Method and system for assessing quality of spot welds

• US 20030234239A1
• Filed: 02/20/2003
• Published: 12/25/2003
• Est. Priority Date: 02/20/2002
• Status: Active Grant

• Current Assignee: FCA US LLC (Stellantis NV)
• Original Assignee: Daimler Chrysler Company LLC (Stellantis NV)
• Inventors: Gr. Maev, Roman, Maeva, Elena Yu, Lee, Hsu-Tung, Titov, Serguei A.
• Granted Patent: US 7,132,617 B2
• Time in Patent Office: Days
• US Class Current: 219/109
• CPC Class Codes: B23K 11/24 Electric supply or control circuits therefor
• CPC Class Codes: B23K 31/006 relating to using of neural networks
• CPC Class Codes: B23K 31/125 Weld quality monitoring
• CPC Class Codes: G01N 2203/0051 Piezoelectric means
• CPC Class Codes: G01N 2203/021 Treatment of the signal; Calibration
• CPC Class Codes: G01N 2203/0296 Welds
• CPC Class Codes: G01N 2291/02854 Length, thickness
• CPC Class Codes: G01N 2291/02863 Electric or magnetic parameters
• CPC Class Codes: G01N 2291/2672 Spot welding
• CPC Class Codes: G01N 29/0681 by acoustic microscopy, e.g. scanning acoustic microscopy
• CPC Class Codes: G01N 29/11 by measuring attenuation of acoustic waves
• CPC Class Codes: G01N 29/4481 Neural networks
• CPC Class Codes: G01N 3/00 Investigating strength properties of solid materials by application of mechanical stress

Abstract

A system and method for assessing the quality of spot weld joints between pieces of metal includes an ultrasound transducer probing a spot weld joint. The ultrasound transducer transmits ultrasonic radiation into the spot weld joint, receives corresponding echoes, and transforms the echoes into electrical signals. An image reconstructor connected to the ultrasound transducer transforms the electrical signals into numerical data representing an ultrasound image. A neural network connected to the image reconstructor analyzes the numerical data and an output system presents information representing the quality of the spot weld joint. The system is trained to assess the quality of spot weld joints by scanning a spot weld joint with an ultrasound transducer to produce the data set representing the joint; then physically deconstructing the joint to assess the joint quality.

Claims

1. A system for assessing the quality of a spot-weld joint between pieces of metal, the system comprising:

an ultrasound transducer for probing a spot-weld joint, the ultrasound transducer transmitting ultrasonic radiation into the spot-weld joint, receiving corresponding echoes, and transforming the echoes into electrical signals;

an image reconstructor connected to the ultrasound transducer for transforming the electrical signals into numerical data representing an ultrasound image;

a neural network connected to the image reconstructor for analyzing the numerical data; and

an output system for outputting information representing the quality of the spot weld joint.

2. The system of claim 1, wherein the neural network comprises:

an input layer including the numerical data representing the ultrasound image;

a hidden neuron layer including a first weight matrix defining numerical relationships between the numerical data representing the ultrasound image and an intermediate output; and

an output layer including a second weight matrix defining numerical relationships between the intermediate output and a final output.

3. The system of claim 1, wherein the neural network comprises:

an input layer having nodes for receiving input data representing the ultrasound image;

first weighted connections connected to the nodes of the input layer, each of the first weighted connections having a coefficient for weighting the input data; and

an output layer having nodes connected to the first weighted connections, the output layer defining numerical relationships representing the quality of the spot weld joint.

4. The system of claim 3, further comprising:

a hidden layer having nodes connected to the first weighted connections, the hidden layer being interposed between the input and output layers; and

second weighted connections connected to the hidden layer nodes and to the output layer nodes, each of the second weighted connections having a coefficient for weighting the outputs of the hidden layer nodes.

5. A method of training a computer system to assess the quality of spot weld joints, the method comprising:

scanning a first spot weld joint with a data acquisition device to produce a data set representing the joint;

physically deconstructing the joint;

assessing the joint quality; and

altering the structure of the computer system to elicit a future response to stimuli based on a comparison of the joint quality to the joint to the data set representing the joint.

6. The method of claim 5, further comprising the step of repeating the above steps with at least a second spot weld joint, wherein the computer system is altered to elicit a future response to stimuli based on the results of all joint quality and data set comparisons.

7. The method of claim 6, wherein the steps of altering the structure of the computer system include altering at least one weight matrix in a neural network.

8. The method of claim 7, wherein the step of altering the computer systems includes altering at least two weight matrices using back propagation.

9. The method of claim 8, wherein the step of back propagation includes the steps of:

computing a first layer neuron activation from a first layer neuron activation function, wherein the first layer neuron activation function defines a mathematical relationship between the computer system input stimuli and an intermediate output data set, the mathematical relationship including information from the first weight matrix;

computing a second layer neuron activation from a second neuron activation function, wherein the second layer neuron activation function defines a mathematical relationship between the intermediate output and a final output data set, the mathematical relationship including information from the second weight matrix;

computing a second layer error by comparing the final output to a desired final output;

computing a first layer error based on the second layer error;

adjusting the second weight matrix based on the second layer error;

adjusting the first weight matrix based on the first layer error; and

interactively repeating the above steps to reduce the second layer error.

10. A method of analyzing an ultrasound image, the method comprising the steps of:

identifying critical data items;

assigning each critical data item a weight parameter specifying the reliability of the data stored in a corresponding data item;

performing surface peak detection; and

altering a weight parameter for a data item based on the results of the surface peak detection to reflect a change in reliability.

11. The method of claim 10, wherein the step of altering the weight parameter for the data item includes analyzing a signal-to-noise ratio determined by the surface peak detection.

12. The method of claim 11, wherein the data items represent ultrasound transducer channels.

13. The method of claim 12, further comprising the step of rejecting a channel from further consideration based on the corresponding weight parameter.

14. A method of analyzing an A-scan ultrasound image to reduce the angular dependence of matrix transducer elements and improve C-scan image quality, wherein the ultrasound image represents a surface sample of a spot-weld joint, the method of comprising the steps of:

locating the positions and amplitudes of surface peaks in the ultrasound image;

measuring a global tilt of the surface sample with respect to the transducer surface, using weighted bilinear regression;

using an empiric tilt-amplitude calibration curve to compute an amplitude drop compensation factor based on the global tilt; and

multiplying the value of each sample point in each A-scan by the amplitude drop compensation factor, such that the factor is applied for all transducer channels;

wherein subsequently acquired C-scan images have more stable amplitudes that are less dependent on the transducer tilt than images acquired using un-compensated transducers.

15. The method of claim 14, further comprising the step of building the calibration curve, wherein building the curve includes the steps of:

using a series of measurements on flat-parallel sheets of metal; and

mapping the amplitude of the signal received from the back face of the sheet.

16. A method of compensating for variations in the velocity of sound in different media in the analysis of ultrasound images, wherein an ultrasound transducer system is used to acquire ultrasound images, the method comprising the steps of:

acquiring a set of A-scan ultrasound images;

shifting each A-scan along its time axis so that the position of surface peaks becomes the same for all ultrasound transducer channels; and

building a C-scan image from the A-scan images, wherein the C-scan image reflects the compensation.