Multi-Mode Predictive Control For Grid-Tied Power Converters With High-Order-Filter In Micro Energy Systems

With the development of science and technology,as well as the society progress,there has been a sharp rise in the demand for clean energy.Renewable energy,e.g.,wind,solar,terrestrial heat,and so on,is recognized as the main source of future power.This may reshape the operating form of electrical power system,i.e.from a centralized to distributed scheme.Microenergy systems(MESs)are an efficient measure for coordinated energy management in power systems.It unifies four energy transmitting steps,i.e.production,consumption,storage,and conversion,within an area and finally converts these energies into electricity.As the extension of microgrids,MESs can adaptively adjust the out/input of sources,loads,and storage,which is efficient and economical.Grid-tied converters(GTCs)are the core equipment of MESs,and usually act as the interface for types of power generation units.In AC MESs,GTCs convert various voltage classes of DC source to AC source(with the same voltage amplitude and frequency).However,GTCs have different working modes in MESs,resulting in problems such as tracking-bias,difficulty of mode switching,and poor cooperative control result.This article will introduce and solve these problems for high-order GTCs(taking LCL-filter as an example),which work under grid-following/grid-forming/grid-supporting modes.For the MESs,GTCs work under the grid-following mode,when connected to the utility power,and generate active/reactive power into the power grid according to the reference.As a high-order filter,LCL-filtered converter has good high-frequency harmonic restraint capability,smaller size,and lower cost than a simple L-or LC-filter at large power ratings.However,as a three-order system,the LCL-filters are easy to cause oscillation and resonance.To solve this,traditional PI control needs cascaded structure and damping method.Finite-control-set model predictive control(FCS-MPC)has become a powerful alternative for power converters,with a simple structure,rapid transient response,and being able to include strong coupling constraints.Besides,FCS-MPC MPC is inherently suitable for MIMO systems by using a straightforward cost function design,and is able to include system nonlinearities and constraints.However,existing multi-variable predictive methods,for the LCL-filtered converter,have steady-state tracking bias in output current.Thus,we establish a mathematical model to describe the underlying system,and propose a simple bias-free FCS-MPC to eliminate the tracking bias.The AC bus of MESs is often worked under weak grid condition.The equivalent model of a weak grid is usually constructed of a stiff grid and a series impedance.At this time,when a number of grid-following GTCs are connected to AC bus,the quality of electric energy of AC bus will worsen.Thus,grid-following GTCs are suitable for strong power grid,while,gridforming ones have better stability.However,the classical mode switching methods use hysteresis or direct switching strategy,which may lead switching oscillating current.To solve it,I designed a flexible mode-switching predictive control based on the adaptive switching function.This method uses the characteristics of full-variable predictive control that are easy to include constraints and change weight factors.Through the hyperbolic tangent curve as the switching function,the proposed method can switch the working modes at the controllable switching time.Besides,it has a low switching ripple with high speed.GTCs,under grid-supporting mode,typically use droop control.However,this may lead to the output global grid voltage(amplitude and frequency)error in GTCs,along with the inaccurate power sharing problem.For MESs,it will also affect the economic efficiency or the working life of distributed sources.To solve these problems,I optimize the cooperative control(secondary control)strategy of converter clusters,and propose a distributed model predictive control(DMPC)method.Taking a four-node parallel converter cluster as an example,I verify the proposed method can compensate for output voltage error,and achieve accurate power sharing under load-switching condition.Besides,compared with classical strategies,the proposed DMPC method simplifies the structure of secondary control and effectively improves the dynamic performance of the underlying system.In summary,this work is proposed to improve the control strategies of GTCs,to achieve high-performance,high-efficiency,and high-reliability operation in MESs.