MPC-based intra-day rolling optimization scheduling strategy for traction power supply system

Abstract

The invention discloses an intra-day rolling optimization scheduling strategy of a tractive power supply system based on MPC, and belongs to the field of power system optimization scheduling. The method comprises the following steps: constructing a state space model of a source network load storage tractive power supply system; a constraint condition is preset, the state space model of the traction power supply system is used as a prediction model, the minimum tracking deviation is used as a target function, and an MPC controller is constructed; the MPC controller is used for intra-day rolling optimization, after rolling optimization control at the current moment is executed, the actual running state vector of the system is collected for feedback correction and serves as an initial value of rolling optimization at the next moment, deviation correction is conducted on the running state of the hybrid energy storage system at the next time period, and closed-loop optimization control is formed. According to the method, the problems of influence of uncertain high-frequency components existing in photovoltaic output and traction load on system optimization operation in actual daily operation and difficulty in realization of dynamic balance of system traction power supply capacity and load demand in the actual daily operation process are solved.

Description

Technical Field

[0001] The present invention belongs to the field of optimal dispatching of power systems, and particularly relates to an intraday rolling optimal dispatching strategy for a traction power supply system based on MPC. Background Art

[0002] With the rapid development and wide application of photovoltaic power generation technology, the high intermittency of photovoltaic output and the uncertainty of traction load have increased the complexity of the operation of the power supply system. By implementing intraday rolling optimization based on MPC, a better balance can be achieved between the unstable output of photovoltaic power generation and the changing traction load, effectively performing peak shaving and valley filling, and at the same time using energy storage systems to enhance the stability and power supply reliability of the power grid. Therefore, it is necessary to study an intraday rolling optimal dispatching strategy for the "source-network-load-storage" traction power supply system based on MPC to solve problems such as the impact of the uncertain high-frequency components of photovoltaic output and traction load on the optimal operation of the system during intraday actual operation, and the difficulty in achieving the dynamic balance between the traction power supply capacity and load demand of the system during the intraday actual operation process. Summary of the Invention

[0003] Aiming at the above deficiencies in the prior art, an intraday rolling optimal dispatching strategy for a traction power supply system based on MPC provided by the present invention solves the problems of the impact of the uncertain high-frequency components of photovoltaic output and traction load on the optimal operation of the system during intraday actual operation and the difficulty in achieving the dynamic balance between the traction power supply capacity and load demand of the system during the intraday actual operation process.

[0004] To achieve the above invention purpose, the technical solution adopted by the present invention is: an intraday rolling optimal dispatching strategy for a traction power supply system based on MPC, including:

[0005] Construct a state space model of the source-network-load-storage traction power supply system; preset constraint conditions, and use the state space model of the source-network-load-storage traction power supply system as a prediction model, and construct an MPC controller with the minimum tracking deviation as the objective function;

[0006] Use the MPC controller for intraday rolling optimization. After the rolling optimization control is executed at the current k moment, collect the actual operation state vector of the system for feedback correction, and use it as the initial value for the finite-time domain rolling optimization at the k + 1 moment to correct the deviation of the operation state of the hybrid energy storage system in the next time period, thereby forming a closed-loop optimization control.

[0007] Further, the expression of the state space model of the source-network-load-storage traction power supply system is:

[0008] x(k + 1) = Ax(k) + Bu(k) + Dd(k)

[0009] y(k) = Cx(k)

[0010] x(k) = [P tr_L (k), P tr_R (k), P α (k), P β (k), P b (k), P sc (k), S b (k), S sc (k), Q α (k), Q β (k)] T

[0011]

[0012]

[0013] y(k) = [P grid (k), S b (k), S sc (k)] T

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021] P tr_L (k) = P tr_L,+ (k) - P tr_L,- (k), P tr_R (k) = P tr_R,+ (k) - P tr_R,- (k)

[0022]

[0023] where is the state variable of the source-network-load-storage traction power supply system at time is the state variable of the source-network-load-storage traction power supply system at time is Control variables of the time source-network-load-storage traction power supply system; is Output vector of the time source-network-load-storage traction power supply system; Both are coefficient matrices; is the current time; is Interactive power between the train on the left power supply arm and the public power grid at time ; is Interactive power between the train on the right power supply arm and the public power grid at time ; is at time Active power generated by the phase change converter; is at time Active power generated by the phase change converter; is Total power of the battery energy storage system at time ; is Total power of the supercapacitor energy storage system at time ; is Capacity of the lithium battery at time ; is Capacity of the supercapacitor at time ; is at time Reactive power generated by the phase change converter; is at time Reactive power generated by the phase change converter; is the transpose; is the difference symbol; is Active power of the phase change converter for the control output at time ; is Active power of the phase change converter for the control output at time ; is Reactive power of the phase change converter for the control output at time ; is Reactive power of the phase change converter for the control output at time ; is Active power of the train on the left power supply arm at time ; is Active power of the train on the right power supply arm at time ; is Reactive power of the train on the left power supply arm at time ; is Reactive power of the train on the right power supply arm at time ;​​​​ is the active power of the photovoltaic at time is the active power generated by the power grid at time is the predicted active power of the train on the left power supply arm at time predicted at time is the time sampling period; is the number of time sampling periods; is the number of sampling period sequences; is the initial day-ahead predicted value of the active power of the left power supply arm at time is at time calculated at time the actual value of the active power of the left power supply arm at time is at time predicted at time the active power of the train on the right power supply arm at time is the initial day-ahead predicted value of the active power of the right power supply arm at time is at time calculated at time the actual value of the active power of the right power supply arm at time is at time predicted at time the reactive power of the train on the left power supply arm at time is the initial day-ahead predicted value of the reactive power of the left power supply arm at time is at time calculated at time the actual value of the reactive power of the left power supply arm at time is at time predicted at time the reactive power of the train on the right power supply arm at time is the initial day-ahead predicted value of the reactive power of the right power supply arm at time is at time calculated at time the actual value of the reactive power of the right power supply arm at time is at time predicted at time the active power of the photovoltaic at time is the initial day-ahead predicted value of the photovoltaic power generation at time is at time predicted at time the power generation power of the photovoltaic at time is ​The predicted total power of the battery energy storage system at a certain moment; is The power increment of the battery energy storage system at a certain moment; is The total power of the supercapacitor energy storage system at a certain moment; is The power increment of the supercapacitor energy storage system at a certain moment; is The interface power from the power grid to the traction power supply system at a certain moment; is The positive power transmitted from the power grid to the traction power supply system at a certain moment; is The negative power transmitted from the power grid to the traction power supply system at a certain moment; is The predicted total power of the battery energy storage system at a certain moment; is The discharge power of the battery energy storage system at a certain moment; is The charging power of the battery energy storage system at a certain moment; is The total power of the supercapacitor energy storage system at a certain moment; is The discharge power of the supercapacitor energy storage system at a certain moment; is The charging power of the supercapacitor energy storage system at a certain moment; is The power transmitted to the left power supply arm through the V / v traction transformer at a certain moment; is The power returned from the left power supply arm to the common power grid at a certain moment; is The power transmitted to the right power supply arm through the V / v traction transformer at a certain moment; is The power returned from the right power supply arm to the common power grid at a certain moment; is The power transmitted from the left power supply arm to the converter at a certain moment; is At a certain moment The power transmitted from the converter to the left power supply arm; is At a certain moment The power transmitted by the converter to the right power supply arm; is At a certain moment, the power transmitted from the right power supply arm to the converter.

[0024] Furthermore, the constraint condition is:

[0025] P grid (k + sΔt * |k) = P tr_L (k + sΔt * |k) + P tr_R (k + sΔt * |k)

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] Among them, P grid (k + sΔt * (k + sΔt|k) is the actual active power value of the power grid at the moment of k + sΔt calculated at time k; k is the current time; Δt * is the time sampling period; s is the number of time sampling periods; i is the sampling period sequence number; P * (k + sΔt tr_L (k + sΔt|k) is the interactive power between the train on the left power supply arm and the public power grid predicted at time k for the moment of k + sΔt * ; P * (k + sΔt tr_R (k + sΔt|k) is the interactive power between the train on the right power supply arm and the public power grid predicted at time k for the moment of k + sΔt * (k + sΔt|k) is the interactive power between the train on the right power supply arm and the public power grid predicted at time k for the moment of k + sΔt * ; To calculate the actual active power value of the train on the left power supply arm at time \(k + s\Delta t\); \(P\) * α (k + s\Delta t * |k) is the predicted active power of the α - converter at time \(k + s\Delta t\); * To predict the active power of the train on the right power supply arm at time \(k + s\Delta t\) at time \(k\); * \(P\) β (k + s\Delta t * |k) is the predicted active power of the β - converter at time \(k + s\Delta t\); \(P\) * b (k + s\Delta t * |k) is the predicted active power of the lithium - battery at time \(k + s\Delta t\); \(P\) * sc (k + s\Delta t * |k) is the predicted active power of the super - capacitor at time \(k + s\Delta t\); * To predict the active power of the photovoltaic at time \(k + s\Delta t\) at time \(k\); * \(Q\) tr_L (k + s\Delta t * |k) is the calculated actual power value of the left power supply arm of the traction substation at time \(k + s\Delta t\); * To predict the reactive power of the train on the left power supply arm at time \(k + s\Delta t\) at time \(k\); * \(Q\) α (k + s\Delta t * |k) is the predicted reactive power of the α - converter at time \(k + s\Delta t\); \(Q\) * tr_R (k + s\Delta t * |k) is the calculated actual power value of the right power supply arm of the traction substation at time \(k + s\Delta t\); * To predict the reactive power of the train on the right power supply arm at time \(k + s\Delta t\) at time \(k\); * \(Q\) β (k + s\Delta t * |k) is the predicted reactive power of the β - converter at time \(k + s\Delta t\); * To predict the capacity of the battery energy storage at time \(k + s\Delta t\) at time \(k\); * To predict the capacity of the battery energy storage at time \(k+(s - 1)\Delta t\) at time \(k\); \(\eta\) * b Is the charge - discharge efficiency of the lithium - battery; The charging and discharging power of the battery energy storage predicted at time k for k + sΔt * The rated capacity of the battery energy storage is; The capacity of the supercapacitor predicted at time k for k + sΔt * is; The capacity of the supercapacitor predicted at time k for k + (s - 1)Δt * is; η sc The charging and discharging efficiency of the supercapacitor The charging and discharging power of the supercapacitor predicted at time k for k + sΔt * is; The rated capacity of the supercapacitor; η b,ch The charging efficiency of the battery energy storage The total power prediction value of the battery energy storage system at time k; η b,dis The discharging efficiency of the battery energy storage; η sc,ch The charging efficiency of the supercapacitor The total power of the supercapacitor energy storage system at time k; η sc,dis The discharging efficiency of the supercapacitor The minimum capacity of the battery energy storage The maximum capacity of the battery energy storage The minimum capacity of the supercapacitor The maximum capacity of the supercapacitor The rated power of the battery energy storage system; P b (k) is the total power of the battery energy storage system at time k; ΔP b (k + sΔt * |k) is the power increment of the battery energy storage system within the future sΔt * time period; The rated power of the supercapacitor energy storage system; P sc (k) is the total power of the supercapacitor energy storage system at time k; ΔP sc (k + sΔt * |k) is the power increment of the supercapacitor energy storage system within the future sΔt * time period; The actual value of the active power of the power grid calculated at time k for k + sΔt * is; M g The transmission line capacity limit The power transmitted by the traction transformer to the trains on the left power supply arm predicted at time k for k + sΔt * is; The power transmitted by the traction transformer to the trains on the right power supply arm predicted at time k for k + sΔt * is; is the active power of the train on the left power supply arm at time k; is the active power of the train on the right power supply arm at time k; a i , b i and c i are all coefficients of the regular dodecagon linearized power flow constraint; is the rated capacity of the α converter; is the rated capacity of the β converter; is the power of the left power supply arm of the traction substation predicted at time k + sΔt * ; is the upper limit value of the voltage unbalance degree; S d is the short-circuit capacity of the external power grid; is the power of the right power supply arm of the traction substation predicted at time k + sΔt * ; Δ is the difference symbol.

[0041] Furthermore, the expression of the objective function is:

[0042] J = min{A1 + A2}

[0043]

[0044]

[0045] where J is the objective function; A1 is the deviation of the power input to the traction substation and the state of charge of the hybrid energy storage system; A2 is the change in the output power of the energy storage; N is the set of s; s is the number of time sampling periods; ω s is the weight coefficient of the power input to the traction substation in the day-ahead stage; f1 is the weighted error term of the squared deviation of the grid power; ω b is the weight coefficient of the state of charge tracking error of the battery energy storage system; f2 is the weighted error term of the squared deviation of the state of the energy storage system; ω sc is the weight coefficient of the state of charge tracking error of the supercapacitor system; f3 is the weighted error term of the squared deviation of the state of the supercapacitor; is the expected power of the power grid at time k + sΔt * ; Δt * is the time sampling period; is the active power of the power grid predicted at time k + sΔt * at time k; is the expected capacity of the battery energy storage at time k + sΔt * ; is the capacity of the battery energy storage predicted at time k + sΔt * at time k; is the expected capacity of the supercapacitor at time k + sΔt * ; The capacity of the supercapacitor at the predicted k + sΔt moment at time k * ; λ b and λ sc are both weight coefficients of the output power increment of the hybrid energy storage system issued by the MPC controller; ΔP b (k + sΔt * |k) is the change in the output power of the battery energy storage system relative to the previous time period; ΔP sc (k + sΔt * |k) is the change in the output power of the supercapacitor energy storage system relative to the previous time period.

[0046] Furthermore, the expression for the initial value of the finite - horizon rolling optimization at the k + 1 moment is as follows:

[0047]

[0048] where is the initial value at the k + 1 moment predicted by the MPC controller after rolling optimization and execution of the lithium - battery energy storage at time k; is the actual value deterministically obtained by the lithium - battery energy storage according to the actual data of photovoltaic output and traction load before the arrival of the k + 1 moment after the MPC controller issues a control command in the k time period; is the initial value at the k + 1 moment predicted by the MPC controller after rolling optimization and execution of the ultra - large capacitor energy storage at time k; is the actual value deterministically obtained by the ultra - large capacitor energy storage according to the actual data of photovoltaic output and traction load before the arrival of the k + 1 moment after the MPC controller issues a control command in the k time period; is the initial value at the k + 1 moment predicted by the MPC controller after rolling optimization and execution of the grid energy storage at time k; is the actual value deterministically obtained by the grid energy storage according to the actual data of photovoltaic output and traction load before the arrival of the k + 1 moment after the MPC controller issues a control command in the k time period.

[0049] The beneficial effects of the present invention are as follows: Using the day - ahead IGDT robust operation plan and the intra - day source - load prediction as the model - prediction input, through finite - horizon rolling optimization and actual - state information feedback correction, the predicted control outputs the intra - day output adjustment plan of the hybrid energy storage system to reduce the impact of uncertain factors on the stable and economic operation of the system. Description of the Drawings

[0050] Figure 1 is the flowchart of the method of the present invention.

[0051] Figure 2 is the flowchart of the intra - day time - domain rolling optimization scheduling of the traction power supply system based on MPC in the embodiment of the present invention.

[0052] Figure 3 This is the simulation result diagram of the incoming power of the traction substation before and after rolling optimization in the embodiment of the present invention.

[0053] Figure 4 This is the diagram of the day-ahead reference trajectory and the intraday rolling feedback control result of the state of charge of the battery and the super capacitor in the embodiment of the present invention.

[0054] Figure 5 This is the comparison diagram of the effects of intraday rolling optimization and non-rolling optimization in the embodiment of the present invention. Detailed implementation manners

[0055] The following describes the detailed implementation manners of the present invention so that those skilled in the art of the present technology can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed implementation manners. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0056] As Figure 1 and Figure 2 shown, in an embodiment of the present invention, an intraday rolling optimization scheduling strategy for a traction power supply system based on MPC includes:

[0057] Construct a state space model of the source-network-load-storage traction power supply system; preset constraint conditions, and use the state space model of the source-network-load-storage traction power supply system as a prediction model, and construct an MPC controller with the minimum tracking deviation as the objective function;

[0058] Use the MPC controller for intraday rolling optimization. After the rolling optimization control is executed at the current k moment, collect the actual operation state vector of the system for feedback correction, and use it as the initial value for the finite-time domain rolling optimization at the k + 1 moment to correct the deviation of the operation state of the hybrid energy storage system in the next time period, so as to form a closed-loop optimization control.

[0059] In this embodiment, the method is based on the day-ahead IGDT robust operation plan, and performs rolling optimization with the minimum tracking deviation as the goal to obtain the optimized incoming power of the substation and the SOC of the hybrid energy storage system, and only issues and executes the intraday adjustment plan at the current moment; then, taking the actual operation state of the traction power supply system as feedback, gradually update the initial value of the rolling optimization, and correct the charging and discharging plan of the intraday energy storage system and the power purchase plan of the traction substation, so as to achieve closed-loop control. The present invention solves the problems such as the influence of the uncertain high-frequency components of the photovoltaic output and the traction load in the intraday actual operation on the optimal operation of the traction power supply system, and the difficulty in achieving the dynamic balance between the traction power supply capacity and the load demand in the intraday actual operation process.

[0060] MPC is an optimal control algorithm for uncertain systems. Compared with traditional control methods, MPC predicts the dynamic behavior of the system for a period of time in the future through a prediction model, uses an optimization method to calculate the optimal control input according to preset goals and constraints, and corrects it based on the real-time state information of the system at each sampling moment to respond to the changes in the system state in real time. Essentially, it is a closed-loop optimal control method centered on model prediction, rolling optimization, and feedback correction. Based on the idea of model predictive control, this invention takes tracking the daily planned value of the traction substation as the intraday optimization goal, divides the entire scheduling cycle into multiple time intervals, and corrects and modifies the output plan of the hybrid energy storage system in advance through three links: model prediction, rolling horizon optimization, and feedback correction, so as to reduce the impact of the high-frequency components of the random fluctuations of the intraday photovoltaic output and traction load on the system operation state.

[0061] As the basis of the MPC algorithm, the prediction model is used to predict the state and output of the system at future moments based on the state information of the system at the current moment and the control input at future moments. Therefore, in the MPC algorithm, the prediction model does not have a specific form and can be a fuzzy model, a neural network model, a transfer function, a state-space equation, etc. More emphasis is placed on the prediction function of the model. According to the prediction results, it is judged whether the system operation constraints and corresponding performance indicators are met, so as to facilitate the decision-maker to adjust and optimize the control strategy to meet the system output goal.

[0062] The expression of the state-space model of the source-network-load-storage traction power supply system is as follows:

[0063] x(k + 1) = Ax(k) + Bu(k) + Dd(k)

[0064] y(k) = Cx(k)

[0065] x(k) = [P tr_L (k), P tr_R (k), P α (k), P β (k), P b (k), P sc (k), S b (k), S sc (k), Q α (k), Q β (k)] T

[0066]

[0067]

[0068] y(k) = [P grid (k), S b (k), Ssc (k)] T

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] P tr_L (k) = P tr_L,+ (k) - P tr_L,- (k), P tr_R (k) = P tr_R,+ (k) - P tr_R,- (k)

[0077]

[0078] where is the state variable of the source-grid-load-storage traction power supply system at time is the state variable of the source-grid-load-storage traction power supply system at time is the control variable of the source-grid-load-storage traction power supply system at time is the output vector of the source-grid-load-storage traction power supply system at time are all coefficient matrices is the current time is the interactive power between the train on the left power supply arm and the public power grid at time is the interactive power between the train on the right power supply arm and the public power grid at time is at time the active power generated by the phase converter is at time the active power generated by the phase converter is the total power of the battery energy storage system at time is Total power of the supercapacitor energy storage system at a certain moment; is Capacity of the lithium battery at a certain moment; is Capacity of the supercapacitor at a certain moment; is a certain moment Reactive power generated by the phase converter; is a certain moment Reactive power generated by the phase converter; is the transpose; is the difference symbol; is Active power of the phase converter controlled and output at a certain moment ; is Active power of the phase converter controlled and output at a certain moment ; is Reactive power of the phase converter controlled and output at a certain moment ; is Reactive power of the phase converter controlled and output at a certain moment ; is Active power of the train on the left power supply arm at a certain moment; is Active power of the train on the right power supply arm at a certain moment; is Reactive power of the train on the left power supply arm at a certain moment; is Reactive power of the train on the right power supply arm at a certain moment; is Active power of the photovoltaic at a certain moment; is Active power generated by the power grid at a certain moment; is at a certain moment predicted Active power of the train on the left power supply arm at a certain moment; is the time sampling period; is the number of time sampling periods; is the sampling period sequence number; is Initial day-ahead prediction value of the active power of the left power supply arm at a certain moment; is at a certain moment calculated Actual value of the active power of the left power supply arm at a certain moment; is at a certain moment predicted Active power of the train on the right power supply arm at a certain moment; is the initial day-ahead predicted value of the active power of the right power supply arm at at the actual value of the active power of the right power supply arm calculated at at the reactive power of the trains on the left power supply arm predicted at is the initial day-ahead predicted value of the reactive power of the left power supply arm at at the actual value of the reactive power of the left power supply arm calculated at at the reactive power of the trains on the right power supply arm predicted at is the initial day-ahead predicted value of the reactive power of the right power supply arm at at the actual value of the reactive power of the right power supply arm calculated at at the active power of the photovoltaic power predicted at is the initial day-ahead predicted value of the photovoltaic power generation at at the power generation of the photovoltaic power predicted at is the predicted total power of the battery energy storage system at is the power increment of the battery energy storage system at is the total power of the supercapacitor energy storage system at is the power increment of the supercapacitor energy storage system at is the interface power from the power grid to the traction power supply system at is the positive power transmitted from the power grid to the traction power supply system at is the negative power transmitted from the power grid to the traction power supply system at is the predicted total power of the battery energy storage system at is the discharge power of the battery energy storage system at is​​​​​​​ The charging power of the battery energy storage system at time is The total power of the supercapacitor energy storage system at time is The discharging power of the supercapacitor energy storage system at time is The charging power of the supercapacitor energy storage system at time is The power transmitted to the left power supply arm through the V / v traction transformer at time is The power returned from the left power supply arm to the public power grid at time is The power transmitted to the right power supply arm through the V / v traction transformer at time is The power returned from the right power supply arm to the public power grid at time is The power transmitted from the left power supply arm to the converter at time is At time The power transmitted from the converter to the left power supply arm is At time The power transmitted from the converter to the right power supply arm is At time The power transmitted from the right power supply arm to the converter.

[0079] In this embodiment, the present invention selects the state space equation as the prediction model of the system. For the convenience of subsequent representation, the incoming power of the traction substation, the charging and discharging power of the hybrid energy storage system, the incoming power of the V / v traction transformer, and the compensation power of the back-to-back converter are combined and represented as:

[0080]

[0081]

[0082] P tr_L (k) = P tr_L,+ (k) - P tr_L,- (k), P tr_R (k) = P tr_R,+ (k) - P tr_R,- (k)

[0083]

[0084] The predicted output of the MPC controller is the hybrid energy storage power correction amount ΔP HESSThe system control sequence formed, where the main control variable is the output power increment of the hybrid energy storage, can be expressed as:

[0085]

[0086]

[0087] In the present invention, considering both the prediction accuracy and the prediction duration, the intraday traction load prediction data and the photovoltaic power generation data are simulated by representing them as the superposition of the day-ahead prediction value and the random disturbance, that is:

[0088]

[0089]

[0090]

[0091] Therefore, the main state variables x(k), control variable u(k), disturbance variable d(k), and output variable vector y(k) of the system are defined as follows:

[0092] x(k) = [P tr_L (k), P tr_R (k), P α (k), P β (k), P b (k), P sc (k), S b (k), S sc (k), Q α (k), Q β (k)] T

[0093]

[0094]

[0095] y(k) = [P grid (k), S b (k), S sc (k)] T

[0096] The state-space model of the "source-network-load-storage" traction power supply system can be expressed as:

[0097] x(k + 1) = Ax(k) + Bu(k) + Dd(k)

[0098] y(k) = Cx(k)

[0099] The constraint conditions are:

[0100] Pgrid (k + sΔt * |k) = P tr_L (k + sΔt * |k) + P tr_R (k + sΔt * |k)

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] Among them, P grid (k + sΔt * |k) is the actual value of the active power of the power grid at the moment of k + sΔt calculated at time k; k is the current time; Δt * is the time sampling period; s is the number of time sampling periods; i is the sampling period sequence number; P * tr_L tr_L (k + sΔt * |k) is the interactive power between the train on the left power supply arm and the public power grid at the moment of k + sΔt predicted at time k; P * tr_R tr_R (k + sΔt * |k) is the interactive power between the train on the right power supply arm and the public power grid at the moment of k + sΔt predicted at time k; * is the actual value of the active power of the power grid at the moment of k + sΔt calculated at time k is the actual value of the active power of the power grid at the moment of k + sΔt calculated at time k* Actual value of the active power of the train on the left power supply arm at a certain moment; P α (k + sΔt * |k) is the active power of the α converter predicted at time k for k + sΔt * at a certain moment; is the predicted value of k + sΔt at time k * Actual value of the active power of the train on the right power supply arm at a certain moment; P β (k + sΔt * |k) is the predicted value of k + sΔt at time k * Active power of the β converter at a certain moment; P b (k + sΔt * |k) is the predicted value of k + sΔt at time k * Active power of the lithium battery at a certain moment; P sc (k + sΔt * |k) is the predicted value of k + sΔt at time k * Active power of the super capacitor at a certain moment; is the predicted value of k + sΔt at time k * Active power of the photovoltaic at a certain moment; Q tr_L (k + sΔt * |k) is the calculated value of k + sΔt at time k * Actual value of the power of the left power supply arm of the traction substation at a certain moment; is the predicted value of k + sΔt at time k * Reactive power of the train on the left power supply arm at a certain moment; Q α (k + sΔt * |k) is the predicted value of k + sΔt at time k * Reactive power of the α converter at a certain moment; Q tr_R (k + sΔt * |k) is the calculated value of k + sΔt at time k * Actual value of the power of the right power supply arm of the traction substation at a certain moment; is the predicted value of k + sΔt at time k * Reactive power of the train on the right power supply arm at a certain moment; Q β (k + sΔt * |k) is the predicted value of k + sΔt at time k * Reactive power of the β converter at a certain moment; is the predicted value of k + sΔt at time k * Capacity of the battery energy storage at a certain moment; is the predicted value of k + (s - 1)Δt at time k * Capacity of the battery energy storage at a certain moment; η b is the charge and discharge efficiency of the lithium battery; is the predicted value of k + sΔt at time k* The charging and discharging power of the battery energy storage at a certain moment; is the rated capacity of the battery energy storage; is the capacity of the supercapacitor predicted at the k-th moment for k + sΔt * moment; is the capacity of the supercapacitor predicted at the k-th moment for k + (s - 1)Δt * moment; η sc is the charge-discharge efficiency of the supercapacitor; is the charge-discharge power of the supercapacitor predicted at the k-th moment for k + sΔt * moment; is the rated capacity of the supercapacitor; η b,ch is the charging efficiency of the battery energy storage; is the predicted total power value of the battery energy storage system at the k-th moment; η b,dis is the discharging efficiency of the battery energy storage; η sc,ch is the charging efficiency of the supercapacitor; is the total power of the supercapacitor energy storage system at the k-th moment; η sc,dis is the discharging efficiency of the supercapacitor; is the minimum capacity of the battery energy storage; is the maximum capacity of the battery energy storage; is the minimum capacity of the supercapacitor; is the maximum capacity of the supercapacitor; is the rated power of the battery energy storage system; P b (k) is the total power of the battery energy storage system at the k-th moment; ΔP b (k + sΔt * |k) is the power increment of the battery energy storage system within the future sΔt * period; is the rated power of the supercapacitor energy storage system; P sc (k) is the total power of the supercapacitor energy storage system at the k-th moment; ΔP sc (k + sΔt * |k) is the power increment of the supercapacitor energy storage system within the future sΔt * period; is the actual value of the active power of the power grid calculated at the k-th moment for k + sΔt * moment; M g is the transmission line capacity limit; is the power predicted at the k-th moment for k + sΔt * moment that the traction transformer delivers to the trains on the left power supply arm; is the power predicted at the k-th moment for k + sΔt * moment that the traction transformer delivers to the trains on the right power supply arm; is the active power of the trains on the left power supply arm at the k-th moment; is the active power of the train on the right power supply arm at time k; a i , b i and c i are all coefficients of the linearized power flow constraint of the regular dodecagon; is the rated capacity of the α converter; is the rated capacity of the β converter; is the power of the left power supply arm of the traction substation predicted at time k + sΔt * ; is the upper limit value of the voltage unbalance degree; S d is the short-circuit capacity of the external power grid; is the power of the right power supply arm of the traction substation predicted at time k + sΔt * ; Δ is the difference symbol.

[0116] In this embodiment, in the intraday stage, considering the prediction errors caused by the random fluctuations of the intraday photovoltaic output and the traction load, the active and reactive power balance constraints of the traction power supply system are:

[0117] P grid (k + sΔt * |k) = P tr_L (k + sΔt * |k) + P tr_R (k + sΔt * |k)

[0118]

[0119]

[0120]

[0121] For all time periods within the prediction horizon N f , the dynamic operation constraints of the battery and the supercapacitor are:

[0122]

[0123]

[0124]

[0125]

[0126]

[0127] Considering the operation deviation of the traction load in the intraday stage, the power of the traction substation incoming line, the incoming line powers of the left and right power supply arms of the traction transformer are respectively limited by the transmission line transmission capacity and the traction transformer capacity, and can be expressed as the following constraints:

[0128]

[0129]

[0130] For the back-to-back converter, since it needs to provide active and reactive power compensation to the system simultaneously and still needs to meet the capacity constraint of the back-to-back converter during the intraday stage, the intraday operation constraints are as follows:

[0131]

[0132]

[0133] In addition, to ensure that the unbalance degree of the three-phase voltage of the traction substation does not exceed the limit value during the actual intraday operation process, the three-phase voltage unbalance constraint also needs to be considered during the intraday stage, which is expressed as:

[0134]

[0135]

[0136] The expression of the objective function is:

[0137] J = min{A1 + A2}

[0138]

[0139]

[0140] where J is the objective function; A1 is the deviation of the incoming power of the traction substation and the state of charge of the hybrid energy storage system; A2 is the change in the output power of the energy storage; N is the set of s; s is the number of time sampling periods; ω s is the weight coefficient of the incoming power of the traction substation in the day-ahead stage; f1 is the weighted error term of the squared deviation of the grid power; ω b is the weight coefficient of the state-of-charge tracking error of the battery energy storage system; f2 is the weighted error term of the squared deviation of the state of the energy storage system; ω sc is the weight coefficient of the state-of-charge tracking error of the supercapacitor system; f3 is the weighted error term of the squared deviation of the state of the supercapacitor; is the expected power of the grid at the future time k + sΔt * ; Δt * is the time sampling period; is the active power of the grid at the time k + sΔt predicted at time k * ; is the expected capacity of the battery energy storage at the future time k + sΔt * ; is the capacity of the battery energy storage at the time k + sΔt predicted at time k *The capacity of the battery energy storage at a certain moment; For the future k + sΔt * The expected capacity of the supercapacitor at the moment; Is the predicted capacity of the supercapacitor at the moment k + sΔt at time k; λ * ; λ b And λ sc Are both weight coefficients of the output power increment of the hybrid energy storage system issued by the MPC controller; ΔP b (k + sΔt * |k) is the change in the output power of the battery energy storage system relative to the previous time period; ΔP sc (k + sΔt * |k) is the change in the output power of the supercapacitor energy storage system relative to the previous time period.

[0141] In this embodiment, since the prediction model is a prediction for a finite time domain, the obtained open-loop optimal solution cannot act on the system at one time. Therefore, a finite time domain "prediction + optimization" strategy of rolling forward in time is adopted, that is, online optimization is solved at each sampling moment, and the first control quantity of the optimal control sequence is applied to the controlled object. At the next sampling moment, the optimization problem is reconstructed and solved to continuously update the control sequence, correct the influence of uncertainty and disturbance on the system, and adjust in real time to achieve better control effects.

[0142] The day-ahead deterministic optimization aims to minimize the system operation cost, and the uncertainty IGDT robust optimization aims to solve the uncertainty within a certain fluctuation range, that is, the low-frequency component of the uncertain factors, under the preset cost target. However, in the intraday optimization and scheduling stage, there are inevitably high-frequency components of the uncertainty of photovoltaic power generation and traction load, which affect the operation state of the traction power supply system. Therefore, it is necessary to continuously feedback and correct the output plan of the hybrid energy storage system and the external power grid supply plan formulated in the day-ahead scheduling stage to achieve the dynamic balance of the intraday traction power supply capacity and load demand. In addition, to ensure the economic benefits of intraday optimal operation, it is required that the intraday operation plan can track the reference operation trajectory set in the day-ahead optimization stage. Therefore, the optimization objectives of the intraday scheduling model of the present invention mainly include the following two parts: i) Minimize the deviation between the actual operation curve (i.e., the incoming power of the traction substation and the state of charge of the hybrid energy storage) and the day-ahead reference trajectory in the intraday stage; ii) Minimize the adjustment amount of the hybrid energy storage output.

[0143] In this embodiment, the commercial solver CPLEX is used to solve the above MPC rolling optimization control problem. The first control quantity u1 of the obtained optimal control sequence is applied to the system to obtain the system state x(k + 1) at the next moment, and the optimization control problem is continuously solved in a loop until the control is completed.

[0144] The expression for the initial value of the finite time domain rolling optimization at the moment k + 1 is:

[0145]

[0146] Among them, is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the lithium battery energy storage at the k-th moment; is the actual value deterministically obtained by the lithium battery energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th time period; is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the ultra-large capacitor energy storage at the k-th moment; is the actual value deterministically obtained by the ultra-large capacitor energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th time period; is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the grid energy storage at the k-th moment; is the actual value deterministically obtained by the grid energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th time period.

[0147] In this embodiment, since the rolling optimization belongs to an advanced control strategy and is inevitably affected by prediction errors, it is necessary to introduce a feedback correction link, that is, to collect the actual operating state of the system before the arrival of each sampling moment and feedback it to the MPC controller, and use the error between the real measurement value and the predicted value to correct the future output of the controller, and correct the charging and discharging plan of the energy storage system within the day and the power purchase plan of the traction substation, so as to achieve the effect of closed-loop control and improve the robustness and anti-interference ability of the system.

[0148] Therefore, after the rolling optimization control is executed at the current k-th moment in the present invention, the actual operating state vector x′(k) of the system is collected before the arrival of the (k + 1)-th moment for feedback correction, which is used as the initial value for the next round of finite-time domain rolling optimization, and the deviation of the operating state of the hybrid energy storage system in the next time period is corrected, so as to form a closed-loop optimization control.

[0149] In this embodiment, taking a traction substation of an electrified railway in the western region as an example, a case study is carried out based on the software platform MATLAB R2019b (with the built-in optimization toolbox YALMIP and the optimization solver CPLEX).

[0150] Taking a traction substation of an electrified railway in the western region as an example, the intra-day optimization time step is set to 15 s, and the prediction horizon and the control horizon are equal and N f = N c= 12. To balance the tracking performance of the incoming power of the traction substation and the state of charge of the hybrid energy storage system, and considering the life loss caused by the frequent adjustment of the battery power, the supercapacitor output is preferentially adjusted to compensate for the imbalance between the source and load supply and demand. The weight coefficients of each deviation tracking in the present invention are respectively set as ω s = 0.6, ω b = 0.35, ω sc = 0.05. Tables 1 - 3 give the parameters of the traction substation, the photovoltaic - hybrid energy storage system, and the electricity charging standard. Among them, the electricity fed back to the power grid adopts the charging scheme of "reverse sending and positive counting", that is, a penalty electricity charge is levied for the electricity returned to the power grid. In addition, the operation and maintenance cost price of the photovoltaic - energy storage system is taken as 0.05 yuan / kWh.

[0151] Table 1 Two - part tariff parameters

[0152] Item Parameter Value Electricity Price per kWh / (yuan / kW·h) 0.55 Demand Price per kW per Month / (yuan / kW·month) 38 Feedback Electric Energy Billing Scheme <![CDATA[c fed = 0.8c buy >

[0153] Table 2 Traction substation parameters

[0154]

[0155]

[0156] Table 3 Hybrid energy storage system parameters

[0157] Hybrid Energy Storage System Parameters Battery Supercapacitor Rated Capacity / MWh 1.5 0.15 Rated Power / MW 2 5 State of Charge Limit [0.2,0.8] [0.05,0.95] Initial / Final State of Charge 0.5 / 0.5 0.5 / 0.5 Charge / Discharge Efficiency 0.8 / 0.8 0.95 / 0.95

[0158] Comparing the incoming power of the traction substation before and after rolling optimization, the simulation results are as Figure 3 shown; comparing the day - ahead reference trajectory and the in - day rolling feedback control results of the state of charge of the battery and the supercapacitor, the simulation results are as Figure 4 shown. It can be seen that the overall change trends of the SOC of the battery and the supercapacitor in - day are the same as the day - ahead reference trajectory, verifying the effectiveness of the proposed rolling optimization control strategy.

[0159] To illustrate the superiority of the day - ahead - in - day rolling optimization scheduling strategy proposed in the present invention, the effect comparison between in - day rolling optimization and non - rolling optimization is given, and the simulation results are as Figure 5 shown.

[0160] Furthermore, the specific operation effect comparison of the traction substation between in - day rolling optimization and non - rolling optimization is given, and the results are shown in Table 1. Among them, the ideal value is defined as the optimized operation result under the condition of 100% accurate prediction of the photovoltaic output and the traction load, and the calculated ideal operation effect. Ave is the average of the fluctuations of the incoming power of the traction substation within the optimization time domain, and the specific calculation method is as follows:

[0161]

[0162] First, from Figure 3 It can be clearly seen from the results that the intraday traction substation incoming power curve (orange line) after rolling optimization has the same trend as the day-ahead reference incoming power trajectory (blue line), that is, the charge and discharge power adjustment plan of the hybrid energy storage system issued by the MPC controller can effectively alleviate the interactive power fluctuation between the traction substation and the external power grid caused by the prediction error of photovoltaic output and the operation deviation of traction load during actual operation in the day. That is to say, the rolling optimization control strategy proposed in the present invention can achieve effective tracking, ensuring the economic benefits during the actual operation of the traction power supply system in the intraday stage, that is, the effectiveness of the strategy.

[0163] Secondly, by comparing Figure 4 in (a) and (b), it can be seen that the overall changing trends of the SOC of the battery and the supercapacitor in the day are the same as the day-ahead reference trajectories, verifying the effectiveness of the proposed rolling optimization control strategy. Specifically, the battery has a better tracking effect and only loses SOC by discharging additionally in some periods, while the capacitor provides additional response capacity by quickly switching the charge and discharge states to suppress the uncertainty effects caused by the short-term fluctuations of photovoltaic output and traction load. Therefore, the change of the capacitor SOC curve is relatively more frequent and drastic than that of the battery, but it also realizes the tracking of the day-ahead reference curve in general trend. This is because the power density of the battery is low. If it is used to respond to the short-term power fluctuations caused by uncertainty factors in the intraday stage, frequent charge and discharge operations need to be carried out at the expense of the battery life. Therefore, in the intraday stage, the supercapacitor with high power density acts as the main role to quickly respond to the short-term and frequent fluctuations of uncertainty factors, while the battery only acts as an auxiliary role to cooperate with the supercapacitor to jointly handle the unbalanced power caused by the random fluctuations of the source-load in the day. This further proves the rationality of the present invention in controlling the tracking weight coefficients of the charge and discharge states of the battery and the supercapacitor in the objective function, and then controlling the additional charge and discharge times of the battery during the intraday operation stage.

[0164] Then, from Figure 5 (a), it can be seen that the intraday rolling optimization can effectively track the day-ahead plan reference value, better suppress the incoming power fluctuation of the traction substation caused by the short-term fluctuations of the source-load in the day, and improve the economic benefits of the intraday plan adjustment of the traction substation. From Figure 5 (b) and (c), it can be seen that compared with no rolling optimization, the active power fluctuations of the left and right power supply arms after intraday rolling optimization are more gentle. This is because the rolling feedback correction model proposed in the present invention makes real-time adjustment and optimization according to the latest operation data and states in each control cycle, and then precisely controls the working mode of the back-to-back converter and the charge and discharge strategy of the hybrid energy storage system, realizing better coordinated cooperation between the two, so as to effectively alleviate the sudden impact of the traction load and the random fluctuation of photovoltaic power generation, and significantly reduce the impact of the short-term impact fluctuations of uncertainty factors on the stable operation of the system in the day.

[0165] Finally, as can be seen from Table 1, after the intraday rolling optimization, the electricity charge and demand charge of the system are respectively 845.84 yuan and 999.97 yuan higher than the ideal cost under accurate prediction. This is because the intraday rolling optimization aims to suppress the short-term random fluctuations of photovoltaic and traction loads, and is essentially an optimization problem in the local time domain. As a result, in some periods, the supercapacitor is difficult to provide fast response capacity to effectively achieve peak shaving and valley filling of traction loads due to its state of charge and rated charge-discharge power limitations, which leads to an increase in the operating cost of the traction substation in the corresponding periods. In addition, since the battery loss of SOC participates in the frequent intraday adjustments in some periods, the battery life loss cost also increases, but it is still within an acceptable range. Compared with the situation without intraday rolling optimization, although the day-ahead and intraday rolling optimization sacrifices a part of the operation and maintenance cost and life loss cost of the energy storage system, the total operating cost is still relatively low. Generally speaking, the total operating cost of the system after intraday rolling optimization is 4.04% higher than the ideal value, but 1.1% lower than the total operating cost without rolling optimization, and the average relative deviation of the traction substation incoming line power is 58.83% lower, indicating that the rolling optimization strategy proposed in the present invention can effectively track the day-ahead operation plan, that is, under the condition of low intraday cost adjustment, it can effectively cope with the high-frequency components of the source-load random fluctuations to a certain extent and achieve the dynamic balance of the traction power supply capacity and load demand on the short time scale within the day.

[0166] In summary, compared with the deterministic model and the method without rolling optimization, the proposed intraday rolling optimization scheduling strategy of the traction power supply system based on MPC in the present invention can, while ensuring the economic operation of the system, predictably adjust the charge-discharge plan of the hybrid energy storage system to cope with the upcoming energy fluctuations, effectively reduce the drastic fluctuations of the substation incoming line power caused by the short-term impact fluctuations of the source-load within the day, with a reduction of up to 58.8%, and achieve the dynamic balance of the traction power supply capacity and load demand during the actual operation within the day.

Claims

1. An intraday rolling optimal scheduling strategy for a traction power supply system based on MPC, characterized in that, Including: Constructing a state - space model of the source - network - load - storage traction power supply system; Presetting constraint conditions, and taking the state - space model of the source - network - load - storage traction power supply system as a prediction model, and taking the minimum tracking deviation as an objective function to construct an MPC controller; Using the MPC controller for intraday rolling optimization. After the rolling optimization control is executed at the current k - th moment, the actual operating state vector of the system is collected for feedback correction, which is used as the initial value for the finite - horizon rolling optimization at the (k + 1)-th moment to correct the deviation of the operating state of the hybrid energy storage system in the next time period, thereby forming a closed - loop optimization control.

2. The intraday rolling optimal scheduling strategy for the traction power supply system based on MPC according to claim 1, characterized in that, The expression of the state - space model of the source - network - load - storage traction power supply system is: Among them, is the state quantity of the source-grid-load-storage traction power supply system at time is the state quantity of the source-grid-load-storage traction power supply system at time is the control quantity of the source-grid-load-storage traction power supply system at time is the output quantity vector of the source-grid-load-storage traction power supply system at time are all coefficient matrices; is the current time; is the interactive power between the train on the left power supply arm and the public power grid at time is the interactive power between the train on the right power supply arm and the public power grid at time is at time the active power generated by the phase converter; is at time the active power generated by the phase converter; is the total power of the battery energy storage system at time is the total power of the supercapacitor energy storage system at time is the capacity of the lithium battery at time is the capacity of the supercapacitor at time is at time the reactive power generated by the phase converter; is at time the reactive power generated by the phase converter; is the transpose; is the difference symbol; is the active power of the phase converter for the control output at time is is the active power of the phase converter for the control output at time is is the reactive power of the phase converter for the control output at time is is the reactive power of the phase converter for the control output at time is is Active power of the train on the left power supply arm at a certain moment; is Active power of the train on the right power supply arm at a certain moment; is Reactive power of the train on the left power supply arm at a certain moment; is Reactive power of the train on the right power supply arm at a certain moment; is Active power of the photovoltaic at a certain moment; is Active power generated by the power grid at a certain moment; is the predicted active power of the train on the left power supply arm at the moment; is the time sampling period; is the number of time sampling periods; is the number of sampling period sequences; is the initial day-ahead prediction value of the active power of the left power supply arm at the is at the moment when the actual value of the active power of the left power supply arm at the moment is calculated; is at the moment when the predicted active power of the train on the right power supply arm at the moment; is the initial day-ahead prediction value of the active power of the right power supply arm at the is at the moment when the actual value of the active power of the right power supply arm at the moment is calculated; is at the moment when the predicted reactive power of the train on the left power supply arm at the moment; is the initial day-ahead prediction value of the reactive power of the left power supply arm at the is at the moment when the actual value of the reactive power of the left power supply arm at the moment is calculated; is at the moment when the predicted reactive power of the train on the right power supply arm at the moment; is the initial day-ahead prediction value of the reactive power of the right power supply arm at the is at the moment when the actual value of the reactive power of the right power supply arm at the moment is calculated; is at the moment when the The active power of photovoltaic at a certain moment; is the initial day-ahead prediction value of photovoltaic power generation at a certain moment; is the predicted photovoltaic power generation at a certain moment; is the total power prediction value of the battery energy storage system at a certain moment; is the power increment of the battery energy storage system at a certain moment; is the total power of the supercapacitor energy storage system at a certain moment; is the power increment of the supercapacitor energy storage system at a certain moment; is the interface power from the power grid to the traction power supply system at a certain moment; is the positive power transmitted from the power grid to the traction power supply system at a certain moment; is the negative power transmitted from the power grid to the traction power supply system at a certain moment; is the total power prediction value of the battery energy storage system at a certain moment; is the discharge power of the battery energy storage system at a certain moment; is the charging power of the battery energy storage system at a certain moment; is the total power of the supercapacitor energy storage system at a certain moment; is the discharge power of the supercapacitor energy storage system at a certain moment; is the charging power of the supercapacitor energy storage system at a certain moment; is the power transmitted to the left power supply arm through the V / v traction transformer at a certain moment; is the power returned from the left power supply arm to the common power grid at a certain moment; is the power transmitted to the right power supply arm through the V / v traction transformer at a certain moment; is the power returned from the right power supply arm to the common power grid at a certain moment; is the power transmitted from the left power supply arm to the converter at a certain moment; is at a certain moment the power transmitted from the converter to the left power supply arm at a certain moment; is at a certain moment The power transmitted by the converter to the right power supply arm; is the power transmitted by the right power supply arm to the converter at the moment.

3. The intraday rolling optimal scheduling strategy for the traction power supply system based on MPC according to claim 1, wherein The constraint conditions are: P grid (k + sΔt * |k) = P tr_L (k + sΔt * |k) + P tr_R (k + sΔt * |k) Among them, P grid (k + sΔt * |k) is the actual active power value of the power grid at the k + sΔt * moment calculated at the k moment; k is the current moment; Δt * is the time sampling period; s is the number of time sampling periods; i is the sampling period sequence number; P tr_L (k + sΔt * |k) is the interactive power between the train on the left power supply arm and the public power grid predicted at the k moment at the k + sΔt * moment; P tr_R (k + sΔt * |k) is the interactive power between the train on the right power supply arm and the public power grid predicted at the k moment at the k + sΔt * moment; is the actual active power value of the train on the left power supply arm calculated at the k moment at the k + sΔt * moment; P α (k + sΔt * |k) is the active power of the α - converter predicted at the k moment at the k + sΔt * moment; is the active power of the train on the right power supply arm predicted at the k moment at the k + sΔt * moment; P β (k + sΔt * |k) is the active power of the β - converter predicted at the k moment at the k + sΔt * moment; P b (k + sΔt * |k) is the active power of the lithium - battery predicted at the k moment at the k + sΔt * moment; P sc (k + sΔt * |k) is the active power of the super - capacitor predicted at the k moment at the k + sΔt * moment; is the active power of the photovoltaic predicted at the k moment at the k + sΔt * moment; Q tr_L (k + sΔt * |k) is the actual power value of the left power supply arm of the traction substation calculated at the k moment at the k + sΔt * moment; is the reactive power of the train on the left power supply arm predicted at the k moment at the k + sΔt * moment; Q α (k + sΔt * |k) is the reactive power of the α - converter predicted at the k moment at the k + sΔt * moment; Q tr_R (k + sΔt * |k) is the actual power value of the right power supply arm of the traction substation at the moment of k + sΔt * ; is the reactive power of the train on the right power supply arm predicted at the moment of k for k + sΔt * ; Q β (k + sΔt * |k) is the reactive power of the β converter predicted at the moment of k for k + sΔt * ; is the capacity of the battery energy storage predicted at the moment of k for k + sΔt * ; is the capacity of the battery energy storage predicted at the moment of k for k + (s - 1)Δt * ; η b is the charge and discharge efficiency of the lithium battery; is the charge and discharge power of the battery energy storage predicted at the moment of k for k + sΔt * ; is the rated capacity of the battery energy storage; is the capacity of the supercapacitor predicted at the moment of k for k + sΔt * ; is the capacity of the supercapacitor predicted at the moment of k for k + (s - 1)Δt * ; η sc is the charge and discharge efficiency of the supercapacitor; is the charge and discharge power of the supercapacitor predicted at the moment of k for k + sΔt * ; is the rated capacity of the supercapacitor; η b,ch is the charging efficiency of the battery energy storage; is the total power prediction value of the battery energy storage system at the moment of k; η b,dis is the discharge efficiency of the battery energy storage; η sc,ch is the charging efficiency of the supercapacitor; is the total power of the supercapacitor energy storage system at the moment of k; η sc,dis is the discharge efficiency of the supercapacitor; is the minimum capacity of the battery energy storage; is the maximum capacity of the battery energy storage; is the minimum capacity of the supercapacitor; is the maximum capacity of the supercapacitor; is the rated power of the battery energy storage system; P b (k) is the total power of the battery energy storage system at the moment of k; ΔP b (k + sΔt * |k) is the power increment of the battery energy storage system within the future sΔt * time period; is the rated power of the supercapacitor energy storage system; P sc (k) is the total power of the supercapacitor energy storage system at time k; ΔP sc (k+sΔt * |k) is the future sΔt * The power increment of the supercapacitor energy storage system during the time period; To calculate k+sΔt at time k * The actual value of active power of the power grid at the moment; M g delivering capacity limits for transmission lines; k+sΔt predicted at time k * The power delivered by the traction transformer to the left supply arm train at all times; k+sΔt predicted at time k * The power delivered by the traction transformer to the right supply arm train at all times; is the active power of the left power supply arm train at time k; is the active power of the right power supply arm train at time k; a i 、b i and c i are the coefficients of the regular dodecagon linearized power flow constraints; is the rated capacity of the α converter; is the rated capacity of the β converter; k+sΔt predicted at time k * The power of the left power supply arm of the traction substation at all times; is the upper limit of voltage imbalance; S d is the short-circuit capacity of the external power grid; k+sΔt predicted at time k * The power of the right power supply arm of the traction substation at this moment; Δ is the sign of the difference.

4. The intraday rolling optimal scheduling strategy for the traction power supply system based on MPC according to claim 1, characterized in that The expression of the objective function is: J = min{A1 + A2} Among them, J is the objective function; A1 is the deviation of the incoming power of the traction substation and the state of charge of the hybrid energy storage system; A2 is the change in the output power of the energy storage; N is the set of s; s is the number of time sampling periods; ω s is the weight coefficient of the incoming power of the traction substation in the day-ahead stage; f1 is the weighted error term of the square of the deviation of the grid power; ω b is the weight coefficient of the state-of-charge tracking error of the battery energy storage system; f2 is the weighted error term of the square of the deviation of the state of the energy storage system; ω sc is the weight coefficient of the state-of-charge tracking error of the supercapacitor system; f3 is the weighted error term of the square of the deviation of the state of the supercapacitor; is the expected power of the grid at the future time k + sΔt * ; Δt * is the time sampling period; is the active power of the grid predicted at time k for the time k + sΔt * ; is the expected capacity of the battery energy storage at the future time k + sΔt * ; is the capacity of the battery energy storage predicted at time k for the time k + sΔt * ; is the expected capacity of the supercapacitor at the future time k + sΔt * ; is the capacity of the supercapacitor predicted at time k for the time k + sΔt * ; λ b and λ sc are both weight coefficients of the increment of the output power of the hybrid energy storage system issued by the MPC controller; ΔP b (k + sΔt * |k) is the change in the output power of the battery energy storage system relative to the previous time period; ΔP sc (k + sΔt * |k) is the change in the output power of the supercapacitor energy storage system relative to the previous time period.

5. The intraday rolling optimal scheduling strategy for the traction power supply system based on MPC according to claim 1, characterized in that The expression of the initial value for the finite - horizon rolling optimization at the (k + 1)-th moment is: Among them, is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the lithium battery energy storage at the k-th moment; is the actual value deterministically obtained by the lithium battery energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th period; is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the ultra-large capacitor energy storage at the k-th moment; is the actual value deterministically obtained by the ultra-large capacitor energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th period; is the initial value at the (k + 1)-th moment of the predicted output of the MPC controller after rolling optimization and execution of the grid energy storage at the k-th moment; is the actual value deterministically obtained by the grid energy storage according to the actual data of the photovoltaic output and the traction load before the arrival of the (k + 1)-th moment after the MPC controller issues a control instruction in the k-th period.

Images