Planning the polishing path is an important part of the polishing process. Reasonable selection of the polishing path can improve surface quality and reduce the mid-spatial frequency (MSF) error. In this paper, a six-directional pseudorandom consecutive unicursal polishing path and the corresponding generation method are proposed. The proposed path is especially useful for suppressing the MSF error of the surface and realizing consecutive uniform coverage. Compared with traditional polishing paths such as scanning, spiral, and fractal, the proposed path has the following characteristics: multi-directionality, high randomness, smoothness, and consecutiveness. In order to verify that the proposed path has the ability to suppress the MSF error, a group of contrast experiments were conducted and the measurement data was processed by the empirical mode decomposition (EMD) method and the fast Fourier transform (FFT) method. The results show that the fluctuation amplitude of the MSF error and the fluctuation area with MSF error are obviously reduced by using the proposed path. Moreover, the corresponding wavelength of the crests in the spectrum shows multi-values. The experimental results show that the proposed path has high randomness and the ability to suppress the MSF error.

In deterministic polishing, solving the dwell time is one of the key factors. Usually, dwell time is solved by the tool influence function (TIF) and residual error. In previous research, single-point TIF (sTIF) is usually used in calculation, but it is not consistent with the TIF in the actual polishing process. In addition, when using the linear equation to solve the dwell time, a large TIF matrix results in a normal computer not having enough memory for calculation. In order to solve the above problems, a continuous TIF (cTIF) that changes with the polishing path is proposed: first, by the discretization method to simulate the continuous movement of the polishing tool in actual polishing, and then an optimized grouped least squares (LSQR) orthogonal decomposition algorithm is proposed to solve the dwell time. In this paper, an x − y polynomial free-form surface with different initial residual errors ( R M S = 30 n m , P V = 120 n m ; R M S = 70 n m , P V = 280 n m ; and R M S = 100 n m P V = 400 n m ) were simulated by the proposed algorithm, respectively. The final residual error was R M S = 1.8 n m , P V = 13.3 n m ; R M S = 2.6 n m , P V = 10.1 n m ; and R M S = 2.8 n m , P V = 17.4 n m , respectively. The convergence rate of RMS and PV basically reached 95%, and the validity of the algorithm is proved.

Electrorheological (ER) polishing, as a new ultra-precision super-effect polishing method, provides little damage to the workpiece surface and is suitable for polishing all kinds of small and complex curved surface workpieces. In this paper, an ER polishing tool with an annular integrated electrode is developed. The orthogonal experiments are carried out on the six influencing factors of ER polishing which include the applied voltage, the abrasive particle size, the abrasive concentration, the polishing gap, the polishing time and the tool spindle speed. The influence order of these six factors on the ER polishing is obtained. On this basis, the effect of a single process parameter of ER polishing on surface roughness is studied experimentally.

A complex surface measurement is important for quality control and manufacturing processes. Articulated arm coordinate measuring machines (AACMMs) are widely used in measuring the complex surface. However, the AACMMs that are currently used always require manual operation, which reduces efficiency and introduces operator errors. This study presents a measuring device with a 6−DOF industrial robot holding a contact probe, which realizes the automation measurement of a complex surface and eliminates operator errors compared with the traditional measurement process of an AACMM. In order to explore the source of the measuring errors of the device, the influence of three measurement parameters (approaching velocity, contact angle, and measurement position) on the measurement error of the device is analyzed in this paper. A calibration ball measurement experiment is conducted for each parameter. The results show that the optimal approaching velocity of the measuring device is around 2 mm/s, the probe should be as perpendicular as possible to the surface being measured during the measurement, and the maximum measurement error at different positions is 0.1979 mm, with a maximum repeatability error of 0.0219 mm. This study will help improve the automation measuring errors of the AACMM by utilizing an industrial robot to hold the probe, pushing for a wider application of the AACMM.

According to the structural characteristics of the blisk, a new adaptive belt tool system for blisk finishing is developed. The pneumatic servo system, which is composed of the cylinder, the servo valve, and the force sensor, is used to control the polishing force. Due to the strong nonlinearity of the pneumatic system, a two-dimensional fuzzy proportion, integral, derivative (PID) controller is developed for the pneumatic force control. The proposed controller adjusts the proportional, integral, and differential parameters of the traditional PID controller in real time through the error and error rate so as to optimize the control performance of the pneumatic system. Compared with the PID control, the steady-state error of the fuzzy PID control is reduced by 0.03 s and the overshoot is reduced by 4 %, which reveals the superiority of the fuzzy PID control algorithm for the nonlinear system. Finally, the experiments of polishing the blisk sample and the real blisk are carried out by the proposed belt tool system. The results show that the polishing process is very stable and the roughness after polishing is less than 0.4 µm, which proves the effectiveness of the proposed new belt tool system and the fuzzy PID controller.