JPH1131579A - Inspection method for induction hardened part - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for inspecting induction hardened parts which automates a process of inspecting induction hardened parts by fuzzy inference, prevents dispersion and oversight in the inspection by workers and saves labor. SOLUTION: In inspecting a work 4 hardened by high frequency heating, a work 4 being heated for hardening is image picked up from the predetermined direction so as to detect lightness F of a surface of the work in the heating condition of the predetermined hardening pressure, current and time. A work 4, which is cooled by the coolant after the heating for hardening, is image picked up from the predetermined direction so as to detect a density L of the surface of the work in the heating condition and the condition of the predetermined cooling temperature and the quantity of the coolant. Degree of match for the lightness data F of the surface of the work be heated for hardening and the density data L of the surface of the work after the cooling for hardening in relation to the reference data are discriminated on the basis of the fuzzy inference, so as to judge the degree of match per each data in totality.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting induction hardened parts.

[0002]

2. Description of the Related Art In order to partially increase the hardness of a workpiece, there is a step of partially quenching by high-frequency heating. In the quenching step, for example, as shown in FIG. 3A, an annular quenching coil (2) is externally fitted to the shaft-like work (1), and the work (1) is rotated on the coil (2) while rotating the work (1). After applying quenching voltage and current for a predetermined time and quenching and heating by high frequency induction heating, the work (1) is cooled with a predetermined temperature and amount of cooling water.
Is to quench. Examples of the work include a work having a bent portion, for example, a crankshaft, in addition to the axial work (1).

[0003] At this time, the quenching quality is affected by heating conditions such as quenching voltage, current and time, and cooling conditions such as cooling temperature and amount, and the surface (1a) to be heated of the work (1) is determined according to each of these conditions. ) Changes to a predetermined color. For example, during quenching and heating, the workpiece heated surface (1a) glows red as shown in FIG. 3A, and after quenching and cooling, the same heated surface (1a) as shown in FIG. 3B. Turns blackish.
Therefore, during quenching heating and after quenching and cooling, an operator visually checks the discoloration state of the appearance of the work-heated surface (1a) to inspect the quenching condition.

[0004]

According to the above-mentioned inspection means for induction hardened parts, the worker visually inspects and checks the discoloration of the work appearance during quenching heating and after quenching and cooling. There is a problem that the inspection results vary, and that oversight occurs.

It is an object of the present invention to provide a method of inspecting induction hardened parts by automating the inspection process of the induction hardened parts by fuzzy inference, preventing variation and oversight of the inspection by the operator, and saving labor. It is.

[0006]

According to the present invention, when inspecting a work hardened by high-frequency heating, the work being hardened and heated is imaged from a predetermined direction to obtain a predetermined hardening voltage.
A step of detecting the lightness of the work surface under current and time heating conditions; and imaging the work cooled by a cooling liquid after quenching heating from a predetermined direction, and shading the work surface under the above heating conditions and a predetermined cooling temperature and liquid amount. And the degree of coincidence of the brightness data of the work surface during quenching and heating and the density data of the work surface after quenching and cooling with the reference data are determined by fuzzy inference, and the respective degrees of coincidence are comprehensively determined and the induction hardening work is performed. Inspection step.

[0007]

1 (a) to 1 (d) and FIG. 2 show an embodiment of an inspection method for an induction hardened part according to the present invention.
This will be described below with reference to (a) and (b). First, Figure 1
As shown in (a), after the work (4) is inserted into the hardening coil (3), a predetermined hardening voltage (V) and current (I) are applied to the hardening coil (3) for a predetermined time (t). The work (4) is quenched and heated. Therefore, when the work (4) during quenching and heating is imaged by a camera (5) from a predetermined direction, for example, from above, as shown in FIG. 1B, the work surface at a predetermined hardening voltage (Va) and current (Ia) is obtained. An analog measurement data curve (Da) showing the temporal change of the lightness (F) is drawn.

Therefore, the measured data curve (Da) is image-processed and divided into predetermined time coordinates (ta) (tb)..., And the analog measured data curve (Da) is digitized. Then, the lightness (Fa) (Fb)... And the reference lightness (Fra) (Frb) at each time coordinate (ta) (tb).
Are determined by fuzzy inference. Further, the quality (F) of the lightness (F) is determined by fuzzy determination over all time coordinates (ta) (tb).

At this time, the judgment is made by the judgment means based on the probability. For example, as shown in FIG. 2A, a membership function (M) of a fuzzy set of lightness (F) at a predetermined voltage (Va), current (Ia), and time (t) during quenching heating, where (M) ZRa) is reference data, (PSa)
(NSa) Each fuzzy set｝ of deviation data and the determination probability (D) are set for each time coordinate (ta) (tb). Thus, for example, the lightness (Fa) at time (ta)
(Aa)
(Ba) is detected. Then, the higher the fitness (Aa) and the smaller the fitness (Ba), the closer to the reference data, the comparison is made with the determination probability (Da).
When a>Da> Ba, the brightness (Fa) at the time coordinate (ta) is determined to be normal, and the determination operation is performed for all time coordinates (ta) (tb). From the results of all the determinations, the quality (F) of the lightness (F) is determined based on, for example, the number of normal determinations or the multiplied value of the total fitness, and the quality of the quenching heating state is also determined. Further, the same determination can be made based on other quenching voltage (V) and current (I), and the optimal heating condition can be selected.

After the quenching and heating of the work (4) is cooled with a predetermined cooling temperature and a predetermined amount of cooling water, the work (4) is entirely surrounded from the side. For example, as shown in FIG. 1 (c), three cameras (6), (7) and (8) surround the entire periphery of the work (4) from the side with the dotted line shown as a visual field boundary for each camera. Therefore, when the work side surface is imaged from all directions, as shown in FIG. 1D, the spatial variation of the density (L) of the work surface under a predetermined heating condition, cooling temperature (Θa) and quantity (Qa) is obtained. , An analog measurement data curve (Db) is drawn. (Ea) (Eb) (Ec)
Indicates the fields of view by the cameras (6), (7) and (8), respectively, and covers all the fields of view with three fields of view (Ea), (Eb) and (Ec). Alternatively, the number of cameras may be increased to four, five, or the like to cover the entire visual field with four or more visual fields.

Then, the measured data curve (Db) is image-processed and divided into predetermined visual field coordinates (Xa) (Xb).
An analog measurement data curve (Db) is digitized. Then, in each of the visual field coordinates (Xa) (Xb)..., The contrast (La) (Lb)... And the reference contrast (Lra) (Lr)
b) The degree of coincidence (Va) (Vb)... with fuzzy inference is determined. Further, the quality of the shading (L) is determined by fuzzy determination over the entire visual field coordinates (Xa) (Xb).

At this time, the judgment is made by the judgment means based on the probability. For example, as shown in FIG. 2 (b), predetermined heating conditions (voltage, current, time) at the time of quenching heating and the density (L) after quenching and cooling at a predetermined cooling temperature (Θa) and a predetermined amount (Qa). Fuzzy set membership function (N) {where (ZRb) is reference data, (PSb) (NSb) each fuzzy set of outliers}, and determination probability (G) is set for each view coordinate (Xa) ... I do. Therefore, for example, the respective degrees of conformity (Ab) and (Bb) with respect to the reference data and the deviation data of the gray level (La) in the visual field coordinates (Xa) are detected. Then, as the fitness (Ab) is larger and the fitness (Bb) is smaller, the reference data is closer to the reference data. Therefore, these are compared with the determination probability (Ga), and for example, Ab>Ga> Bb
, It is determined that the shading (La) in the visual field coordinates (Xa) is normal, and the determination operation is performed in all visual field coordinates (Xa) (Xb).
Do about ... From the results of all the determinations, for example, the density (L) using the number of normal determinations or the multiplied value of the total fitness as a guide
Is determined, and the quality of the quenching quality is also determined.
Further, the same judgment is made based on other heating and cooling conditions, and the optimum heating and cooling conditions can be selected.

Further, the degree of quenching is comprehensively determined from the determination results of the lightness (F) and the shading (L). Note that the greater the number of divisions of time and field coordinates, the higher the accuracy of the determination.

[0014]

According to the present invention, when inspecting an induction hardened part, the brightness of the part surface during quenching heating and the density of the part surface after quenching and cooling are measured, and the degree of coincidence with reference data is determined by fuzzy inference. After the judgment and inspection, the inspection process of the induction hardened parts is automated and labor saving,
In addition, it is possible to prevent variation and oversight of the inspection by the operator, and the inspection accuracy is greatly improved.

[Brief description of the drawings]

FIG. 1A is a side sectional view showing a workpiece during induction hardening and heating and an imaging camera thereof. 2B is a graph illustrating a temporal change in the brightness of the workpiece captured by the camera in FIG. (C) is a top view showing the imaging camera of the workpiece after quenching and cooling. (D) is a graph showing the change in the density of the work imaged by the camera of FIG.

FIG. 2A is a membership function of the brightness of a workpiece during quenching and heating. (B) is a membership function of the density of the work after quenching and cooling.

FIG. 3A is a perspective view of a workpiece and a quenching coil during induction hardening and heating. (B) is a side view showing the work after quenching and cooling.

[Explanation of symbols]

 4 Work F Lightness L Shading

Claims (1)

[Claims] When inspecting a work hardened by high-frequency heating, a step of detecting the brightness of the work surface under heating conditions of a predetermined hardening voltage, current and time by imaging the work being hardened and heated from a predetermined direction; A step of imaging the work cooled by the cooling liquid after the quenching and heating from a predetermined direction to detect the density of the work surface at the above-mentioned heating conditions and the predetermined cooling temperature and liquid amount; and brightness data and quenching of the work surface during the quenching and heating. Inspecting the induction hardened part by determining the degree of coincidence of the density data of the work surface after cooling with the reference data by fuzzy inference, and comprehensively judging each degree of coincidence to inspect the induction hardened workpiece. Method.