Intellectual Properties

​Vector control in power transmission and distribution systems
​

In today’s electric power systems, AC/DC and DC/AC power converters are employed to interface between DC and AC systems (Fig. 1) . Today, this kind of power electronic converters is the primary component poised to revolutionize the electric power transmission and distribution systems. A critical issue for energy conversion and interface between DC and AC systems is the control of the power converters, the converters indicated by green boxes in Fig. 1. Traditionally, this type of converters is controlled by using the standard vector control approach. However, recent studies have revealed drawbacks of conventional control methods that result in low power quality, inefficient power generation and conversion, loss of electricity, and obstructions to the healthy growth of smart grids. 

Fig. 1. A microgrid with inverter-interfaced distributed energy sources

​I. Direct Current Vector Control

​Related intellectual properties:

  1. Shuhui Li, Xingang Fu, and Ishan Jaithwa, “Systems and methods for providing vector control of a grid connected converter with resonant circuit grid filters ,” U.S. Patent, No. 10,333,390; Granted on June 25, 2019.
  2. Shuhui Li and Timothy A. Haskew, “Control of a Permanent Magnet Synchronous Generator Wind Turbine,” U.S. Patent, No. 9,391,554; Granted on July 12, 2016.
  3. Shuhui Li and Timothy A. Haskew, “Intelligent Power Converter Control for Grid Integration of Renewable Energies,” U.S. Patent, No. 9,124,140; Granted on September 1, 2015.
  4. Shuhui Li and Timothy A. Haskew, “Converter Control of Variable Speed Wind Turbines,” U.S. Patent, No. 8,577,508; Granted on November 5, 2013​.

Main characteristics:

  1. A PI-based vector control technology
  2. Overcome the limitations of conventional vector control methods
  3. Improved system stability and reliability
  4. Enhanced power quality
  5. Ability to maintain proper operation considering constraints of physical systems
  6. Ability to maintain normal operation under weak grid conditions
  7. A vector control strategy with low computational cost
Conventional standard vector control for a L-filter converter scheme
Direct current vector control for a L-filter converter scheme

​II. Artificial neural network Vector Control

​Related intellectual properties:

  1. ​Shuhui Li, Yang Sun, Malek Ramezani, and Yang Xiao, (2018) “Artificial Neural Networks for Autonomous Volt/Var Control of Inverter Interfaced Units,” United States Patent Application 20200064782.
  2. Shuhui Li, Xingang Fu, Ishan Jaithwa and Raed Suftah, “Systems and methods for providing vector control of a grid connected converter with resonant circuit grid filters ,” U.S. Patent, No. 10,333,390; Granted on June 25, 2019.
  3. Shuhui Li, Michael Fairbank, Xingang Fu, Donald Wunsch, and Eduardo Alonso, “Systems, Methods and Devices for Vector Control of Permanent Magnet Synchronous Machines Using Artificial Neural Networks,” U.S. Patent, No. 9,754,204; Granted on September 5, 2017.
  4. Shuhui Li, Donald C. Wunsch, and Michael Fairbank, “Vector Control of Grid-Connected Power Electronic Converter using Artificial Neural Networks,” U.S. Patent, No. 9,379,546; Granted on Jun 28, 2016.

Main characteristics:

  1. A neural network controller trained based on approximate dynamic programming
  2. Compared to conventional vector control methods, neural network control approach produces the fastest response time, lowest overshoot, and, in general, the closest to ideal performance.
  3. Reduce the harmonics injected into the grid 
  4. Improve power system reliability and stability 
  5. Enhance the energy conversion efficiency 
  6. Reduce sizes of energy conversion systems and components
  7. Able to determine a best and reliable approach to control the power converter when there is a trend for the converter to operate beyond its physical constraints
  8. Potential to connect a converter to the AC grid without the need for pre-synchronization
  9. Ability to maintain normal operation under weak grid conditions
  10. More computational complex than PI-based control
Fig. 2. Conventional standard (top) and neural network (enclosed by red and blue boxes) vector control illustration for AC/DC and DC/AC power converters

Demand response and smart home energy management systems

The global market for grid-connected residential solar PV installations coupled with energy storage is predicted to grow tenfold to reach more than 900 MW in 2018, making smart management and control of residential PV and energy storage increasingly important. To electric utilities, proper management of PV and energy storage will help to maintain grid reliability, resiliency, and power quality while minimizing curtailment of available solar power. To home owners, effective use of PV and energy storage will help consumers to save money off the electricity bill. For mutual benefit, residential PV and energy storage should be managed and controlled holistically according to dynamic electricity prices and tariffs in a demand response framework.  

Fig. 3. Integration of PV, battery, supercapacitor and converters

Related intellectual properties:

  1. Shuhui Li, Yixiang Gao, Weizhen Dong, and Bing Lu, (02/25/2022) “Intelligent Building Energy Management in Joint Day-Ahead and Real-Time Electricity Markets,” Filed in Febuary to The University of Alabama, Pending.
  2. Shuhui Li and Yang Xiao, “Cyber Physical Systems: Smart and Secure Home Energy Management System,” Filed in July 2014 to The University of Alabama, Approved by The University of Alabama on October 13, 2014.​
  3. Shuhui Li, Dong Zhang, and Min Sun, “Systems and methods for modeling energy consumption and creating demand response strategies using learning-based approaches,” U.S. Patent, No. 9,817,375​; Granted on Nov 14, 2017.

Main characteristics:

  1. A decoupled DR mechanism by classifying home appliances into fixed, deferrable and regulate-able loads
  2. A special technique that treats fixed loads as future fixed loads, deferrable loads as future fixed load patterns, and PVs as a future fixed generation resource
  3. Special data-driven learning mechanisms to learn energy consumption models of home appliances, house energy performance, energy generation characteristics of PVs, and energy storage properties of batteries.
  4. A mechanism to update energy consumption models daily so as to accurately capture the thermal or energy consumption behavior of a house or appliance at present seasons, users, and weather conditions.
  5. An optimization method capable of decoupling a complicated and multi-objective optimization problem into small independent optimization problems.   
  6. An optimal demand response technique that is obtained based on weather prediction, day-ahead or real-time electricity price and the learning-based energy consumption model of the house and appliances.
Fig. 4. Learning based demand response and home energy management system

High performance motor control using artificial neural networks

An electric motor is a device that converts mechanical energy to electrical energy or electrical energy to mechanical energy. It is widely used in broad motion and control applications. Today, the primary electric machines used in diverse control applications are AC induction machines and  AC permanent magnet synchronous machines. Presently, the primary control technique for these ac electric machines is the standard vector control technology. 
The IPs in this categories are to provide power, defense, automation and process industries with the highest performance and to improve efficiency, reliability and durability of electric motors and energy conversion systems. 

Fig. 5. Motor Control in power, defense, automation and process industries

Related intellectual properties:

  1. Shuhui Li, Weizhen Dong, and Yixiang Gao, “Neural-Network Based MTPA, Flux-Weakening and MTPV for IPM Motor Control and Drives,” United States Patent Application pending, 2021.
  2. Xingang Fu and Shuhui Li, “Systems, methods and devices for vector control of induction machines using artificial neural networks,” Granted on April 27, 2021, as U.S. Patent No. 10,990,066.
  3. Shuhui Li, Michael Fairbank, Xingang Fu, Donald Wunsch, and Eduardo Alonso, “Systems, Methods and Devices for Vector Control of Permanent Magnet Synchronous Machines Using Artificial Neural Networks,” U.S. Patent, No. 9,754,204; Granted on September 5, 2017.
  4. Shuhui Li, Xingang Fu, Hoyun Won, and Yang Sun, “Systems, Methods and Devices for Approximate Dynamic Programming Vector Controllers for Operation of IPM Motor in Linear and Over Modulation Regions,” U.S. Patent 10,367,437; Granted on July 30, 2019.
  5. Shuhui Li and Michael Fairbank, (2019) “Systems, Methods and Devices for Neural Network Control for IPM Motor Drives,” United States Patent Application 20200266743​.

Main characteristics:

  1. ​An optimal neural network controller trained based on approximate dynamic programming
  2. Compared to conventional standard vector control methods, neural network control produces the fastest response time, lowest overshoot, and, in general, the closest to ideal performance.
  3. Reduce the harmonics injected into the motor 
  4. Improve motor reliability and stability especially in over modulation ranges
  5. Enhance the energy conversion efficiency 
  6. Reduce sizes and cost of electric motors and power converters
  7. Ability to maintain normal operation under over modulation conditions
  8. More computational complex than PI-based control
Induction motor conventional and neural network vector control
PMSM conventional and neural network vector control

High performance control of dc/dc converters using artificial neural networks
​

​Related intellectual properties:

  1. Shuhui Li, Xingang Fu, and Weizhen Dong, (2018) ​”Systems, Methods and Devices for Control of DC/DC Converters and A Standalone DC Microgrid  Using Artificial Neural Networks,” United States Patent Application 20190296643.​
  2. Shuhui Li, Weizhen Dong,  Xingang Fu, and Michael Fairbank, (2019) “Control of a Buck dc/dc Converter Using Approximatie Dynamic Programming and Artificial Neural Network,” United States Patent Application 20200251986.

Main characteristics:

  1. ​More computational complex than PI-based control​
  2. ​An optimal neural network controller trained based on approximate dynamic programming
  3. Compared to conventional standard vector control methods, neural network control produces the fastest response time, lowest overshoot, and, in general, the closest to ideal performance.
  4. Improve dc/dc converter reliability and stability 
  5. High performance using low switching frequency and sampling rate
  6. Good ability to handle maximum inductor current constraint
  7. Good ability to handle maximum duty ratio constraint