Trus joist's plant systems were as old as the Georgia hills surroundmg it. By layering in new Wonderware software, this leading provider of engineered lumber products got something they never had befare: real-time manufacturing intelligence. The resutt-the plant chopped engineering costs by 50%. Today, plant performance-and customer satisfaction-is up. Way up. And ptant downtime is way down.
Trus joist's plant systems were as old as the Georgia hills surroundmg it. By layering in new Wonderware software, this leading provider of engineered lumber products got something they never had befare: real-time manufacturing intelligence. The resutt-the plant chopped engineering costs by 50%. Today, plant performance-and customer satisfaction-is up. Way up. And ptant downtime is way down.
Trus joist's plant systems were as old as the Georgia hills surroundmg it. By layering in new Wonderware software, this leading provider of engineered lumber products got something they never had befare: real-time manufacturing intelligence. The resutt-the plant chopped engineering costs by 50%. Today, plant performance-and customer satisfaction-is up. Way up. And ptant downtime is way down.
ELECTRONICA IT ORIZABA ELECTRONICA OTF 9 852 BOX 94300 ORIZABA VERA CRUZ 94300 MEXICO Powering inteWgent plant dedsions in real time. Trus JOlst's Colbert plant systems were as old as the Georgia hills surroundmg it. By layering in sorne new Wanderware software, this leading provider of engineered lumber products got something they never had befare: real-time manufacturing intelligence. '7: The faundation of their new Wonderware solution, Industrial Application Server- built on the ArchestrA software architecture - enabled plant engineers to standardize _ applications to cut development time by 60%, and easily integrate a mixture of legacy plant systems, while preserving existlng technology investments. The resutt-the plant chopped engineering costs by 50%. Wonderware vsualization. historian, Web reporting, trending and downtime analysis now deliver the right information, to the right person, in the rght context, at the right time. This enables Trus Joist to identify problems they never knew they had, and make intelligent decisions to correct Invensys them. Today, plant performance-and customer satisfaction- is up. Way up. And ptant downtime is way down. To share your Wonderware application success story with the 'orld or to e.e. otner 'iUCCe.'i'ie.'i, 'li'iit wo n d e rwa re. co mIs u ccess weight's motion in eilher direction. Tlle spring force is proporlional to lhe weighl's position Xp",<c J. lhe frictionaJ rce is proportionaJ to th weight's velocity (Xp,acc), and lh force of grc vily is proportional lo the weiocrht's ac eleralion IX J. Combinin o cr these .. with the displacement f lhe handle (X eonl ,,) yields the governing equation shown in Figure 5. This is a econd arder relationship because it conlains both a ingle differentialion operator and a dOllble differ ntiation operalor. Uulike lhe firsl order process in th previotls example, this second arder process wiU sorne times oscillate ev n if the conlTol effort does not. Specifically, if the friclion is low enough, the spring is sliff enollgh, and the weigbl is heavy Figure 3 First order brewing process step response -lIT J T - -- ---- - - --=.-:.;- -.. -;.;--.--.....--j The brewing process will respond with an expo nentially decaying rise in the process variable (T",oee.. ) when the control effort (Teonrrol ) changes abruptly from zero to T at time zero. enough, the process will be underdamped and a imple lug wiJI cau lbe weighl to bounce as sbown in Figure 6. Conversely, an overdamped second order process wiU respond without oscilo lations as shown in Figure 7. Other processes The relatiollships shown in Figures 3, 6, and 7 are nol unique to brewing beer amI bouncing springs. They appl y to all first and sec ud order plocesses-just about everythiug lhat can be controlled with a lraditional PID (proportionaJ integral-derivativelloop. QnJy lhe values oI 'i, and (J.l" change from one process to the next. Al! fust order step responses look Ws:e Figure 3, and all second order step responses lo k like either Fi re 6 or Figure 7. Only the scaJe changes. In the first order case, the time constant 'i determines the duralion of the step response by defining exactly when the process variable will reach 63% of its final vaJue (which is to say, , = when t = 'i). The smaller the T ploe v lue of 'i, the sooner that 63% point will be reached, and conversely. Similarly, tbe values of the damping ratio s and the natural frequency w" completely del r mine tbe duration al1d amplitude of an under damped second ord r step response. As shown in Figure 6, lhe inverse of w serves as the time n constant for the decaying exponential term e '''"T. The product w" (that is, aw"l serves as the frequency oE the inusoidal term sin(aw" t + <\. The phase of that sinusoid is <1' and its amplitude is l/a, both of which are deter mined by lhe magnitude of Feedback control These relationships are particularly conve nient for feedback controller designo They al10w a controller lo predict lhe step response of any first order process with a known time constant or any second order process wilh a known damping ralio and natural frequency. Th controller can thcn be configured to apply just the right series of steps required lo drive Figure 4 the process variable towards the desired set Weight-ona point along a desired tra spring process jectory. Just how the values of The object of this 'i, and w" can be trans child's toy is to X control 1 laled into appropriate bounce the weight conlr ller parameters, up and down. The position of the weight such as P, l, and D, is often a matter of signifi (X",oce.. ) can be con cant technical complexi tm1led by manipulat ly. Fortunately, there are ing the position of Xprocess 1 formulas available for the handle (Xco"tro). many of the simplest case. See, for e ample, Figure 5 Governing equation the famous Ziegler , Nichols tuning rules for P"'" 1'" =>1l Xl:olllml first order proc sses in "L op Tuning Fundamen This second order relationship governs how lals," Control Engineering, the position of the weight (X ) wil/ process .luly 2003, at www.con to the position of the handle (Xeonlro)' Xproeess is troleng.com. the time derivative of Xprocess and X is its proeess Determining the val second derivative. The fixed values of the ues of 'i, s, and t" that damping ratio Sand the natural frequency w" b st represent th behav depend on the physical properties of the ior of the process in ques process - the mass of the weight. the viscosity tion can also be a chal of the frictional force, and the strength of the lenge. First principies spring. analysis uses lhe laws of www.controleng.com CONTROL ENGINEERING DECEMBER 2004 35