Derivative mode filter

The PID form implemented usually includes a derivative mode filter such as a first-order filter to eliminate noise, which would be written in the time domain as Eqs. (90), where ep(t) is the filtered error and 0.05 < a < 0.2 is the dimensionless filter constant [7]. 

The PID controller may also use the rate of change of the measured variable (for example, h(t)) instead of the error e(t) to eliminate set point change derivative kick. Derivative kick is mitigated if a derivative mode filter is used. 

Figure 12.24 shows the dynamic response of P, PI, and PID controller types to a step-change in the input of the first-order plus dead time (FOPDT) process of Figure 12.22 with parameters Kp = 10 min m , r = 20 min, 0 = 2 min. For the FOPDT example the tuning for the P controller is Kc = 0.595 min m, for the PI controller it is Kc = 10 min m , r/ = 19.7 min, and for the PID controller it is Kp — 0.691 min m , t/ = 25.6 min, tp — 0.725 min. The derivative mode filter was used for the PID controller with a filter constant of a = 0.1. The control loop was simulated numerically for Figure 12.24. It can be seen that the P controller produces a long-term offset, which the PI controller eliminates, but with some overshoot of the set point. The addition of the derivative action for the PID controller eliminates the overshoot and produces the best controller performance.

Another common need is to increase resolution, and sometimes spectra are routinely displayed in the derivative mode (e.g. electron spin resonance spectroscopy) there are a number of rapid computational methods for such calculations that do not emphasize noise too much (Section 3.3.2). Other approaches based on curve fitting and Fourier filters are also very common. 

The Derivative mode is sometimes referred to as rate because it applies control action proportional to the rate of change of its input. Most controllers use the process measurement, rather than the error, for this input in order to prevent an exaggerated response to step changes in the setpoint. Also, noise in the process measurement is attenuated by an inherent filter on the Derivative term, which has a time constant 1/8 to 1/10 of the Derivative time. Even with these considerations, process noise is a major deterrent to the use of Derivative mode. 

Fig. 17. 2-D spectral-spatial ESRI contour (top) and perspective (bottom) plots of HAS-derived nitroxides in ABS2H after 70 h (a) and 643 h (b) of irradiation by the Xe arc in the weathering chamber, presented in absorption. The spectral slices a, b, c, and d for the indicated depths are presented in the derivative mode these slices were obtained from digital (nondestructive) sections of the 2-D image. %F is shown for a, b, c, and d slices in (a) and for a, c, and d slices in (b). Both 2-D images were reconstructed from 83 real projections, Hamming filter, 2 PSA iterations, L = 4.5 mm, AH = 70 G, and were plotted on a 256 X 256 grid.

If measurement noise combined with a large ratio of derivative time to sampling period (Tjy/At) causes an overactive derivative mode, then the error signal must be filtered before calculating the derivative action (see Chapter 17). 

This expression describes the fastest and most important mode of transport in groundwater. In fact, an important task of the hydrologist is to develop models to predict the effective velocity u (or the specific flow rate q). Like the Darcy-Weis-bach equation for rivers (Eq. 24-4), for this purpose there is an important equation for groundwater flow, Darcy s Law. In its original version, formulated by Darcy in 1856, the equation describes the one-dimensional flow through a vertical filter column. The characteristic properties of the column (i.e., of the aquifer) are described by the so-called hydraulic conductivity, Kq (units m s"1). Based on Darcy s Law, Dupuit derived an approximate equation for quasi-horizontal flow ... 

To investigate the relationship between the reaction driven v7 mode and the subsequent protein motions along the dissociative pathway, further modulations of the frequency of the v7 mode by the surrounding intramolecular and protein bath fluctuations were found using an instantaneous frequency (IF) analysis. The IF was derived from the data by applying a Gaussian filter around the v7 mode in the Fourier spectrum. An inverse Fourier transform produced the time trace TT(t) given by ... 

Formally, the sum of random electromagnetic-field fluctuations in any set of bodies can be Fourier (frequency) decomposed into a sum of oscillatory modes extending through space. The "shaky step" in this derivation, already mentioned, is that we treat the modes extending over dissipative media as though they were pure sinusoidal oscillations. Implicitly this treatment filters all the fluctuations and dissipations to imagine pure oscillations only then does the derivation transform these oscillations into the smoothed, exponentially decaying disturbances of random fluctuation. 

Reactions were also conducted in a stirred slurry reactor operating in continuous mode with the solution injected by means of a pump and the catalyst retained by a filter at the liquid outlet [19,21]. A fixed-bed reactor was used by Kimura [71,72,74] for the oxidation of glycerol and derivatives. 

An important hybrid approach has also developed in recent years that makes use of block-structured or modular models. These models are composed of parametric and/or nonparametric components properly connected to represent reliably the input-output relation. The model specification task for this class of models is more demanding and may utilize previous parametric and/or nonparametric modeling results. A promising variant of this approach, which derives from the general Volterra-Wiener formulation, employs principal dynamic modes as a canonical set of filters to represent a broad class of nonlinear dynamic systems. Another variant of the modular approach that has recently acquired considerable popularity but will not be covered in this review is the use of artificial neural networks to represent input-output nonlinear mappings in the form of connectionist models. These connectionist models are often fully parametrized, making this approach affine to parametric modeling, as well. 

---