Performance evaluation of decentralized wireless sensing and control in civil structures

Transcription

1 Performance evaluation of ecentralize wireless sensing an control in civil structures Yang Wang *a, R. Anrew Swartz b, Jerome P. Lynch b, Kincho H. Law a, Chin-Hsiung Loh c a Dept. of Civil an Environmental Engineering, Stanfor Univ., Stanfor, CA 9435 b Dept. of Civil an Environmental Engineering, Univ. of Michigan, Ann Arbor, MI 499 c Dept. of Civil Engineering, National Taiwan Univ., Taipei, Taiwan, 6 R.O.C. ABSTRACT A structural control system consists of sensors, controllers, an actuators integrate in a single network to effectively mitigate builing vibration uring external excitations. The costs associate with high-capacity actuators an system installation are factors impeing the wie sprea aoption of structural control technology. Wireless communication can potentially lower installation costs by eliminating coaxial cables an offer better flexibility an aaptability in the esign of a structural control system. This paper introuces a prototype wireless sensing an control unit that can be incorporate in a real-time structural control system. Tests are conucte using a 3-story half-scale laboratory structure instrumente with magnetorheological ampers to valiate the feasibility of the wireless structural control system. This paper also aresses the serious issue of time elay an communication range inherent to wireless technologies. Numerical simulations using ifferent ecentralize structural control strategies are conucte on a -story steel structure controlle by semi-active hyraulic ampers. Keywors: structural control, wireless communication, embee computing, ecentralize output feeback control.. INTRODUCTION For the past three ecaes, significant avances have been mae in structural control []. Harware technologies an control theories are now wiely available for exploration in civil structural applications. It is reporte that between 989 an 3, about 5 builings an towers have been instrumente with various types of structural control systems with evient reuction in structural ynamic responses being observe []. Structural control systems can be categorize into three major types: passive control (e.g. base isolation), active control (e.g. active mass ampers), an (c) semi-active control (e.g. semi-active variable ampers). Passive control systems utilize passive energy issipation evices to mitigate structural vibration. These types of control systems require minimum amount of operational power an complexity, but they lack aaptability to real-time external excitations. On the other han, active control systems are aaptable to external excitations by ajusting actuation forces through a realtime feeback control loop, but the active control evices are usually power hungry. The operation of these active control evices often epens on power lines in the builing, which may not be reliable uring extreme natural events. Semi-active control effectively combines the merits of active an passive control techniques. Semi-active control evices are power-efficient an their actuation forces are ajustable through feeback control. Examples of semi-active actuators inclue active variable stiffness (AVS) evices, semi-active hyraulic ampers (SHD), electrorheological (ER) ampers, an magnetorheological (MR) ampers. Similar to passive control systems, semi-active control systems are inherently stable because their actuators o not apply mechanical energy irectly to the structure. To facilitate feeback control in a semi-active control system, structural sensors are employe to measure the ynamic response of a structure in real-time. Sensor ata are fe into control moules to compute optimal actuation forces. The actuation forces are then applie to the structure through semi-active control evices to mitigate excessive ynamic responses. In traitional semi-active control systems, coaxial cables are installe to connect sensors, controllers, an actuators for feeback control. As the size of the structure increases, the cost of installing the wires also grows. Furthermore, once a cable control system is installe, reconfiguring the system woul require costly rerouting of the cables. To eliminate * wyang98@stanfor.eu; phone ; fax ;

2 the cost an inconvenience of cable installation, wireless communication an embee computing technologies can be viable alternatives in structural control applications. Wireless communication has been extensively explore for use in structural health monitoring [3-4]. However, only a few stuies have been reporte on incorporating wireless communication in real-time feeback structural control [5-6]. In these stuies, sensor ata is transmitte wirelessly to wireless control units where embee control algorithms etermine control actions to be applie to actuators. One important issue about real-time feeback structural control is the latencies for transmitting sensor ata an control commans wirelessly. In orer to achieve goo performance, the communication latency shoul be kept as low as possible. In centralize control, regarless of wire or wireless communication, the control server has to collect ata from all the sensors in the structure. The communication latency increases with the size of the structure an the number of sensors being eploye. Moreover, the centralize control server represents a single-point of failure of the entire control system. To alleviate these issues, ecentralize control strategies can be explore [7]. In ecentralize control architectures, sensors an controllers are istributively place in a structure. Each controller collects ata from the nearby sensors an executes commans on the nearby structural actuators. Therefore, the communication ranges are significantly shorter an the communication latency ecreases since the number of sensors or actuators that each controller has to communicate with is significantly smaller. Decentralization also allows large quantities of lowercapacity actuators to be eploye in the structure, which coul be much cheaper to fabricate than actuators with very high force capacities. As the control ecisions are compute an execute istributively by iniviual controllers, system reunancy an reliability can also be improve. This stuy investigates the feasibility an issues in wireless sensing an ecentralize control. A prototype harware an software system is evelope for wireless structural sensing an control with both centralize an ecentralize control strategies embee in the system. Valiation of the wireless sensing an control system was conucte on a half-scale three-story steel structure instrumente with magnetorheological (MR) ampers. The experiments show that by reucing communication latency, ecentralize control strategies can achieve reliable control performance comparable to their centralize counterparts. To further investigate the feasibility of ecentralize structural control on large-scale structures, numerical simulations of a -story steel-frame structure have been carrie out. Preliminary results from these numerical simulations illustrate the effectiveness of ecentralize structural control strategies.. EXPERIMENTS ON DECENTRALIZED WIRELESS STRUCTURAL CONTROL In this section, a prototype wireless structural sensing an control system is first introuce. Centralize an ecentralize structural control algorithms with time-elaye output feebacks are then escribe. This section also presents the experimental setup an results for the valiation tests of the wireless prototype system on a half-scale threestory steel structure instrumente with magnetorheological (MR) ampers.. A prototype wireless structural sensing an control system Fig. illustrates a 3-story structure instrumente with the prototype wireless sensing an control system. The system consists of wireless sensors an controllers that are mounte on the structure for measuring structural response ata an commaning actuators in real-time. Besies the wireless sensing an control units that are necessary for ata collection an the operation of the actuators, a remote comman server with a wireless transceiver is inclue in the system for experimental purpose. In a laboratory experiment, the server is esigne to initiate the operation of the control system an to log the flow of wireless ata. To initiate the operation, the comman server first broacasts a start signal to all of the wireless sensing an control units. Once the start comman is receive, the wireless units that are responsible for collecting sensor ata start acquiring an broacasting ata at a preset time interval. Accoringly, the wireless units responsible for commaning the actuators receive the sensor ata, calculate esire control forces, an apply control commans within the specifie time interval assigne at each time step. The wireless unit is esigne in such a way that each unit can serve as either a sensing unit, or a control unit, or a unit for both sensing an control. This flexibility is supporte by an integrate harware esign base upon a wireless sensing unit (Fig. b) previously propose for wireless structural monitoring [8]. The three original functional moules inclue in the wireless sensing unit esign are the sensor signal igitizer, the computational core, an the wireless transceiver. To exten the functionality of the wireless sensor for actuation, an off-boar actuation signal generation moule (Fig. c) is esigne an fabricate. The actuation signal generation moule consists of a single-channel 6-bit igital-to-analog converter an other support electronics. The moule receives igital integers from a wireless sensing unit an converts the igital signal into an analog voltage ranging from -5V to 5V. This voltage signal then commans the associate

3 S 3 V 3 Floor-3 C i: Wireless control unit (with one wireless transceiver inclue) 3m C S i: Wireless sensing unit (with one wireless transceiver inclue) D V Floor- T i: Wireless transceiver D i: MR Damper 3m C V i: Velocity meter D V Floor- 3m T Lab experiment comman server D C Floor- (c) Fig.. Overview to the prototype wireless sensing an control system: a 3-story structure controlle by three actuators; package wireless sensing an control unit ( cm 3 ); (c) printe circuit boar of the actuation signal generation moule ( cm ). V structural actuator that applies actuation forces to the controlle structure. Detaile esign of the wireless sensing an control unit an the actuation signal generation moule has been escribe in Ref. [5].. Centralize an ecentralize control algorithms using time-elaye output feeback In this feasibility stuy, an output feeback control algorithm base on linear quaratic regulation (LQR) is employe. The algorithm can be briefly summarize as follows. For a lumpe-mass structural moel with n egrees-of-freeom (DOF) an m actuators, the system state-space equation consiering l time steps of elay can be formulate as: z [ k+ ] = A z [ k] + B p [ k l], where z [ k] Here z [ k] represents the n iscrete-time state-space vector, [ k l] [ k] [ k] x = x& () p is the elaye m control force vector, A is the n n system matrix (containing the information about structural mass, stiffness, an amping), an B is the n m actuator location matrix. For the experiments an simulations presente in this paper, we consier the case where l is equal to, i.e. the time elay is equal to one sampling time step. The primary objective of the time-elay LQR problem is to minimize a cost function J by selecting an optimal control force trajectory p : T T ( [ ] [ ] [ ] [ ]) J = z k Qz k + p k Rp k, where Q an R > () p k= l n n m m In an output feeback control esign, control ecisions are compute base on real-time measurement ata in the q system output vector y [ k]. The output vector is efine by a q n linear transformation, D, to the state-space z k : vector [ ] [ k] = [ k] y D z (3) For example, if the relative velocities on all floors (but not the relative isplacements) are measurable, D can be efine (by letting q = n) as: [ ] D = I (4) n n n n

4 In another example, if inter-story velocities between ajacent floors are measurable, the output matrix D can be efine (by letting q = n) as: L L D [ n n] = L M O O O M L The m q optimal gain matrix is esigne to provie the optimal output feeback control force: [ k] = [ k] n n p y (6) Chung et al. [9] propose the formulation to the above output feeback control problem consiering time elay (l time steps). As a result, a set of couple nonlinear matrix equations can be solve for an optimal output feeback gain matrix. In our implementation, an iterative algorithm put forth by Lunze [] is moifie to solve the matrix equations. The algorithm escribe by Lunze also provies the flexibility to hanle aitional external constraints. In particular, this algorithm can compute a suboptimal control solution for a ecentralize system simply by constraining the structure of to be consistent with the ecentralize architecture []. For example, the following equations illustrate the constraine structures of two ecentralize output feeback gain matrices for a 3-story lumpe-mass structure (n = 3): * * * *, * * = = (7) * * * Combining with the output matrix D efine in Eq. (4) or (5), the pattern in specifies that when computing control ecisions, the actuator at each floor only nees the entry in the output vector y that correspons to that floor. The pattern in specifies that the control ecisions also require information from a neighboring floor..3 Valiation experiments To stuy the potential application of the wireless sensing an control system for ecentralize structural control, valiation tests on a 3-story frame structure instrumente with MR ampers are conucte at the National Center for Research on Earthquake Engineering (NCREE) in Taipei, Taiwan (Fig. a). The floor plan of this structure is 3m m, with each floor weight ajuste to 6, kg using concrete blocks. The inter-story height is 3m. Both the beams an the columns of the structure are constructe with H5 5 7 steel I-beam elements. The three-story structure is mounte on a 5m 5m 6-DOF shake table. The shake table can generate groun excitations with frequencies spanning from.hz to 5Hz. For this stuy, only longituinal excitations are use. Along this irection, the shake table can excite the structure with a maximum acceleration of 9.8 m/s. The excitation has a maximum stroke an force of ±.5m an kn, respectively. The test structure is heavily instrumente with accelerometers, velocity meters, an linear variable isplacement transucers (LVDT) installe on each floor of the structure to measure the ynamic response. These sensors are interface to a high-precision wire-base ata acquisition (DAQ) system native to the NCREE facility; the DAQ system is set to a sampling rate of Hz. A separate set of wireless sensors are installe as part of the wireless control system. For the experimental stuy, three kn MR ampers are installe with V-braces on each story of the steel structure (Fig. b). The amping coefficients of the MR ampers can be change by issuing a comman voltage between V to.v. This comman voltage etermines the electric current of the electromagnetic coil in the MR amper, which in turn, generates a magnetic fiel that sets the viscous amping properties of the MR amper. Calibration tests were first conucte on the MR ampers before mounting them to the structure so that moifie Bouc-Wen amper moels can be formulate for each amper []. In the real-time feeback control tests, hysteresis moel parameters for the MR ampers are an integral element in the calculation of amper actuation voltages. Fig. (c) illustrates a wireless control unit an an off-boar control signal generation moule that work together to comman an MR amper. (5)

5 (c) Fig.. Laboratory setup: the 3-story test structure mounte on the shake table; the MR amper installe between the st floor an the base floor of the structure; (c) a wireless control unit an an off-boar control signal generation moule. Table. Different ecentralization patterns an sampling times for the wireless an wire-base control experiments. Degree of Centralization ain Matrix Pattern Output Matrix Sampling Time/Rate in Eq. (7) Wireless System in Eq. (7) D in Eq. (5).s / 5Hz D in Eq. (5).6s / 6.67Hz 3 N/A D in Eq. (4).8s /.5Hz Wire System 3 N/A D in Eq. (4).5s / Hz For the wireless system, a total of four wireless sensors are installe, following the eployment shown in Fig.. Each wireless sensor is interface to a Tokyo Sokushin VSE5-D velocity meter to measure the absolute velocity response for each floor of the structure, as well as the base. The sensitivity of this velocity meter is V/(m/s) with a measurement limit of ± m/s. The three wireless sensors on the first three levels of the structure (C, C, an C) are also responsible for commaning the MR ampers. Besies the wireless control system, a traitional wire-base control system is installe in the structure for comparative analyses. Centralize an ecentralize velocity feeback control schemes are use for the wire an the wireless control systems. Table summarizes the ifferent patterns of the gain matrix,, the output matrices, D, an the achievable sampling times for the centralize, partially ecentralize an fully ecentralize control strategies (which are enote by egrees of ecentralization, 3, an, respectively). For this test structure, the wire-base system can achieve a sampling rate of Hz, or a time step of.5 s. Mostly ecie by the communication latency of the 4XStream wireless transceivers, the wireless system can achieve a sample rate of.5 Hz (or a time step of.8 s) for the centralize control scheme. This sample rate is ue to each wireless sensor waiting in turn to communicate its ata to the network (about. s for each transmission). An avantage of the ecentralize architecture is that fewer communication steps are neee, thereby reucing the time for wireless communication. To ensure that appropriate control ecisions are compute by the wireless control units, one necessary conition is that the real-time velocity ata use by them are accurate. Rarely experiencing ata losses uring the experiments, our prototype wireless sensor network proves to be robust. In case ata loss happens, the wireless control unit is currently esigne to simply use a previous ata sample. For the experimental results presente herein, the groun excitation is the 94 El Centro NS earthquake recor (Imperial Valley Irrigation District Station) scale to a peak groun acceleration of m/s. To illustrate the performance of ifferent ecentralize schemes with various communication latencies, the same groun excitation is applie to the original uncontrolle structure with four ifferent wire an

6 3 Maximum Inter-story Drifts No Control Wireless # Wireless # Wireless #3 Wire 3 Maximum Absolute Accelerations No Control Wireless # Wireless # Wireless #3 Wire Story Floor Drift (m) Acceleration (m/s ) Fig. 3. Experimental results of ifferent control schemes using the El Centro excitation scale to a peak acceleration of m/s : peak inter-story rifts; peak absolute accelerations. wireless control schemes implemente, as efine in Table. Fig. 3 illustrates the structure s peak inter-story rifts an floor absolute accelerations uring these experimental runs. Compare with the uncontrolle structure, all wireless an wire control schemes achieve significant reuction with respect to maximum inter-story rifts an absolute accelerations. Among the four control cases, the wire centralize control scheme shows better performance in achieving the least peak rifts an secon least overall peak accelerations. This result is rather expecte, because the wire system has the avantages of lower communication latency an utilizes complete sensor ata from all floors. The wireless schemes, although running at longer sampling steps, achieve control performance comparable to the wire system. The fully ecentralize wireless control scheme (case #), results in uniform peak inter-story rifts an the least peak floor accelerations. The results illustrate that in the ecentralize wireless control cases, the higher sampling rate (from lower communication latency) can potentially compensate for the lack of sensor ata from faraway floors. 3. NUMERICAL SIMULATIONS INVESTIATIN THE PERFORMANCE OF DECENTRALIZED STRUCTURAL CONTROL With the ecentralize output feeback control algorithm valiate through experiments, further analysis of ecentralize control strategies is pursue via numerical simulations. A -story benchmark structure esigne for the Structural Engineers Association of California (SAC) project is selecte [3]. The structure represents a typical mi- to high-rise steel-frame builing in the Los Angeles, Southern California region. To simplify the analysis, the builing is moele as an in-plane lumpe-mass structure (Fig. 4). In the following, the ecentralize control architecture for this -story builing is first introuce. Then numerical simulation results for ecentralize control are presente, assuming that the structure is instrumente with ieal actuators that are capable of proucing any esire forces. This section inclues simulation results for the case when the -story structure is installe with semi-active hyraulic ampers [4]. 3. Decentralize control architecture In the numerical simulations, it is assume that both the inter-story isplacements an inter-story velocities between every two neighboring floors are measurable. Similarly, the system state-space equations are formulate such that the state-space vector contains inter-story isplacements an velocities, rather than the isplacements an velocities relative to the groun. The output matrix D is efine as a n n ientity matrix, so that the state-space vector is use for control feeback irectly. The simulations are conucte for ifferent egrees of centralization (DC), as illustrate in Fig. 4(c). The egrees of centralization reflect ifferent wireless network architectures, with each wireless channel representing one wireless subnet. The actuators covere within a subnet are allowe to access the wireless sensor ata within that subnet. For example, the case where DC= implies each wireless channel covers only five stories an a total of four wireless channels (subnets) are utilize. Of all the ifferent egrees of centralization, the case where DC= represents the lowest requirement to the wireless communication range an achieves lowest communication latency as a

7 smaller number of wireless sensors nee to broacast their ata in the subnet. Constraine by this ecentralize information structure, the gain matrix for the case where DC= has the following sparsity pattern: (,) (,5) (,) (,6) =, when DC ( 3,3) ( 3,7) = ( 4,4) ( 4,8) The left submatrix an the right submatrix correspon to the inter-story isplacement an the inter-story velocity components, respectively, of the output vector y ( i, j). Each submatrix is block-iagonal, with every block being a 5- by-5 square matrix. For the case where DC=, each wireless channel covers ten stories an a total of three wireless channels are utilize. Constraine by the overlapping information structure, the gain matrix for DC= has the following sparsity pattern: 4 (,) (,) (,5) (,6) (,) (,) (,3) (,5) (,6) (,7) =, when DC ( 3,) ( 3,3) ( 3,4) ( 3,6) ( 3,7) ( 3,8) = ( 4,3) ( 4,4) ( 4,7) ( 4,8) For the case when DC=3, the number of stories covere by each wireless subnet increases accoringly. That leas to fewer communication subnets an fewer zero blocks in the gain matrices. The case where DC=4 specifies that one 4 (8) (9) F F5 Seismic Mass F.6 x 6 kg F F9. x 6 kg F.7 x 6 kg Damping 5% Natural Damping Inter-story Stiffness F-F x 3 kn/m F6-F x 3 kn/m F-F x 3 kn/m F5-F7 9.3 x 3 kn/m F8-F x 3 kn/m F 7.7 x 3 kn/m 9 x 3.96m F F F5 Ch4 Ch3 F F5 Ch3 F F5 Ch F F5 F Ch F F F Ch Ch Ch F5 F5 F5 Ch F5 F5 Ch 5.49m DC = DC = DC = 3 DC = 4 (c) Fig. 4. Twenty-story SAC builing [3] for numerical simulations: builing elevation; moel parameters of the lumpe mass structure; (c) wireless subnet partitioning for ifferent egrees of centralization (DC).

8 wireless channel covers all twenty floors, which results in a centralize information structure. 3. Numerical simulations using ieal actuators To investigate the effectiveness of the ecentralize control schemes propose, we first assume the -story structure is instrumente with ieal actuators that prouce any esire horizontal force between every two neighboring floors. Various combinations of centralization egrees (DC =,,4) an sampling time steps ranging from.s to.s (at a resolution of.s) are simulate. Aitionally, three groun motion recors are use for each simulation: the 94 El Centro NS recor (Imperial Valley Irrigation District Station), the 995 Kobe NS recor (JMA Station), an the 999 Chi-Chi NS recor (TCU-76 Station), all scale to a peak groun acceleration (PA) of m/s. Two representative performance inices propose by Spencer et al. [3] are aopte: PI max () t () t i ti, = max Earthquakes max ˆ i ti, J LQR, an PI = max Earthquakes Jˆ LQR Here PI an PI are the performance inices corresponing to inter-story rifts an LQR cost inices, respectively. In Eq. (), i () t represents the inter-story rift between floor i (i =,, n) an its lower floor at time t, an max i ( t ) is ti, the maximum inter-story rift over the entire time history an among all floors. The maximum inter-story rift is normalize by its counterpart max ˆ i () t, which is the maximum response of the uncontrolle structure. The largest ti, normalize ratio among the simulations for the four ifferent earthquake recors is efine as the performance inex PI. Similarly, the performance inex PI is efine for the LQR control inex J LQR, as given in Eq. (). When computing the LQR inex over time, a uniform time step of.s is use to collect the structural response ata points, regarless of the sampling time step of the control scheme; this allows one control strategy to be compare to another without concern for the ifferent sampling time steps use in the control solution. Values of the two control performance inices are plotte in Fig. 5 for ifferent combinations of egrees of centralization an sampling time steps. The plots shown in Fig. 5 an 5 illustrate that the egrees of centralization an sampling steps have significant impact on the performance of the control system. enerally speaking, control performance is better for higher egrees of centralization an shorter sampling times. The plots show that except for the case where DC=, the control schemes with certain overlapping information structures achieve comparable performance. To better review the simulation results, the performance inices for the four ifferent control schemes are re-plotte as a function of sampling time in Fig. 5(c) an 5(). As shown in Fig. 5(c), if a partially ecentralize control system with DC equal to can achieve a sampling step of.s an a centralize system can only achieve.4s ue to aitional communication latency, the partially ecentralize system can result in lower maximum inter-story rifts. Similar trens are observe in Fig. 5(), except that the plots are smoother ue to the summation process for computing the LQR inices. 3.3 Numerical simulations using semi-active hyraulic ampers In general, structural actuators have certain capacity limits for actuation forces. Numerical simulations are conucte for ecentralize control where semi-active hyraulic ampers (SHD) are employe on the structure. The arrangement of SHD ampers is shown in Fig. 6; the number of instrumente SHD ampers ecreases graually from 4 to in the higher floors. If the esire amping force is in an opposite irection to the inter-story velocity, as shown in Fig. 6, the amping coefficient is ajuste so that a amper force closest to the esire force is generate. If the esire force is in the same irection to the inter-story velocity, the amping coefficient is set to its minimum value. Each SHD amper is installe as a V-brace unerneath the corresponing floor. To accurately moel the amping force, the Maxwell element propose by Hataa et al. [4] is employe. In a Maxwell element, a ashpot an a stiffness spring are connecte in series, which result in a amping force escribe by the following ifferential equation: () keff pt &() + pt () = keff xt &() () c () t SHD

9 Maximum Drift Among All Stories LQR Inex Performance Inex PI Samping Time(s) 4 3 Degree of Centralization Performance Inex PI Samping Time(s) 4 3 Degree of Centralization Performance Inex PI Maximum Drift Among All Stories Degr Degr Degr 3 Degr 4 Performance Inex PI Degr Degr Degr 3 Degr 4 LQR Inex Samping Time (s) Samping Time (s) (c) () Fig. 5. Simulation results illustrating performance inexes for ifferent sampling time steps an egrees of centralization (DC): 3D plot for performance inex PI ; 3D plot for performance inex PI ; (c) conense D plot for PI ; () conense D plot for PI. F ~ F5 F6 ~ F5 F ~ F5 & x( t) SHD p(t) F6 ~ F Maximum Control Force Maximum Displacement Stiffness of the SHD Maximum Damping Coefficient Minimum Damping Coefficient Maximum Shaft Velocity Power Consumption, kn +/- 6 cm 4, kn/m, kn-s/m, kn-s/m 5 cm/s 7 Watts F ~ F5 F6 ~ F5 F ~ F5 Fig. 6. Instrumentation of semi-active hyraulic ampers (SHD): layout of SHD ampers on the floor plans; key parameters of the SHD amper. where p(t) an & x() t enote the amping force an the inter-story velocity, respectively, k eff represents the effective stiffness of the amper in series with the V-brace, an c () t is the ajustable amping coefficient of the SHD amper. SHD

10 El Centro, PA=5m/s Kobe, PA=5m/s Chi-Chi, PA=5m/s Story Story Story No Control DC=, ms DC=4, 4ms Maximum Drift (m)..4.6 Maximum Drift (m) Maximum Drift (m) Fig. 7. Maximum inter-story rifts for cases where DC= with ms time elay an DC=4 with 4ms time elay. Fig. 7 presents the simulate maximum inter-story rifts when the structure is excite using the same three groun motions as escribe in section 3., except that the PAs are scale up to 5m/s, instea of m/s. To compare the performance of ecentralize versus centralize control, the case where DC= (partially ecentralize) an the case where DC=4 (centralize) are plotte. As shown in Fig. 4, each wireless subnet covers ten floors when DC=, or twenty floors when DC=4. That is, the inuce time elay when DC= is about half of the elay when DC=4, an the time elays of ms an 4ms are assigne, respectively, for these two cases. As shown in Fig. 7, both of the two control schemes significantly reuce the maximum inter-story rifts compare with the uncontrolle case. For the Kobe an Chi-Chi groun motions, the partially ecentralize case where DC= achieves equivalent performance compare with the centralize case where DC=4, while for the El Centro recor, the case with DC= achieves slightly better performance than the case where DC=4. These results illustrate that although ecentralize control has the isavantage of computing control ecisions without complete sensor ata, the lower time elay in ecentralize control coul make the control scheme perform as well as centralize control, given that the centralize case requires longer communication latencies. 4. CONCLUSIONS In this stuy, a prototype wireless structural sensing an control system is evelope, an its performance valiate in real-time feeback control experiments using a 3-story steel structure instrumente with MR ampers. Further simulation analyses are conucte using a -story builing structure instrumente with ieal actuators an SHD ampers by varying the egrees of centralization an the sampling time steps of the control system. Both experimental an simulation results illustrate that ecentralize control strategies may reuce system cost an provie equivalent or even superior control performance, given that their centralize counterparts coul require longer sampling steps ue to communication latencies in a wireless network. 5. ACKNOWLEDMENTS This research is partially fune by the National Science Founation uner grant CMS an the Office of Naval Research Young Investigator Program aware to Prof. Lynch at the University of Michigan. Aitional support is provie by National Science Council in Taiwan uner rant No. NSC Z--3. The authors wish to thank

11 two grauate stuent fellowship programs: the Office of Technology Licensing Stanfor rauate Fellowship at Stanfor University an the Rackham rant an Fellowship Program at the University of Michigan. The authors woul also like to express their gratitue to Dr. Pei-Yang Lin an Mr. Kung-Chun Lu at National Taiwan University, for their generous support for conucting the shake table experiments at NCREE, Taiwan. REFERENCES. T. T. Soong an B. F. Spencer Jr., Supplemental Energy Dissipation: State-of-the-art an State-of-the-practice, Eng. Struct. 4(3), 43-59,.. S. Y. Chu, T. T. Soong, an A. M. Reinhorn, Active, Hybri an Semi-active Structural Control, John Wiley & Sons Lt., E.. Straser an A. S. Kiremijian, A Moular, Wireless Damage Monitoring System for Structures, Report No. 8, John A. Blume Earthquake Eng. Ctr., Stanfor University, Stanfor, CA, USA, J. P. Lynch an K. J. Loh, A Summary Review of Wireless Sensors an Sensor Networks for Structural Health Monitoring, Shock Vib. Dig., Sage Publications, 38(), 9-8, Y. Wang, A. Swartz, J. P. Lynch, K. H. Law, K.-C. Lu, an C.-H. Loh, Wireless Feeback Structural Control with Embee Computing, Proc. of the SPIE th Intl. Symposium on Nonestructive Evaluation for Health Monitoring an Diagnostics, San Diego, CA, USA, Feb 6 - Mar, J. P. Lynch an D. Tilbury, Implementation of a Decentralize Control Algorithm Embee within a Wireless Active Sensor, Proc. of the n Annual ANCRiSST Workshop, yeongju, Korea, Jul -4, N. Sanell, Jr., P. Varaiya, M. Athans, an M. Safonov, Survey of Decentralize Control Methos for Large Scale Systems, IEEE Transactions on Automatic Control, 3(), 8-8, Y. Wang, J. P. Lynch, an K. H. Law, A Wireless Structural Health Monitoring System with Multithreae Sensing Devices: Design an Valiation, Structure an Infrastructure Engineering - Maintenance, Management an Life-Cycle Design & Performance, 3(), 3-, L. L. Chung, C. C. Lin, an K. H. Lu, Time-elay Control of Structures, Earthquake Eng. Struct. Dyn., John Wiley & Sons, 4(5), 687-7, J. Lunze, Feeback Control of Large-scale Systems, Prentice Hall, Hertforshire, UK, 99.. Y. Wang, R. A. Swartz, J. P. Lynch, K. H. Law, K.-C. Lu, an C.-H Loh, Decentralize Civil Structural Control using a Real-time Wireless Sensing an Control System, Proc. of the 4th Worl Conf. on Structural Control an Monitoring (4WCSCM), San Diego, CA, USA, Jul - 3, 6.. P.-Y. Lin, P. N. Roschke, an C.-H. Loh, System Ientification an Real Application of a Smart Magneto- Rheological Damper, Proc. of the 5 Intl. Symposium on Intelligent Control, Limassol, Cyprus, Jun 7-9, B. F. Spencer Jr., R. E. Christenson, an S. J. Dyke, Next eneration of Benchmark Structural Control Problems for Seismically Excite Builings. Prof. of the n Worl Conference on Structural Control, Kyoto, Japan, Jun 8 - Jul, T. Hataa, T. Kobori, M. Ishia, an N. Niwa, Dynamic Analysis of Structures with Maxwell Moel, Earthquake Eng. Struct. Dyn., John Wiley & Sons, 9(), 59-76,.