Implementing real-time machining process control at shop floor has great significance on raising the efficiency and quality of product manufacturing. A framework and implementation methods of real-time machining process control based on STEP-NC are presented in this paper. Data model compatible with ISO 14649 standard is built to transfer high-level real-time machining process control information between CAPP systems and CNC systems, in which EXPRESS language is used to define new STEP-NC entities. Methods for implementing real-time machining process control at shop floor are studied and realized on an open STEP-NC controller, which is developed using object-oriented, multithread, and shared memory technologies conjunctively. Cutting force at specific direction of machining feature in side mill is chosen to be controlled object, and a fuzzy control algorithm with self-adjusting factor is designed and embedded in the software CNC kernel of STEP-NC controller. Experiments are carried out to verify the proposed framework, STEP-NC data model, and implementation methods for real-time machining process control. The results of experiments prove that real-time machining process control tasks can be interpreted and executed correctly by the STEP-NC controller at shop floor, in which actual cutting force is kept around ideal value, whether axial cutting depth changes suddenly or continuously.

Machining efficiency and quality of finished parts can be improved by monitoring, analyzing, and diagnosing the process of product manufacturing. It is hard to model the machining process accurately due to its complexity and variability. Therefore, artificial intelligent algorithms are usually used to build the relational model of machining parameters, cutting tool wear, and quality of finished part, with the purpose of developing machining process controllers that shorten machining time, prevent damage of tools, and improve quality of finished part. Li et al. used back propagation neural network for multiobjective cutting parameters optimization in sculpture parts machining to increase surface quality [

The standard of STEP-NC supports bidirectional data transmission between CAD/CAM systems and CNC systems, which provides a possible way for solving the first bottleneck. Kumar et al. presented a STEP-NC compliant process control framework for discrete components and relevant self-learning algorithms in order to compensate errors and improve the surface quality of finished part [

In this paper, a framework and implementation methods of real-time machining process control based on STEP-NC are studied. STEP-NC data model for real-time machining process control is defined in order to transfer high-level product information between CAD/CAM systems and CNC systems. A software-based STEP-NC controller, which interprets and executes high-level information in STEP-NC files, such as manufacturing features, machining operations, and real-time machining process control functions, is designed and developed. An adaptive control algorithm based on fuzzy logic is also embedded within the software kernel of STEP-NC controller in order to improve the real-time performance of machining process control. The STEP-NC data model, software STEP-NC controller, and implementation methods proposed in this paper can be used to realize real-time machining process control at shop floor. The rest of the paper is organized as follows. In Section

Machining process control methods are usually designed based on automatic control theories such as classical control theory, modern control theory, information theory, system theory, and artificial intelligence theory. Machining process control systems become more integrated, interoperable, and intelligent with the development of computer and network technologies in recent years. The procedure of machining process control consists of three stages, namely, information collection, information processing, and control output. According to real-time property, machining process control can be classified into three categories.

Machining process controllers (MPC), which analyze the process condition and adjust machining parameters, can be implemented at process planning stage or shop floor stage as shown in Figure

Implementation of machining process controller.

Fluctuation of cutting force has great influence on machining system stability, cutting tool life, dimensional accuracy, and surface quality of finished part. It is necessary to select proper machining parameters with constraint of maximum or optimal cutting force. In this paper, the control of cutting force at specific direction is chosen as controlled object to implement real-time machining process control. As shown in Figure

Side milling with end mill.

Edge cutting coefficients

STEP-NC data transfer standard describes the machining process plan with entities such as

EXPRESS-G diagram of real-time machining process control.

Entity

Entity

STEP-NC file contains high-level information of product manufacturing without low-level tool path, in which entity

C++ classes for STEP-NC real-time machining process control information.

CNC systems should be able to interpret the STEP-NC files that contain machining process control operations directly, acquire machining process condition data in real-time, and compute interpolation points and optimized machining parameters simultaneously. Most commercial CNC systems have limited interfaces for STEP-NC interpretation and real-time adaptive control. In this paper, A STEP-NC controller based on open architecture software CNC kernel is proposed and developed in order to implement real-time machining process control at shop floor. An integrated data model based on STEP-NC is used to describe the information of geometry, technology, process planning, and machining process condition, which makes all stages of product manufacturing process traceable. The STEP-NC controller interprets STEP-NC files directly while communicating with sensors without external data acquisition and analysis system. Adaptive control algorithms can be embedded into interpolation calculation procedure in order to optimize machining parameters. The architecture of machining process control system that consists of three subsystems is shown in Figure

Architecture of machining process control system based on STEP-NC.

STEP-NC interpretation module is responsible for interpreting and executing STEP-NC files that contains real-time machining process control functions directly. Most of the manufacturing features derived from entity

Realization methods of real-time machining process condition monitor along with algorithms for real-time machining process control are studied and developed based on the open STEP-NC controller proposed and built in Section

Procedure of real-time process control based on STEP-NC.

The complexity of cutting process makes it hard to be modeled accurately with state-space equations. Artificial intelligence algorithms, which are able to handle unpredictable, nonlinear, multivariable, and incertitude controlled objects, can be a feasible solution. However, most artificial intelligence algorithms are used for off-line machining process optimization due to the limitation of computational complexity. Fuzzy logical algorithm, which has high computational efficiency, is suitable for real-time machining process control. In this paper, a fuzzy control algorithm with self-adjusting factor is proposed to adjust the feed rate in real-time in order to get constant cutting force. The principle of proposed control algorithm, which is designed by modifying the conventional fuzzy control algorithm, is illustrated in Figure

Constant milling force control algorithm based on fuzzy logic.

The error and error variety rate of ideal cutting force and actual cutting force can be calculated as

Then the error and error variety rate are quantified according to the domain of fuzzy rule input as^{−1}), ^{−1}).

The quantitative factors

Membership function of fuzzy rule.

The output fuzzy language values

The self-adjusting factor ^{2}), ^{2}), and ^{2}).

Factors

Experimental part for validation of fuzzy control algorithm.

The test results are shown in Figure ^{2} and more steadily when ^{2}. The change of cutting force signal is more smooth when ^{−1}. So in this paper, adjusting factors ^{−1}, and 4.0 mm/s^{2}. The results indicate the fuzzy control algorithms proposed in this paper can work properly without building the model of metal cutting process with state-space equations.

Selection of adjusting factors.

0.9-10.8-8

0.9-7.2-8

0.9-10.8-4

0.9-7.2-4

Experiments are designed and carried out to verify the function of proposed STEP-NC controller for real-time machining process control. The experiment platform consists of Qier XKV715 vertical milling machine, Kistler 9257B dynamometer, and industrial computer is shown in Figure

Experiment platform.

Two test parts made of aluminum alloy are machined on the experiment platform as shown in Figure

ISO-10303-21;

HEADER;

FILE_DESCRIPTION((

FILE_NAME(

FILE_SCHEMA((

ENDSEC;

DATA;

/

/

/

#23 = MACHINING_WORKINGSTEP(

#29 = PLANAR_FACE(

#45 = SIDE_FINISH_MILLING($,$,

#50 = MILLING_TECHNOLOGY(100.0,.TCP.,

#53 = CONST_FORCE_MILLING(.F.,.F.,.T.,

/

/

/

#54 = CARTESIAN_POINT(

#55 = CARTESIAN_POINT(

#56 = MILLING_FORCE_SAVE(#23,#54,#55,1000.0,

ENDSEC;

END-ISO-10303-21;

Parts and cutting condition of experiments. (a)

Results of real-time machining process control.

Conventional machining

Constant force milling

Conventional machining

Constant force milling

Results of real-time machining process control are presented in Figure

This study has focused on the information exchanging mechanism, implementation method, and system scenario of real-time machining process control. An advanced solution method for optimization problems in manufacturing is proposed for implementing real-time machining process control at shop floor. Two key issues for implementing real-time machining process control are studied and solved. The issue of exchanging high-level product manufacturing data between CAPP systems and CNC systems is solved by extending STEP-NC standard and building an open STEP-NC controller that interpret STEP-NC data directly at shop floor. The issue of implementing real-time machining process control at shop floor is solved by integrating adaptive control algorithm with interpolation algorithm in software CNC kernel. Cutting force at specific direction is chosen to be the controlled object of real-time machining process control to demonstrate the newly proposed STEP-NC data model and implementation methods. A fuzzy control algorithm with self-adjusting factor is proposed to keep the cutting force constant by adjusting feed rate in real-time. The verification of proposed STEP-NC data model and implementation method is carried out on an open CNC platform. The experiment results indicate that the STEP-NC controller is able to interpret STEP-NC file with real-time machining process control functions correctly and keep the cutting force at specific direction close to ideal value. More research works on performance of real-time machining process control algorithms will be carried out based on the proposed implementation method in the future.

The authors declare that they have no competing interests.

This study is financially supported by National Science and Technology Major Projects of China (Grant no. 2013ZX04013-011).