- \n
- Intake Gate<\/li>\n
- Flume and Spillways<\/li>\n
- Reservoirs<\/li>\n
- Penstock with Valve<\/li>\n<\/ul>\n
While HPL contained all necessary components as an out-of-the-box solution, specific components were customized to allow for more flexible and accurate modeling based on available data from the plant. With little or no flowrate data available for the plant, the model was built using the physical components in the library with adjustable coefficients as needed.<\/p>\n
Intake Gate<\/h4>\n
The intake gate is modeled as a linear valve with the gate velocity as an input (<\/span><\/span>Figure 1<\/span><\/span>). The valve is characterized based on an estimate of the gate position at maximum flow.\u00a0<\/span><\/span><\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/Intake-Gate.png\")
Figure 1: Intake gate model<\/figcaption><\/figure>\n Flumes and Spillways<\/h4>\n
The upper and lower flumes as well as the spillway were modeled as reservoirs with varying elevation. A custom friction model was also implemented and integrated into the reservoir model. Since the focus of the controls development was the avoidance of spill, it was important that the flume system flowed correctly at different depths. To ensure this critical system characteristic, operator setpoint data was used for the upper flume depth as a function of flowrate. A unit test model of the upper flume was used to calibrate the roughness coefficient for the friction model.<\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/Flumes-and-Spillways.jpg\")
Figure 2: Calibration model for upper flume<\/figcaption><\/figure>\n Reservoir<\/h4>\n
The reservoir for the basin and forebay are based on the open volume component in the Modelon HPL. However, since the reservoirs in the actual flume system were an irregular shape due to topological variations of the geography, a custom component was created for the reservoirs in the flume system.<\/p>\n
Penstock with Valve<\/h4>\n
Th<\/span><\/span>e system model includes a dynamic model of the forebay and the penstock with a specified downstream pressure at the outlet.<\/span><\/span>\u00a0Based on the location of the penstock connection to the forebay,\u00a0<\/span><\/span>hydrostatic pressure drives the flow in the penstock.\u00a0<\/span><\/span>Th<\/span><\/span>e<\/span><\/span>\u00a0penstock\u00a0<\/span><\/span>valve is a linear\u00a0<\/span><\/span>valve that is character<\/span><\/span>ized to deliver the maximum plant flow at a flow command of 1<\/span><\/span>.\u00a0<\/span><\/span><\/p>\n
Control System Design<\/h4>\n
To evaluate the proposed control\u00a0<\/span><\/span>algorithm<\/span><\/span>,\u00a0<\/span><\/span>Modelon<\/span>\u00a0<\/span><\/span>designed\u00a0<\/span><\/span>a control system\u00a0<\/span><\/span>for integration with the plant<\/span><\/span>.\u00a0<\/span><\/span>\u00a0<\/span><\/span>T<\/span><\/span>he\u00a0<\/span><\/span>penstock control\u00a0<\/span><\/span>system<\/span><\/span>\u00a0<\/span><\/span>includ<\/span><\/span>ed<\/span><\/span>\u00a0the following\u00a0<\/span><\/span>operating\u00a0<\/span><\/span>modes<\/span><\/span> (see Figure 4)<\/span><\/span>:<\/span><\/span><\/p>\n
- \n
- Open-loop<\/li>\n
- Continuous control<\/li>\n
- Steady-state detection<\/li>\n<\/ul>\n
This controller allows the system to be simulated in both normal and emergency modes including a realistic specification of the driving elements for the system response, such as the maximum gate opening velocity, in a manner similar to the actual plant.<\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/Controls-Development.jpg\")
Figure 4: Penstock control including open loop, continuous control, and discrete control with switching<\/figcaption><\/figure>\n Control System Testing<\/h4>\n
To\u00a0<\/span><\/span>verify the control system, a virtual test plan was developed to prove out the physical model and calibrate the proprietary control algorithm<\/span><\/span>.\u00a0 The<\/span><\/span>\u00a0<\/span><\/span>test plan\u00a0<\/span><\/span>included\u202flevel setpoint step tests, varying oscillation tests, and max flow step tests.\u00a0<\/span><\/span>\u00a0<\/span><\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/image48.gif\")
Figure 5: This GIF shows the system response to a large change in the input flowrate. The system is operating at a low level and then the gate is opened as fast as possible to input more flow into the system. The water level depth (shown in red) changes as the water flows through the system. Waves are visible as the system responds to the new input flow. The max height is shown in red.<\/figcaption><\/figure>\n Control System<\/span><\/span>\u00a0Commissioning<\/span><\/span>\u00a0<\/span><\/h4>\n
After the control system was verified using the model, the controller was commissioned on the plant. A comprehensive test plan was established to ensure signal integrity, controller response, and actuator bandwidth and response over a wide range of operating conditions. Figure 6 shows one of the results from a max flow test from the virtual validation testing conducted prior to commissioning. This test is one of the most severe for the system and showed that the control scheme was able to manage the water level and avoid spillage.<\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/Step-Test-Simulation.png\")
Figure 6: System response to max flow step test at high level<\/figcaption><\/figure>\n Figure 7 shows results <\/span><\/span>from a\u00a0<\/span><\/span>max flow<\/span><\/span>\u00a0test during the commissioning.\u00a0<\/span><\/span>The\u00a0<\/span><\/span>commissioning data is provided at\u00a0<\/span><\/span>5-minute<\/span><\/span>\u00a0intervals, where the intake setpoint is representative of the intake system flow.\u00a0<\/span><\/span>Note that this test is not identical to the one conducted in the virtual validation as boundary conditions\u00a0<\/span><\/span>could not be controlled in\u00a0<\/span>exactly the same<\/span>\u00a0way but is close enough for qualitative comparison of the system response between the model and the actual system.<\/span><\/span>\u00a0<\/span><\/p>\n
![\"Hydroelectric](\"https:\/\/modelon.com\/ja\/wp-content\/uploads\/sites\/3\/2022\/06\/Step-Test-Commisioning-Results.jpg\")
Figure 7: System response to max flow step test during plant commissioning<\/figcaption><\/figure>\n Tests concluded that the model-based\u202fcalibration\u202fdeveloped during analytic work was successfully validated \u2013 efficiently managing spillage risk.<\/p>\n
Summary<\/h4>\n
With the assistance of Modelon, the U.S. based energy company successfully designed, tested, and implemented a control system that minimized the environmental impact of their hydroelectric power plant. Through this work, a few different best practices to improve future model-based controls development processes were identified:<\/p>\n