According to one aspect, a system and method is featured for controlling a process parameter of a thermal processing system by estimating an arc voltage between the tip of the plasma arc torch and a metallic workpiece and controlling the process parameter based on the estimated arc voltage. For example, particular embodiments can include adjusting the height of a plasma torch based on an estimated arc voltage.
According to another aspect, a system and method is featured for estimating an arc voltage in a thermal processing system that includes a switch mode power supply for providing an arc current to generate a plasma arc between the tip of the plasma arc torch and a metallic workpiece.
The PWM control circuit block 200 provides a gate signal T3PWM to the switch Q1 to control its duty cycle, and thus the output current IARC through the plasma arc load RLD. As shown, the PWM control block 200 includes a current reference block 210, an error control block 220, a feedback current sensor 240, a PWM comparator block 230, and an arc voltage estimation module 250.
An operator of the system manually sets block 210 to a desired current reference IREF at which to maintain the output current IARC. The output current IARC is monitored using the current sensor 240, such as a Hall current sensor. The current sensor 240 transmits a feedback current IFB to an input of the error control block 220. The error control block 220 can be implemented, for example, as a standard proportional-integral-derivative controller (PID controller) known to those skilled in the art. The error control block 220 compares the feedback current IFB against the desired current reference IREF and outputs a modulating error signal, Error.
The error signal, Error, is then input to the PWM comparator block 230 where it is sampled and used to generate the appropriate gate signal T3PWM that adjusts the duty cycle of the switch mode power supply 300, thereby correcting for the error in the output current. The PWM comparator block 230 and the arc voltage estimation module 250 can be realized using a digital signal processor (DSP), such as TMS320LF2407A from Texas Instruments. These control blocks can also be realized using a combination of one or more suitably programmed or dedicated processors (e.g., a microprocessor or microcontroller), hardwired logic, Application Specific Integrated Circuit (ASIC), or a Programmable Logic Device (PLD) (e.g, Field Programmable Gate Array (FPGA)) and the like.
In order to generate the appropriate gate signal T3PWM, the PWM comparator block 230 compares an instantaneous error sample T3CMPR with a carrier wave signal T3CNT. The carrier wave signal can be generated as a sawtooth or triangular carrier wave with its frequency ranging anywhere from hundreds of Hertz to MegaHertz depending on the application. In a plasma cutting application, the frequency of the carrier wave signal is typically around 15 kHz. The comparator amplifies the difference between the two signals and produces a gate signal T3PWM whose average value over one switching cycle of the carrier wave signal T3CNT is equal to the value of the instantaneous error sample T3CMPR. Application of the gate signal to the switch Q1 adjusts the duty cycle to drive and maintain the output current IARC at a desired steady state value.
According to a first embodiment, the method for arc voltage estimation is based on the principle that inductor voltage drop is zero at constant arc current IARC. This implies that the average dc voltage at the input of the inductor L1 is equal to the average value of the arc voltage VARC. Thus, an estimate of the average arc voltage VARC can be determined by calculating the product of the steady state duty cycle DSS of the switch Q1 and a dc input voltage VIN according to equation (1):
V
ARC
=D
SS
*V
IN (1)
At step 410, the arc voltage estimation module 250 obtains the dc input voltage VIN. As shown in
The dc input voltage VIN can also be determined from the input ac voltage VACIN (not shown). The input ac voltage VACIN is an ac voltage from which the dc input voltage VIN can be derived, for example, through a rectifier stage. The arc voltage estimation module 250 can determine the dc input voltage VIN from the peak value of the input ac voltage VACIN. The dc input voltage VIN can also be derived from the root mean square (RMS) value of the input ac voltage VACIN. Other methods for translating an input ac voltage to a dc input voltage can also be implemented. Because the input ac voltage VACIN can have a magnitude in the range of hundreds of Volts, signal conditioning circuitry 310 is used to scale down the input ac voltage VACIN to a voltage suitable for processing by the arc voltage estimation module 250.
At step 420 of
According to a second embodiment, the method for arc voltage estimation additionally accounts for the voltage drop in the inductor according to the Equation (2) below.
where current sample Is is a present sample of the inductor current, current sample Is*z−1 is a preceding sample of the inductor current, and L is the inductance of the inductor L1. In this example, Equation (2) is discretized using a backwards Euler transform. However, other discretization transforms known to those skilled in the art can also be used. For example, another discretization transform is the Tustin transform (also referred to as the “Bilinear Z” transform) Other substitutions can be possible.
At step 540, the arc voltage estimation module 250 calculates the estimate of the arc voltage VARC based on the difference between the voltage applied to the input of the inductor L1 from step 520 and the calculated voltage drop across the inductor from step 530. For example, after substitution of Equation (3) into Equation (2), the arc voltage estimate VARC can be obtained from the following:
Equation (4) provides an accurate estimate of arc voltage but in practice is sensitive to noise in the current measurement Is and requires low pass filtering that significantly affects the estimate. Also Equation (4) implicitly assumes that the output voltage changes so slowly as to be essentially constant throughout the PWM switching period and makes a sudden step change at the sampling instant. In the case of plasma arc loads, this assumption generally does not hold. Rather, the voltage across a plasma arc can be highly dynamic with rapid changes relative to typical PWM switching periods.
According to a third embodiment, the accuracy of the arc voltage estimate can be further improved by starting with the assumption the arc voltage VARC changes throughout the PWM switching period. Many different profiles can be assumed for the change in arc voltage VARC, including linear, parabolic, exponential profiles, for example.
At step 640, the arc voltage estimation module 250 models the change in arc current through the inductor based on the voltage applied to the input of the inductor and the time varying profile of the expected variations in arc voltage. At step 650, the arc voltage estimation module 250 derives a model of the arc voltage based on the model of the change in arc current through the inductor. At step 660, the arc voltage estimation module 250 calculates the arc voltage estimate from the model derived in step 650.
Although not so limited, the following is an example of a method for estimating the arc voltage according to the third embodiment. For the purpose of example only, the arc voltage is assumed herein to vary linearly throughout the PWM switching period. However, as previously discussed, the variation in arc voltage over time can be modeled as linear, parabolic, exponential, or using any other mathematical or statistical representation.
The following table includes a description of terms discussed in following example for estimating the arc voltage according to the third embodiment.
A single switching period begins with a current sample and ends with a current sample. The average voltage applied to the output circuit (i.e., the inductor and the load) is:
V
applied
=D*V
IN (1)
The basic equation for the voltage across an inductor is:
Converting to a discrete form we obtain:
The change in the Arc Voltage between sampling instants is:
ΔVarc=Varc−Varc*z−1=Varc*(1−z−1) (4)
The rate of change of the Arc Voltage is:
The change in the Arc Current between sampling instants is:
ΔIs=Is−Is*z−1=Is*(1−z−1) (6)
Assuming a linear uniformly changing Arc Voltage:
Simplifying:
Back substituting
Back substitutingΔVarc=Varc*(1−z−1) (11)
Solving for VARC:
Solving for a recursive, implementable form:
This technique is extendable to other models of arc voltage behavior including, for example, parabolic models in which the arc voltage varies with t2.
With respect to all of the embodiments, one or more of the steps described can be combined as known to those skilled in the art.
While this invention has been particularly shown and described with references to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.
This application claims the benefit of U.S. Provisional Application No. 60/825,470, filed on Sep. 13, 2006. The entire teachings of the above application are incorporated herein by reference. This application relates to co-pending U.S. patent application Ser. No. ______, (Attorney Docket No. HYP-068) entitled “LINEAR, INDUCTANCE BASED CONTROL OF REGULATED ELECTRICAL PROPERTIES IN A SWITCH MODE POWER SUPPLY OF A THERMAL PROCESSING SYSTEM,” filed concurrently herewith. The entire teachings of the above application are incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
60825470 | Sep 2006 | US |