values or station reliability variations, is also performed.

Under production conditions cutting tools often fail under several failure modes, the occurrence of a single one only for a given operation being rather exceptional. In light of this observation a stochastic model is developed, considering as causes of tool failure both wear and fracture processes. Machining economics are then analyzed with a probabilistic approach, deriving distribution functions of profit rate.

The effect of tool-life scatter on the uncertainty of parameters of a typical tool-life model is analyzed using the joint confidence region approach. Some apparent contradictions concerning tool-life test data are explained in terms of their information content in the light of personal probability concepts.

The design problem of series production systems refers to the choice of an efficiency-optimal set of values for station parameters (production rate, life and maintenance period) and intermediate storage capacities. Useful criteria for system design can only be obtained if based on a dynamic model with a reduced number of states (Buzacottand Hani fin, 1978). Present paper firstly deals with the model definition, in terms of a discrete event dynamic system whose states are driven by the occurrence of events due both to station unreliability and to finite internal storage capacity. Then efficiency evaluation criteria are investigated based on the sensitivity analysis of the obtained dynamic aggregate model (DAM) versus buffer capacities and station life or maintenance period variations.