"A combined PXI and CompactRIO platform
performs real-time data acquisition, signal processing, control algorithms, and interfacing with the control panel and the electronic thrusters."
- Miguel Taboada, Seaplace S.L
The Challenge:
Controlling and keeping the position of a 1,000 m3 split hopper vessel (2,500 t displacement) during the dumping operation within a 10 m tolerance, withstanding a force six wind (25 kN), a long crested sea up to 3 m high, and sea current up to 0.5 kN.
The Solution:
Using a combined PXI and NI CompactRIO platform running on NI LabVIEW software to perform real-time data acquisition, signal processing, and control algorithms, as well as to interface with the control panel and electronic thrusters.
Author(s):
Miguel Taboada - Seaplace S.L
A split hopper vessel is a specialized ship with two articulated semihulls as the main structure. The semihulls form a hopper using a hydraulic system to open and close it. This ship is used in maritime works as a foundation for dam construction to transport and dump materials.la
During the dumping operation, the ship undergoes a two-thirds decrease in displacement and a double increase in wind-exposed area. This causes a rapid variation of wind, currents, and wave forces, leading to an undesired offset. This could lead to a very expensive material replacement during dumping operations.
SEADP is a dynamic positioning (DP) system specially adjusted to operate the split hopper vessels and control the position and heading of the ship throughout the dumping process. It approaches the target and controls and keeps the position and orientation while dumping.
System Description
We acquire data using a GPS, gyro, anemometer, speed log, and draught sensor in the PXI platform with NMEA protocol RS232 and RS422 serial interfaces. The sampling rate ranges from 10 s of the speed log to 10 Hz of the differential global positioning system (DGPS) real-time kinematic (RTK). The PXI hardware also samples the inertial motion unit (IMU) at 75 Hz through an RS232 serial interface. The ship’s thruster system consists of three thrust devices, capable of giving variable force in any planar direction, and a 200 kN overall thrust at 2,100 kW power.
Our solution features two CompactRIO systems working synchronously (interrupt method) with an NI PXI real-time controller at 10 Hz. One performs digital logic interface with SCHOTTEL Transverse Thrusters, revolution and azimuth thrust direction signal acquisition, command signals, and throttles to the diesel engine and thrust azimuth direction.
A client/server TCP architecture carries out the signal monitoring and user input using a touch panel monitor. With LabVIEW, we can run our system in record time.
The control panel contains all the indicators and buttons that make the system operate even during a communication breakdown between the control computer and the monitoring and supervision computer. In addition, operators can use a joystick to manually control the ship’s position and heading. The 3D positioning of the antenna location from GPS and DGPS is transferred to the hopper center to account for roll, pitch, and yaw angles. The ship’s position is given in Universal Transverse Mercator (UTM) coordinates. The anemometer measures the relative speed and wind direction, which feeds an aerodynamic drag force model to estimate the mean wind forces after gust removal.
The IMU records rotation rates and acceleration in body axes at the ship’s center of gravity. Kalman filters calculate the roll and pitch angles while the gyro obtains the yaw angle, or true heading. We use acceleration time series frequency domain processing for HF surge, sway, and yaw movement extraction. We also use roll angle in a recursive least squares (RLS) estimation for the period of wave-induced ship movements.
We use azimuth angle and propeller revolution sensors in thrust devices to estimate the total force and momentum acting upon the ship.
The taking draughts procedure is the ship model’s SEADP recursive estimation starting point contained in SEADP. We use a detailed ship model in a nonlinear state observer and in the LQR controller.
State Observer and Controller
Ship movements in waves, wind, and current are considered to be the superposition of wave frequency movements (0.05 to 0.2 Hz) and low frequency movements caused by wave grouping forces. In practice, it is impossible to counteract the wave movements as the ship weight magnitude force produces them. Due to the fact that these frequencies fall into the thrust device bandwidth, we must perform adequate wave movement filtering to avoid excessive wear and tear.
We achieve wave filtering using a nonlinear state observer that consists of an LF model of the ship motion, thruster system response, and stochastic models of wave motions and environmental disturbances working together to give smooth LF motion and ship velocity estimations.
The controller uses the ship’s LF motion, velocity, and target position deviation to calculate the commanded forces and moments using a linear quadratic regulator (LQR). The feedback controller minimizes a weighted integral of the deviation and power/thrust effort. Furthermore, wind forces estimated in real time are feed-forwarded to increase the controller performance. The controller performs manual/automatic control in surge, sway, and yaw movements, and commands forces in longitudinal and transversal directions. We have to allocate the moment around the gyration point among the different thrust devices. We accomplish this in SEADP using online quadratic programming techniques.
Future SEADP Developments
We are applying the technology used for split hopper vessels to the DP-1 system for an accommodation barge, which acts as a floating hotel moored near offshore platforms. Similar to the previous architecture, the NI PXI real-time controller acquires environmental data from the surroundings and we use CompactRIO to synchronously control the ship's movements. To distribute control in this time-critical system, we added the NI 9144 CompactRIO expansion chassis to communicate synchronously with the PXI controller via real-time Ethernet. The high determinism of remote I/O and tight integration with LabVIEW Real-Time were the main motivations for adding the NI 9144 chassis to the system. This new distributed control architecture increases the system reliability and modularity while reducing the overall cost through standard Ethernet cabling.
SEADP meets Class I DP requirements and any thrust device or controller failure could lead to position loss. Higher Class II and Class III requirements involve physical and logical redundancy and special application features such as online single-failure consequence analysis. We rely on NI hardware and software to overcome these future challenges.
Author Information:
Miguel Taboada
Seaplace S.L
Tel: 91 458 5119
mtaboada@seaplace.es
Print
PDF
