WO2001044681A2

WO2001044681A2 - Hybrid digital-analog controller

Info

Publication number: WO2001044681A2
Application number: PCT/US2000/042182
Authority: WO
Inventors: Clark Radcliffe; Charles B. Birdsong
Original assignee: Board Of Trustees Operating Michigan State University
Priority date: 1999-11-17
Filing date: 2000-11-16
Publication date: 2001-06-21
Also published as: AU4708501A; WO2001044681A3; WO2001044681A9

Abstract

A Hybrid Digital-Analog controller (10) configured to dynamically adjust the gains of the control parameters of an analog controller (20) to optimize the control parameters of a monitored controlled variable. The controller (10) receives a sampled signal representative of the controlled variable, and creates a digital control signal (40) based on the sampled signal. The analog controller (20) has a plurality of digitally variable resistors (45) which control the gains of the analog controller (20).

Description

HYBRID DIGITAL-ANALOG CONTROLLER

Cross Reference to Related Applications This application claims the benefit of priority of U.S. provisional application Serial No. 60/165,989 filed November 17, 1999, the entire contents of which are hereby incorporated by reference into the present application.

FIELD OF INVENTION This invention relates to an apparatus and method for adjusting the gains of a controller, and more specifically to an apparatus and method for adjusting the gains of an analog proportional-integral differential controller using a digital processor.

BACKGROUND OF THE INVENTION Feed back and feed forward controllers ("controllers") are commonly employed to maintain temperature, humidity, pressure, and flow rates for heating, ventilating, and air-conditioning equipment. They are also used in a variety of other systems such as vibration control, chemical process industry, and to control mechanical servos. In such systems, the controller adjusts the control action based on a feedback signal indicative of the "controlled variable". The feedback signal is generated by a sensor disposed to monitor the controlled variable. A feed forward signal can be used to adjust the controlled action by monitoring controller inputs. The object of such controllers is to control the system in such a way as to maintain the controlled variable, as sensed by the feedback signal, at a desired level (the "setpoint"). For example, the controller of a vibrating system attempts to maintain the amplitude of the vibrations in the system at a specific level. When the actual amplitude of the system deviates from the desired amplitude, the controller must appropriately adjust the control action of an actuator to bring the actual system amplitude back in line with the desired value. Thus, if the feedback signal indicates that the actual amplitude is lower than the desired amplitude, the controller will cause the control action of the actuator to change so that the actual amplitude of the system increases. Likewise, if the feedback signal indicates that the actual amplitude is higher than the desired amplitude, the controller will cause the control action of the actuator to change so that the amplitude of the system decreases.

An ideal feedback control system would be able to maintain the controlled variable at the setpoint based only on the feedback signal. However, actual feedback control systems require additional inputs known as control parameters.

Control parameters are values used by a controller to determine how to control a system based on the feedback signal and the setpoint.

Several methods for controlling a closed-loop system are known. Often, these methods call for manually adjusting the gains in an analog controller after inspection of a systems response. Digital controllers can have their control parameters adjusted in use by various error-checking methods. Digital controllers have the significant problem of being limited by their input sampling rate to systems with response speeds much slower then the digital controller's input sampling rate. The required Digital to Analog ("D/A") and Analog to Digital ("A/D") converters to allow high speed sampling significantly increase system costs.

As these control parameters directly affect the performance and stability of a controller, it is important to determine the appropriate values of these parameters.

However, the appropriate values for these parameters may change over time as the system is used. For example, in a vibrating system, the dynamics of a process may be altered by system damping, inherent nonlinear behavior, ambient variations, environmental changes such as large and frequent disturbances, and unusual operations status, such as failures, startup and shutdown. The process of adjusting the control parameters of a controller to compensate for such system changes is called retuning. If a controller is not retuned, the control response may be poorer than a retuned system. For example, the controlled variable may become unstable or oscillate widely with respect to the setpoint. Thus, to insure adequate performance, controllers must be periodically retuned with new control parameter values.

Many industrial processes create unwanted vibration that can be reduced by careful application of engineering controls. These can be divided into two classes: applications where controls are added to reduce unwanted vibration from the environment outside of the system, and second, applications where controls are added to change the vibrational response of the system to improve performance. The first deals with the attenuation of unwanted environmental vibration. One example is automobile exhaust noise. The automobile engine creates a pressure pulsation in the engine that transmits down the exhaust system where it is injected into the environment through the tail pipe. Humans perceive this pressure pulsation as undesirable noise. The traditional engineering control consists of adding a muffler to an exhaust system that dampens the dynamic pressure variation resulting in reduced environmental excitation of automobile vibration and noise. Other examples of this class of problem and the traditional noise controls include acoustic liners in jet engines, applying Helmholtz resonators to wind tunnels and applying mufflers to pneumatic exhaust nozzles. A second class of vibration control problem deals with adding or removing vibrational resonances in a system to improve performance, product quality, or safety. An example of this type of problem is an industrial robot arm where vibrations cause production errors. Other examples of this class of problem include tuning rotating shafts using rotation dampers. In cases where a narrow frequency band of vibrations exists in a structure

(referred to here as the primary vibration system), a traditional engineering control consists of adding a passive mechanical resonator, typically a Vibration Absorber ("VA"), which are tuned to reduce the unwanted noise/ vibration. The VA is a passive device that consists of mass coupled through a spring to the primary vibrating system. The VA absorbs vibration energy at its resonant frequency decreasing the vibration of the primary vibration system. The VA resonant frequency is determined by the physical parameters of the mass and spring. Past effort by others has been made to physically vary the resonator parameters with time, to achieve a variable tuned device. A driven force actuator incorporated into an VA can be used to create a desired vibrational impedance and thus change the noise reducing properties of the VA without physical changes to the physical mass and spring. Until now, adaptive noise control was prohibitively expensive due to costly digital signal processing computer control.

SUMMARY OF THE INVENTION Accordingly, this invention provides a hybrid-analog controller that overcomes the problems and disadvantages of conventional controllers. The system utilizes a hybrid digital proportional-integral controller to actively control vibrating systems at higher resonant frequencies and at lower cost than ever before possible.

The controller described below incorporates the key developments in semiconductor technologies that now make it possible to implement high speed active noise and vibration control at a fraction of the costs of previous equipment. The controller is described in a non-limiting example as a PI D controller being combined with a passive semi-active resonator). It is envisioned the techniques and apparatus disclosed herein is equally applicable to other controllers, e.g., a lead-lag controller. The device is effective in reducing vibration or acoustic noise from propagating in a component. The SR consists of a passive resonator with the addition of a sensor and controller driven compensated actuator on one surface of the vibrating member. By varying the controller gains, the vibrational or acoustic impedance of the SR can be varied from a nominal value, and the resonant frequency and amplitude of the resulting vibrational/acoustic filter can be changed on line. The system described herein provides a solution of using active control to modify the acoustic response of a passive acoustic resonator or vibrational resonator in real time. A successful working model of the Semi-active Resonator ("SR") can be used in systems such as compressed air lines to continuously reduce plant noises associated with an air compressor and thus provided improved safety and comfort for manufacturing plant operators. Importantly, the SR can be used to quiet industrial processes that produce noise, which varies in frequency over time.

This system would function as a self-contained device with few moving parts and integrate smoothly with the primary vibrational or acoustic system, thus eliminating the complexity of changing the physical dimensions of the acoustic system during operation.

The SR system disclosed represents a powerful new tool in tuning resonant systems. The SR adds a nominal resonance to any acoustic or vibrational system and allows the resonance frequency and damping to be charged real-time, continuously over a range of frequencies. It has advantages over other technologies in that it does not add significant mechanical complexity, the design places the sensitive sensor and actuator away from the direct path of the process, and it is fault tolerant. Four SRs are envisioned: the first utilizes a Helmholtz resonator; the second utilizes a quarter wave resonator; the third utilizes known spring mass passive absorber; the fourth utilizes piezo-electric actuators and sensors . Each of the resonator systems are converted to "semi-active" resonators by utilizing actuators and incorporating a sensor into resonator cavity or compliant element. The actuator and sensors being coupled to the controller to allow changes on the resonances of the SR system.

BRIEF DESCRIPTION OF THE DRAWINGS The various advantages of the present invention will become apparent to one skilled in the art upon reading the following specification and by reference to the drawings in which: Figure 1 represents a schematic representation of the preferred embodiment of the present invention;

Figure 2 represents a schematic representation of a feed back controller with the hybrid-digital system of the present invention;

Figure 3 represents a schematic diagram of a digitally adjusted PID controller; Figure 4 represents a single semi-active Helmholtz resonator of the current invention;

Figure 5 is a representation of the semi-active Helmholtz resonator coupled to an acoustic duct;

Figure 6 is a schematic representation of the closed-loop semi-active resonator control system with an adaptive gain scheduler;

Figure 7 is a schematic representation of the closed-loop semi-active acoustic resonator control system;

Figure 8 is a schematic of the hybrid analog/digital controller of the current invention; Figure 9 is a schematic representation of the closed-loop semi-active vibrational resonator coupled to an vibrating body and accompanying control system; and

Figure 10 is a representation of a piezo-electric actuator coupled to a compliant vibrating beam. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Figure 1 shows the configuration of the Hybrid Digital-Analog Controller 10 which has two (2) components: a digital control gain analyzer 15 and an analog controller 20. Digital control analyzer 15 uses controller inputs 25 and analog controller output 30 as well as auxiliary input signals 35 to determine the PID control gains 40 that are optimal for the current operating conditions of the system under control. Digital control analyzer 15 then adjusts the gains on the analog controller 20 to match those optimal gains. Hybrid Digital-Analog Controller 10 combines the flexibility of digital control with the low cost and high speed of analog control elements. Control gains 40 of analog controller 20 are changed through the application of digital signals to digitally controlled resistors 45 (Figure 3). This system has application to vibration control, acoustic noise control, process control, mechanical servo control as well as other analog control systems whose optimal gains change with operating conditions. While fixed-gain PID controller is the most widely employed analog controller design, Hybrid Digital-Analog Controller 10 substantially improves the performance of these controllers by allowing the adaptation of PID control gains through the use of digitally controlled potentiometers in analog op-amp PID control circuits.

A crucial factor in the efficiency and performance of Hybrid Digital-Analog Controller 10 is the accuracy with which digital control analyzer 15 determines the new PID gains 40 after any given change to the inputs 25, 30, and 35. It is envisioned that the digital control analyzer 15 would contain algorithms or a look-up table to set the analog controller 20 gains based upon the values of the input or inputs 25, 30, and 35). Automatic control technology is widely employed in both the industrial sector and for consumer goods. More accurate, responsive and reliable control is possible through the application of feedback (Figure 2). Feedback control actions, a(t) applied to the plant, G_P from controller, G_c are based on the error e(t) = r(t)-m(t) that is the difference between desired response, r(t) and measured response, m(t). The use of feedback increases accuracy of the control response, c(t) but makes the system response and stability very dependent on the design of the controller. The hybrid controller's digital control analyzer uses controller control input, e(t) and output, a(t) as well as well as other auxilary inputs to configure the controller's analog control elements. This ability to digitally reconfigure itself in response to changing operating conditions allows controls designed to optimize control response over a variety of operating conditions. Previous analog controls are unable to reconfigure themselves in this way.

According to a preferred embodiment of the present invention (Fig. 3), at least one measured signal is used to characterize the closed loop response. The signal may be controller output 30, controller input 25, or auxiliary 35 signal. From these parameters, the optimal gains of analog controller 20 are determined and adjusted. Specifically, the gain of analog controller 20 is adjusted using digitally variable resistors 45 located within proportional 50, integral 55, and derivative 60 branches of analog controller 20.

In one implementation of Digital-Analog PID (DA/PID) controller 10, the digital signal analyzer is a micro-controller such as the Basic Stamp ™, BS1-IC by Parallex, Inc., which monitors operating conditions at a relatively slow rate (10-100 times/sec) through measurements made on the controller input and any auxiliary input signals. In this implementation, the micro-controller digital control analyzer 15 computes optimal gains and then sets those gains in the analog controller 20 unit using either sequences of pulses, standard serial protocols or other digital signals. Commercially available digital resistors 45a-45c use a variety of digital interface signals to set their internal value. An example of the digitally controlled resistor is the DS1804 NV 100 step trimmer potentiometer from Dallas Semi-conductor. The analog controller 20 using readily available operational amplifiers has gain- bandwidths of 1-10 MHz. These highly responsive amplifiers in the analog circuit would allow the DA/PID controller to control systems with rapidly changing sensor signals far beyond the operational abilities of purely digital control systems. Because micro-controllers and digital resistors 45a-45c are inexpensive components, a DA/PID control system could be implemented with the flexibility of a pure digital controller at a substantially lower cost. Figure 3 shows the electronic configuration of a hybrid control system 10 using digitally controlled resistors 45a- 45c. Illustrated in the figure are three (3) signals paths for the signals: a) proportional to the input e(t) 50; b) proportional to the integral of e(t) 55; and c) proportional to the derivative of e(t) 60. Each signal path includes an operational amplifier circuit 65a-65c. The output/input gain of these amplifiers 65a-65c is varied through the setting of digital resistors 45a-45c in each signal path for the analog controller's 20 proportional 50, integral 55, and derivative 60 gains. These digital resistors 45a-45c are set by the system's digital control analyzer 15 which monitors controller inputs 25 and outputs 30 as well as any required auxiliary signals 35 to determine the analog controller 20 operating conditions and optimize the control gains 40 in response to those operating conditions.

By way of non-limiting example, the DA/PID can be used in the Semi-Active Helmholtz Resonator ("SHR"). The SHR is an acoustic noise control device whose effective noise control frequency is set by changing its acoustic resonant frequency to match the frequency of the acoustic noise to be controlled. The DA/PID controller monitors the frequency of its input signal, e, and sets the PID controller gains to optimal values based on a stored table of PID gain values vs. input signal frequency.

Figure 4 represents a single Helmholtz resonator 118 of the current invention. The Helmholtz resonator has a cavity 120 of a known volume coupled to a neck region 121 for acoustically coupling the resonator to the duct having the sounds to be cancelled. Incorporated in the semi-active Helmholtz resonator 118 is an actuator 122. This actuator is further coupled to a controller 123 and a microphone 124 disposed within the cavity 120. The HR consists of a rigid-wall acoustic cavity 120, or "cavity" with at least one short and narrow orifice, or "neck" 121 through which the fluid filling it communicates with the external medium as shown in Figure 5. It is an acoustic resonant system whose volumetric flow Q, through the neck, to input pressure. P, at the entrance of the neck can be represented by a second order transfer function. Thus, when a HR is added to another acoustic system, pressure oscillations at the HR resonance frequency excite the HR and are thus reflected back to the source. The device is analogous to a mechanical vibration absorber.

A theoretical model of the Helmholtz resonator version of the SVR without adaptive tuning will be described briefly here so that it can be examined for closed- loop adaptive control application. The SVR consists of a Helmholtz resonator with one surface of the cavity replaced by a moving surface as shown in Figure 1. The system can be represented by linear time invariant state equations of the form

where the states are Q.,, the volumetric flow rate from the neck (m³/s) and V, the sum of the volumes introduced through the neck and the inner surface of the cavity(m³). The inputs are P.,, the pressure at the neck inlet to the cavity (N/m²), and Q₂, the volumetric flow rate from the movable surface in the cavity (m³/s). The outputs are Q₁ and P₂, the pressure in the cavity (N/m²). The other parameters are R_a, the acoustic loss that represents viscous and radiation losses (Ns/m⁵), l_a, the acoustic inertia of the mass of air in the resonator neck (Ns²/m⁵), and C_a, the acoustic compliance of the cushion of air in the resonator cavity (m⁵/N).

A proportional-integral (PI) controller,

can be used to generate a complex impedance on the actuator in the resonator cavity, where K_P and A., are the proportional and integral gains respectively. The closed-loop transfer function for Q/P, can be computed by substituting (3) into (1) and (2) and simplifying which gives

This shows that the dynamics of the system depend of the acoustic parameters, R_a, I_α, and C_α, which are fixed, and the controller gains K_P and Kj which can be varied. It can be shown that K, has the effect of modifying the resonant frequency, and K_P has the effect of modifying the acoustic loss, or damping of the system. A range of gains K_P and K₍ that produce a stable system can be found by examining the denominator of (4). The system is stable provided that there are no sign changes in the coefficients of the denominator of (4). This produces the following ranges for the gains:

These bounds on the gains determine the limitation of the controller to tune the system, defining a design space for the controller gains. The limit on K_P , (7) translates into a maximum amount of damping that can be removed from the system. The limit on K_{ (8) translates into a limit on the direction the resonance frequency can be moved from the nominal value. In this case a negative K_I increases the value of the resonance frequency, and (8) indicates that the resonance frequency can not be decreased from the nominal value.

The design space defined by the range of stable gains can be examined using pole placement analysis. A closed form solution can be found to compute the gains K_P and K_j that make the system respond with a desired resonance frequency and damping. . The desired denominator of the transfer function can be represented by

G(s)_desιred = [(s² + 2ζω_ns + ω_n ² ){s + P)\ ( 9 )

The values of the damping ratio, ζ and the natural frequency, ω_n can be chosen.

The pole P can not be set, but becomes an output of the calculation. Setting (9) equal to the denominator of (4) and matching coefficients in s, the values of K_P,

K_{and P can be computed from the resulting linear system of equations. The solution is given by

R_a <Ca^Ra² + CaIa ^{~ ~} 21 _aC _a R_a 0 „C,)

C_{α a} R_α ² + „² - /_αR_αω„ζ)

A gain scheduled control algorithm can be constructed from the above formulations. The objective is to tune the system so that a resonance with a large peak occurs at an arbitrary command frequency, ω_{c C} where the actual resonance frequency ω„ = ω_c The amplitude of the resonance can be controlled by setting the value of ζ. The strategy used here is to place the complex poles at a constant distance from the imaginary axis for all values of ω_c . This can be accomplished by applying the constraint

ζ = */ω„ ( 13 )

where k is a scalar that defines the distance of the complex poles from the imaginary axis assuming that ζ is small.

For simplicity, the control is described below with the resonator neck open to atmospheric pressure instead of connect to a primary acoustic system. The normal operation is with a primary acoustic system producing P₁ at the resonator neck and this input becomes the disturbance to the system. From the control viewpoint the system can be disturbed by either input, P₁ or Q₂. The motivation for disturbing the system through Q₂ is that this approach eliminates the dynamics associated with the primary acoustic system and focuses the analysis on the dynamics of the SVR alone.

The closed-loop transfer function relating the system output P₂ to the input D is computed from (1 ) and (2) and given by (14). Note the denominator of (14) which characterized the system dynamics is identical to the previous implementation (4) where the disturbance was through P This confirms that the previous stability and gain scheduling analysis applies both when the disturbance is through P₁ or Q₂.

The proceeding gain scheduled controller can be applied to the plant model and the time response was simulated for various inputs. A block diagram of the system is shown in Figure 6. Note that the frequency of P₂ must be determined in real time. This can be done by various methods, including computing the peak in the Fourier transform, using a frequency counter, or using an adaptive frequency estimation algorithm. The previous description assumes an ideal actuator, i.e., the actuator velocity, and Q₂ is proportional to the voltage input to the actuator. This is a reasonable assumption for analysis of the fundamental response of the SHR since the analysis can be applied in general and actuator dynamics can be accounted for afterwards. However, the gain scheduling control strategy must account for actuator dynamics.

In the event that the actuator dynamics degrade the effectiveness of the closed-form mapping between the controller gains and the system dynamic response, characterized by the resonant frequency and damping, an empirical controller design can be used. This technique is based on the qualitative information derived from the model: K_l effects the resonant frequency, and K_P effects the damping. The mapping between the controller gains and the system dynamic response can be determined experimentally by fixing K_j (fixing the resonant frequency) and searching for a K_P that produces the desired damping. This process can be repeated until the mapping is complete. Discrepancies between these results and the closed-form solution can be attributed to errors in the model acoustic parameters, or significant actuator dynamics.

Incorporated into each SHR is an actuator, which is responsive to the controller and capable of producing acoustic excitations. Any actuator meeting this description will function with the current system. One actuator known to work with the current system utilizes a standard piezoelectric element. In the embodiment discussed here, a velocity compensated dual voice coil speaker is used as the actuator for the movable surface of the SHR. This actuator compensates the velocity for internal dynamics and pressure interactions on the face of the actuator.

The result is an actuator with nearly constant magnitude and phase between command signal and actuator output within a limited frequency range of 20 to 400 Hz. The actuator model inputs are the command voltage to the actuator and the pressure on the face of the actuator, and the output is Q₂. The actuator is a significant improvement over uncompensated speakers, however finite magnitude and phase variations remain, affecting the closed-loop SVR performance, even within the frequency range of 20 to 400 Hz. The actuator model can be added to the SVR model by feeding the desired Q₂ into the input to the actuator and P₂, the pressure output from the SHR cavity to the input of the actuator, and the actuator output to the SHR.

Figure 5 represents the Semi-active Helmholtz Resonator of the current invention coupled at its neck 121 to the acoustic duct 125. The acoustic duct 125 having a noise source 126 at one end producing a first incident pressure wave 127.

Upon activation of the semi-active Helmholtz resonator, a second pressure 128 from the SVR results in a third reflected pressure wave 129 . The third reflected pressure wave 129 acts to cancel the first pressure wave 127 at union of the SHR with the acoustic duct 125. The final product at the duct end is a fourth pressure wave 130, having a significantly reduced amplitude from that of the first wave 127.

Figure 6 is a schematic representation of the closed-loop semi-active resonator control system with an adaptive gain scheduler. The semi-active resonator consists of a passive resonator 132 tuned to resonate at a nominal frequency. The pressure in the resonator compliant element is measured by a microphone 124. The controller 134 ensures that the relative magnitude and phase of the input velocity to force creates a dynamic impedance that when combined with nominal impedance of the passive resonator, creates an overall impedance that can be tuned in real-time. The PID controller 136 or any equivalent controller tunes the device's overall dynamic impedance. This tuning changes the devices resonant frequency to increase the resonant vibration mode amplitude enhancing the acoustic noise reduction effect. The semi-active resonator control algorithms consist of two components each with different time scales: one fast 138 and one slow 161.

In the fast time scale, the PID controller 136 component, takes the signal 138 from the microphone 124 and generates an output 142 that is the sum of a proportional, integral and/or derivative signal, gains K_P, K and K_D respectively. This output 142 is then fed to the actuator 122 that inputs a velocity to the resonator compliant element. The PID controller 136 gains are not held fixed as in traditional PID controllers. The gains are tuned in real time by a slow time scale control analyzer tuning component 140. It is envisioned that the PID controller is implemented as a tunable analog controller. This allows for a simple means to have the gains tuned in real time. As is known, data acquisition and delays caused by purely digital computer hardware allows only the control of vibrational and acoustic frequencies in the order of less than 1000 cycles per second.

Figure 7 is a schematic diagram of the Hybrid analog/digital proportional- integral controller 150 of the current invention. It includes a digital control analyzer logic circuit 152 that receives the voltage signal from the microphone 124 within the SHR 118. The digital logic circuit 152 estimates the disturbance frequency, determines the required actuation frequency, and adjusts the digital potentiometers (154, 156, 158) located within the analog PID controller 136. The K_P, K_{I t} and K_D K_p, K_j, and K_d gains for the PID controller 136 are changed via a digital serial communications by the digital logic circuit 152. The combination of the digital logic circuit 152 with the analog PID controller 136 allows continuous time transfer function between the PID input and output, and digitally controlled controller gains. The analog portion 136 of the controller allows the system to function at frequencies that are far above the capabilities of state of the art digital signal processors, without introducing sampling delays into the system. Once the potentiometers (154, 156, 158 ) are set by the digital logic circuit 152, the controller is an analog device with no sampling delay introduced. The primary benefit of this system is that it allows for adaptive control of high frequency vibrations without introducing sampling delay, and the cost of the components is dramatically less than high speed digital signal processors.

The slow time scale tuning component of the control algorithm is used to change the gains K_P, K_{! t} and K_D; see Figure 4. The signal is used to compute the dominant frequency 141 of the disturbance force. This can be implemented by various means, i.e., compute the frequency of the peak response of a Fourier transform of the signal: a frequency counter using zero crossings, adaptive frequency estimation, etc. These techniques can be used because it is assumed that the disturbance force frequency changes slowly over time. Therefore, any technique that estimates the frequency faster than the disturbance changes in frequency can be used. The frequency estimation algorithm 139 generates a signal 140 that is a scalar value representing a circular frequency or the disturbance force. This signal is then used to set the gains K_p, K„ and K_d.

The gains K_p, K„ K_d, are chosen such that the overall dynamic impedance of the device 132 minimizes the acoustic transmission. For simple passive resonators, an analytical solution can be found which maps the gains to resonant frequency and damping ratio while ensuring stability. However as the passive resonator becomes more complex and the sensor and actuator dynamics become significant, the analytical solution diverges from experimental results. Therefore, an empirical gain scheduling technique can be used instead, which finds a set of gains for each disturbance frequency that optimizes the amplitude and therefore the acoustic noise reduction. This can be implemented as a look-up table in a gain scheduling controller which takes the disturbance frequency estimate and output gains K_p, K„ and /or K_d, implemented on the fast time scale component.

Figure 9 is a schematic representation of a closed-loop semi-active vibration resonator coupled to an vibrating body and accompanying control system. The dynamics of vibrating body and accompanying passive vibration damper are described by the model equations:

m — K- 2 V^2 C )- F_A (16)

(17) m x₃ = F_A where

F_A = K_A(x₂ -x ) (18)

E, which is the force component from the actuator, is generated as a function of the measured deflection of the compliant attachment with stiffness k₂. This measurement is available from the indicated sensor. The actuator would react against it's own mass, m₃ to produce the necessary force required to vary the effective mass of the passive resonator. In this way, the tuned frequency of the passive resonator could be controlled over the required range of frequency deviations, Δω about the passive resonator tuned frequency, ω .

It is envisioned that a passive absorber with accompanying sensor and actuator would be coupled to the body having the vibrations to be dampened. The controller can be of any suitable type having it's gains modified based on the above equations so that the necessary force required to vary the effective mass of the passive resonator is produced.

Figure 10 is a representation of a piezo-electric actuator coupled to a compliant vibrating beam. The system above does not require any attachment points except the attachment to the primary system. It's sensor is incorporated into the compliant attachment. A variation of this device is the piezo-electric patch actuator/sensor device below. In this case, the beam functions as both compliant attachment to the primary system and passive absorber mass. The piezo patches function as both actuators and sensors. The passive vibration absorber generates absorber forces through flexural resonance of the beam. In another application, the DA/PID can be used in the chemical process industry. The chemical process industry includes many thermal and/or reaction processes where PID controllers are employed to control the rate of those processes. In many cases, optimal PID control gains change with the operating conditions of the process. Without the ability to adapt those gains, controllers are adjusted to conservative values at the sacrifice of process efficiency. The DA/PID controller can be employed to monitor process conditions such as external temperature and humidity along with the control input, e, and set the PID controller gains to optimal values based on a stored tables of PID gain values.

In another application, the DA/PID can be used in mechanical servo controls. Servo controlled mechanisms PID controllers are common. One example is factory production machinery. Fixed gain PID servo controls compromise between servo performance because their gains are set to allow acceptable servo performance over a wide range of operating conditions while avoiding gains which might lead to objectionable performance under specific conditions. In many cases, optimal PID control gains change with the operating conditions of the process. Without the ability to adapt those gains, controllers are adjusted to conservative values at the sacrifice of efficiency. The DA/PID controller can be employed to monitor process conditions such as external temperature and humidity along with the control input, e, and set the PID controller gains to optimal values based on a stored tables of PID gain values.

While the foregoing embodiments have all been described with respect to a semi-active vibration damper, which controls the response of an actuator by monitoring the actuator's output within a system, the gains of the controller may be adapted using auxiliary inputs, feed forward inputs, as well as feedback inputs. Additionally, those skilled in the art upon reviewing disclosure herein will see the applicability of the teaching within other technological areas. It should be appreciated that the Hybrid Digital-Analog Controller disclosed herein may replace those all digital systems that are commercially available. Also, an analog controller disclosed herein may be controlled by a standard computer as taught, for example, in U.S. Patent No. 5,355,305 to Seem et al., or as found in commercially available computer systems having controllable digital output capabilities.

A number of other possible uses and modifications have already been disclosed above. Further changes are clearly possible, as different features and aspects of one embodiment may be combined with other embodiment to provide a controller with the desired features from both. Thus, it to be understood that the present invention is by no means limited to the particular construction herein disclosed and/or shown in the drawings. Instead, the present invention also encompasses many modifications or equivalents within the scope of the disclosures that are fairly covered by the claims set forth below.

Claims

The Invention Claimed Is:

1. A device for reducing vibrations in a member, the device comprising: a passive vibration resonator , the resonator being coupled to the member; an actuator operable to apply forces to the resonator; a sensor operatively coupled to the member serving to detect changes in displacement between the mass and the member, and producing a first signal indicative of the change in displacement between the mass and the member; and a controller operatively coupled to the sensor and the actuator, the controller operatively receiving the first signal and in response to the first signal, the controller generates a second signal to drive the actuator to change the resonant frequency of the passive vibration resonator, thereby changing the vibrational response of the member.

2. The device of claim 1 wherein the passive vibrational resonator is a piezo-electric element.

3. The device of claim 2 wherein the controller contains an analog PID controller.

4. The device of claim 3 wherein the controller further includes at least one digitally variable resistor.

5. The device for changing the vibrational response of a member of claim

4 wherein the controller further contains an adaptive gain scheduler.

6. The device of claim 3 wherein the actuator is a piezoelectric element.

7. The device of claim 2 wherein the controller further contains an adaptive gain scheduler.

8. The device of claim 1 wherein the passive vibrational resonator is a spring mass resonator.

9. The device of claim 8 wherein the controller further contains an adaptive gain scheduler.

10. The device of claim 8 wherein the controller is an analog PID controller.

11. The device of claim 10 wherein the controller further includes at least one digitally variable resistor.

12. The device of claim 10 wherein the controller further contains an adaptive gain scheduler.

13. The device of claim 1 wherein the controller includes a controller utilizing a digital computer.

14. The device of claim 13 wherein the controller includes an electronic PID controller.

15. The device of claim 14 wherein the PID controller further includes at least one digitally variable resistor.

16. The device of claim 15 wherein the actuator contains an electromechanical component.

17. A device for reducing acoustic response of a partially enclosed defined volume, the device comprising: a passive acoustic resonator having a cavity, the cavity being acoustically coupled to the defined volume and further having a passive acoustic resonant frequency; an actuator creating changes in pressure within the cavity; a sensor operatively coupled to the cavity detecting changes in pressure within the cavity and producing a first signal indicative of the change in pressure within the cavity; a digital control analyzer operatively coupled to the sensor, the digital control analyzer being capable of receiving the first signal, the digital control analyzer capable of generating a second gain control; and an analog controller having at least one digitally variable resistor, the controller being operatively coupled to the digital control analyzer and the actuator, the analog controller operatively receiving the second gain control signal and in response to the second gain control signal adjusting the digitally variable resistor, the controller operatively generating a third signal to drive the actuator to change the acoustic impedance of the passive acoustic resonator, thereby changing the acoustic response of the defined volume.

18. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the digital control analyzer is part of a digital computer.

19. The device for changing the acoustic response within a partially enclosed defined area of claim 18 wherein the controller is an analog PID controller.

20. The device for changing the acoustic response within a partially enclosed defined area of claim 19 wherein the analog PID controller has adjustable gains K_p, K_j, and K_d, the digital control analyzer changes the gains in real-time to modify the acoustic response within the partially enclosed defined area.

21. The device for changing the acoustic response within a partially enclosed defined area of claim 20 wherein the actuator is a piezoelectric element.

22. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the controller is an analog PID controller.

23. The device for changing the acoustic response within a partially enclosed defined area of claim 22 wherein the analog PID controller has adjustable gains K_p, K„ and K_d, the digital control analyzer changes the gains in real-time to modify the acoustic response within the partially enclosed defined area.

24. The device for changing the acoustic response within a partially enclosed defined area of claim 23 wherein the digital control analyzer is a micro controller.

25. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the controller has adjustable gains, the digital control analyzer changing the gains in real-time to modify the acoustic response within the partially enclosed defined area.

26. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the actuator is an electromechanical speaker.

27. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the digital control analyzer further contains an adaptive gain scheduler.

28. The device for changing the acoustic response within a partially enclosed defined area of claim 17 wherein the digital control analyzer further contains an adaptive gain scheduler.

29. The device for changing the acoustic response within a partially enclosed defined area of claim 22 wherein the digital control analyzer further contains an adaptive gain scheduler.

30. The device for changing the acoustic response with a partially enclosed defined area of claim 25 wherein the digital control analyzer further contains an adaptive gain scheduler.

31. The device of claim 17 wherein the passive acoustic resonator is a Helmholtz resonator.

32. The device for changing acoustic response within a partially enclosed defined area of claim 17 wherein the passive acoustic resonator has a length selected from the group of: (a) 1/4 times the wave length of the passive acoustic resonant frequency, (b) 3/4 times the wave length of the passive acoustic resonant frequency, and (c) 5/4 times the wave length of the passive acoustic resonant frequency.

33. A hybrid controller for dynamically adjusting the control parameters of an analog controller having a controllable output signal, the apparatus comprising: a sampling circuit to receive a first signal representative of a controlled variable of the process; a plurality of digitally variable resistors, coupled to said analog controller for changing gains in the analog controller, based on the sampled signal; and the analog controller, responsive to changes to the digitally variable resistors to affect the output signal.

34. The hybrid controller of claim 33 wherein the circuit for generating a resistor control signal includes an adaptive gain scheduler.

35. The hybrid controller of claim 33 wherein the analog controller is a PID controller.

36. The hybrid controller of claim 33 wherein the circuit for generating a digital resistor control signal comprises a microprocessor.

37. A method of dynamically adjusting the control parameters of an analog controller having a plurality of digitally variable resistors, the method comprising: (a) generating a second control signal operable to adjust the values of the digitally variable resistors;

(b) sampling the first signal;

(c) generating the second signal based on the sampled first signal;

(d) adjusting the values of the digitally variable resistor; and (e) controlling the output signal.

38. The method of claim 37 wherein changing the resistance of the resistor changes the gains of the analog controller.

39. The method of claim 37 wherein sampling a first signal includes sending an input to the analog controller.

40. The method of claim 37 wherein sampling a first signal is sampling an output of the analog controller.

41. The method of claim 37 wherein sampling a first signal is sampling an auxiliary signal.