ENERGY MANAGEMENT STRATEGY BASED ON DYNAMIC PROGRAMMING FOR DUAL SOURCE TROLLEYBUS

Transcription

1 Liwei Zhang i dr Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja ISSN (Print), ISSN (Online) ENERGY MANAGEMENT STRATEGY BASED ON DYNAMIC PROGRAMMING FOR DUAL SOURCE TROLLEYBUS Liwei Zhang, Yang Yang, Minghe Sun, Hui Liu Original ientific paper The dual source trolleybus of a new odel with battery and super capacitor energy storage syste was designed The related paraeters of dynaical syste were designed on the basis of theory deduction, and the siulation odel was re-developed in ADVISOR software Based on the analysis of work odes of the dual source trolleybus, an optial control strategy based on Dynaic Prograing was designed, which is aied at reducing grid peaking power of on-line ode and energy consuption of off-line ode In addition, the rule based control strategy has also been ipleented and copared to the optiized control strategy The results have deonstrated that the optiization techniques based on Dynaic Prograing can ake full use of the supercapacitors with high power and fast charging/diharging characteristics, and ore effectively distribute the power between the power sources, iprove the dynaic perforance and econoic perforance of the trolleybus Keywords: dual source trolleybus; supercapacitors; energy anageent; optial control; Dynaic Prograing Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja Izvorni znanstveni članak Dizajniran je novi odel trolejbusa s dvojni izvoro napajanja s akuulatoro i super električni kondenzatoro, a s ti povezani paraetri dinaičkog sustava dizajnirani su na teelju teorije dedukcije Siulacijski odel razvijen je u ADVISOR softveru Zasnovana na analizi radnih odova trolejbusa s dvojni sustavo napajanja, dizajnirana je optialna strategija upravljanja na teelju dinaičkog prograiranja čiji je cilj reduciranje vršne snage reže on-line oda i potrošnje energije off-line oda Uz to, upravljačka strategija zasnovana na to pravilu također je ipleentirana i uspoređena s poboljšano strategijo upravljanja Rezultati su pokazali da etode optiizacije teeljene na dinaičko prograiranju ogu u potpunosti iskoristiti super električni kondenzator velike snage, s brzi punjenje i pražnjenje, i učinkovitije rasporediti energiju eđu izvoria energije te poboljšati dinaičke i ekonoske karakteristike trolejbusa Ključne riječi: trolejbus s dvojni izvoro napajanja; super električni kondenzator; upravljanje energijo; optialno upravljanje; dinaičko prograiranje 1 Introduction In recent years, with the strong support of the governent, urban public transport has developed rapidly However, with the urban environental pollution, energy shortages and other issues further intensified, the traditional fuel power bus has been unable to eet the needs of energy-saving eission reduction, so the state began to vigorously proote the new energy bus [1] Nowadays, with the developent of rechargeable battery technology and energy-saving environental protection policy, dual-source trolley bus has been rapidly developed Dual-source trolley bus s proinent features are the use of rail and its own power battery installed to generate electricity through dual-source In the cable network section, the network can generate electricity for the bus and charge the battery at the sae tie, while in the wireless network section, the bus can rely on the power stored in the battery [2, 3] Therefore, not only does the dual-source trolleybus solve the probles of traditional trolleybus controlled by the network layout, which has inherent flaws in energy saving and environental protection, but also those of new energy bus Trolleybus driving conditions are ore coplex, especially in the urban environent Frequent start, acceleration or deceleration often produces a large peak power to the grid causing a lot of load pressure However, the existing power battery cannot withstand the vehicle acceleration or braking generated by the peak power, and in the regeneration brake process, the battery is also difficult to withstand the brake feedback pulse current, causing the waste of braking energy Besides, the pulse current will seriously daage the lithiu-ion battery and reduce its life [4] With the rapid developent of energy storage technology, super capacitor is a new energy storage device with higher power, good safety perforance, long cycle life and other outstanding advantages Therefore, this paper uses lithiu battery - super capacitor hybrid energy storage syste as the auxiliary power source of trolleybus Giving full play of lithiu batteries and super capacitors in the specific energy and specific power of their respective advantages, can not only reduce the peak power grid and lithiu battery power, reduce the charge and diharge frequency of lithiu battery, which will extend its life, but can also provide recovery of vehicle braking generated by the regenerative kinetic energy efficiently, thus iproving vehicle energy efficiency [5] There are two kinds of power sources in the dualsource trolleybus: online operating ode and off-line operating ode To ake the two power sources coordinated, a reasonable vehicle energy anageent strategy is required to coordinate the distribution of power requireents between the different power sources At present, ore control strategies are based on the rule based control algorith According to the engineering experience and experient, the working area of the working parts and the control rules of the vehicle are deterined [6] This control ethod relies on the accuracy of the rules can t ake the perforance of the vehicle to the best Dynaic prograing ethod is a global optiization algorith, which is suitable for the optial control proble under ulti-constrained and nonlinear conditions It has the characteristics of theory copleteness, applicability and easy engineering, which has becoe a hotspot in recent years [7, 8] Based on the odel of vehicle siulation, this paper ais to reduce the Tehnički vjesnik 24, 5(2017),

2 Energy anageent strategy based on dynaic prograing for dual source trolleybus peak power and vehicle energy consuption of the power grid The dynaic prograing ethod is used to obtain the energy anageent strategy under specific working conditions, and it is optiized and copared with the traditional control strategy by eans of siulation coparison analysis Table 1 Paraeters of dual source trolleybus Paraeter Unit Value Full load quality t 16 Windward area 2 7 Maxiu speed k/h 55 Rolling resistance coefficient Air resistance coefficient Wheel radius Dual source trolleybus structure and paraeter atching Dual-source trolleybus is powered by the grid and auxiliary energy storage devices, the power syste structure and the paraeters of the coponents have to eet the need of vehicle's dynaic perforance and the standard of driving range According to the design Liwei Zhang et al requireents, the selected dual source trolleybus paraeters are shown in Tab 1 21 Syste structure The hybrid energy storage device, which is coposed of super capacitor and lithiu battery, serves as trolley bus auxiliary power source is shown in Fig 1 Power syste is consists of grid, lithiu battery, super capacitor, DC/DC converter, inverter, otor and echanical transission syste The power source of the vehicle is the power grid, battery and super capacitor The power supply is provided by the pantograph for the vehicle The rated voltage of the power supply in Beijing is DC 650V The lithiu battery is directly connected to the DC bus and can be charged through the grid The super capacitor is connected to the DC bus through a bidirectional DC/DC converter When the tra is traveling on a section without overhead contact, the energy is supplied by the energy storage syste The inverter converts the regulated DC voltage into an AC voltage to drive the AC otor The drive syste is used to change the speed ratio and adjust the speed of the output to the wheel Substation Inverter Wheel AC power supply Battery Braking resistor M Mechanical transission Rectifie r bridge Bidirectional DC/DC converter Electric drive syste S Super capacitor Figure 1 The structure of dual source trolleybus 22 Motor paraeters In the process of trolley bus operating, the response power of the vehicle's deand power is supplied by driving the otor Therefore, it is necessary to deterine the power of the otor according to the vehicle paraeters and the dynaic requireents first The rated power of the otor shall be greater than the deand power of the vehicle at the axiu speed which is: P ( crvgvax + rvax ACd ) (1) η 2 T Where P is the otor rated power, c r is the rolling resistance coefficient, v is the vehicle ass, g is the gravitational acceleration, ρ is the air density, v ax is the axiu vehicle speed, A is the windward area of the vehicle, C d is the air resistance coefficient, η T is powertrain efficiency The peak power of the otor should be greater than the required power for the axiu acceleration of the vehicle, which is: P ( vaaxv + crvgv + rv ACd ) (2) η 2 T Where, P _ax is the otor peak power, a ax is the axiu acceleration of the vehicle The otor speed n and torque T e are calculated as follows: n T e vig = (0, 377r) 9550P = n (3) 1440 Technical Gazette 24, 5(2017),

3 Liwei Zhang i dr Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja Where i g is the transission ratio, r is the wheel radius The existing bus transission ratio is 433, 4875, 5134, 557 and so on The iniu transission ratio i g_in should eet the requireents of the axiu speed of the trolleybus, calculated and cobined with the actual situation of peranent agnet synchronous otor to deterine the rated speed of the drive otor 2000 r/in, and the axiu speed 6000 r/in Based on the above analysis, this paper chooses MC_AC187 otor in ADVISOR siulation software to eet the vehicle power deand The specific paraeters are shown in Tab 2 Table 2 Paraeters of MC_AC187 otor Maxiu power (kw) Rated power (kw) Maxiu torque (N ) Rated torque (N ) Maxiu rotating speed (r/in) Rated rotating speed (r/in) Capacity optiization of energy storage syste The paraeter design of the energy storage syste is a coproise optiization process, that is, in addition to eet the basic energy and power requireents of the vehicle,we should also iniize the input cost [9-12] As for the dual source trolleybus, the access line running can be achieved by online charging, therefore the vehicle travel distance and off-line travel distance directly affect the energy storage syste capacity configuration Cobined with the characteristics of lithiu batteries and super capacitors, lithiu batteries prefors as the ain energy source eet the vehicle capacity requireents, while super battery as a secondary energy source eets the power needs of vehicles Therefore, this paper not only uses the idea of ain and auxiliary energy source capacity and power independent design, but also considers the off-line coparison of the running route of trolleybus to design the paraeters of hybrid energy storage syste Tab 3 is the Paraeters of the selected battery and super capacitor Table 3 Paraeters of battery and super capacitor Paraeter Lithiu battery Super capacitor Model (V) Saft_LSH14 Maxwell_PC Rated voltage (Wh/kg) Power density (kw/kg) Monoer quality (kg) Monoer price (yuan) At an average speed v a = 30 k/h, the average energy consuption per kiloetre of the trolleybus Q a is: Q a Pa = = ( crvg + rva ACd ) = 15, kw h/k v h 2 a T (4) Where L is the full distance travelled, D is the off-line travel distance In order to achieve the axiu use of wire network, in addition to charging during the grid section, there is no need to spend additional parking tie to charge the lithiu battery So the energy consued in the off-line section should be copensated in the cable section Running tie for a single bus is generally 2 hours, when it is considered that the axiu cycle is 7 ties a day, that the balance of the lithiu battery is 30%, and that a single cycle of energy difference is 10%, there is forula: D X (7) D ax It can be drawn that the condition lithiu battery capacity should eet is as follow: 90D E batt (8) L D + 6 E batt Taking the off-line ratio k = D/L, there is: 90k 1 k + 6 L (9) Fig 2 is a three-diensional relationship diagra which shows the relationship aong off-line ratio k, travel distance L and lithiu battery iniu capacity E in, we can see that the greater the off-line ratio, the greater the required lithiu battery capacity Set the lithiu battery capacity as E batt, if the lithiu battery is fully charged, the axiu off-line ileage D ax on theory of the vehicle is: Ebatt 2 D ax = = Ebatt (5) Q 3 a The charging rate of lithiu battery by network is 05C, the vehicle at an average driving speed under the charge energy percentage X is: Figure 2 The diensional diagra of k, L and E in X L D L D = = (6) Taking the 103 trolleybus in Beijing as an exaple, the full travel distance L = 267 k, off-line distance D = 69 k, off-line ratio k = 026, the iniu capacity of Tehnički vjesnik 24, 5(2017),

4 Energy anageent strategy based on dynaic prograing for dual source trolleybus lithiu battery is 239 kw h, so the aount of lithiu battery onoer is at least n 1 = 239/0032 = 7468 Taking the ethod of lithiu battery connection into account,we use 125 onoer in series, the rated voltage is U b_no = V = 450 V, and then take 3 series in parallel, thus n 1 = 750 Super capacitor has to eet the requireents of energy and power when it coes to the conditions that the vehicle has the axiu acceleration According to the otor peak power derived above, the super capacitor power has to eet: P_ax P = 208 kw (10) η Aong the, η is the otor efficiency Super capacitor energy to eet the vehicle t = 10 s peak power, is: E P_ax t = 0 58 kw h h (11) Considering the power and energy requireents, the aount of supercapacitor onoers is n 2 = 190 Using a single series connection, the super capacitor group rated voltage U _no = V = 475 V When considering the power output capacity and energy storage capacity of the energy storage syste, the aount and cost of the single unit in the single lithiu battery storage ode and the ixed energy storage ode are shown in Tab 4 Energy storage ethod Lithiu battery Mixed energy storage Table 4 Cost coparison of different energy storage Lithiu battery aount Super capacitor aount Total cost (Million yuan) As can be seen fro Tab 4, due to its low density, when perforing as the vehicle energy storage device, lithiu battery syste costs ore than ixed energy syste when there is a need to eet the requireent of vehicle power as well as that of storage at the sae tie The super capacitor in hybrid energy storage syste is responsible for providing high deand for vehicle power Lithiu battery is responsible for eeting the requireents of the vehicle driving range Therefore, the use of hybrid energy storage syste can reduce costs and iprove econoic efficiency Liwei Zhang et al perforance can optiize the ais The control requireents are: to ensure the full power of the vehicle under the preise of giving full play to the advantages of super capacitor, such as, high energy density, charge and diharge in a fast speed when there is a high current In the on-line operating ode, reduce the peak power of the grid; in the off-line operating ode, reduce the ipact of high current on the battery, iprove the charge and diharge efficiency, extend the battery life, axiize the recovery of braking energy and iprove vehicle econoy The global optiization anageent strategy is a ulti-source syste energy allocation strategy developed by the application optiization ethod and the optial control theory The strategy includes linear prograing ethod, nonlinear prograing ethod, dynaic prograing ethod, Multi-target tracking algorith and genetic algorith Genetic algorith is a coputational odel for siulating the natural selection and genetic echanis of Darwin's theory of biological evolution It is a ethod to search the optial solution by siulating the natural evolutionary process Arango et al proposed a atheatical odel and developed a heuristic procedure based on a genetic algorith to solve non-linear probles Nishiura et al developed a heuristic procedure based on a genetic algorith to obtain a good solution with considerably sall coputational effort This algorith can also be used to solve the proble of production coordination [13, 14] Multi-target tracking algorith is based on depth learning First, the GoogLeNet + LSTM odel is used for target detection to obtain accurate target detection results On the basis of this, a ethod of extracting the depth feature directly fro the feature ap based on the target detection ethod is proposed The traditional feature of the depth feature can reflect the appearance of the detection target ore accurately than the traditional feature such as ale invariant feature transfor (SIFT), so it can iprove the accuracy of the target tracking algorith [15] 31 Analysis of operation ode of dual - source trolleybus According to the difference aong power supply for energy work of dual-source trolleybus, the operating ode of dual-source trolley bus is divided into on-line operating ode and off-line operating ode, as is shown in Fig 3 On-line operating Off-line operating On-line operating 3 Energy anageent strategy based on dynaic prograing The energy anageent strategy is an indispensable part of efficient operation of dual-source trolleybus It allocates the power between the grid and the energy storage equipent according to the power deand of the vehicle under different road conditions, so that the vehicle Figure 3 The operating ode of dual source trolleybus Fig 4 shows that the power flow of the trolleybus is in the on-line operating ode At this tie, the trolleybus 1442 Technical Gazette 24, 5(2017),

5 Liwei Zhang i dr Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja is connected with the catenary, the vehicle is powered by the power grid and the super capacitor The super capacitor is connected with the DC bus through the bidirectional DC/DC converter, power relationship is shown in Eq (12) Aong the, P req is the deand power of the vehicle, P grid is the power of grid, P is the power of the super capacitor When the vehicle is running in the traction ode, the grid provides steady power required for noral driving When the vehicle accelerates, the super capacitor provides the different power of the power grid and the deand power, and reduces the average power of the grid output When the vehicle decelerates while braking, the super capacitor charging, recycling regenerative braking energy In addition, if the battery charge state is lower than the off-line driving deand, the grid needs to charge the battery to ensure the energy needs of off-line operating ode P t) = P ( t) + P (12) req ( grid the field of operational research It is a nuerical algorith based on Bellan's optiization principle for solving ulti-stage decision probles [16, 17] The ulti-stage decision proble is divided into N stages by tie At each stage, it is necessary to ake a decision, and the decision variable u k of the given stage is deterined The state variable x k of the k + 1 stage x k+1 is also deterined, the k stage to k + 1 stage of the state transfer law is called state transition equation, expressed as: x k + 1 = fk ( xk, uk ) (14) where k = 0, 1,, N 1 The control vector of the syste is: { u u,, u } u U 0, 1 N 1 k k u = (15) The goal of dynaic prograing is to find the best control strategy so that the su of the objective function values at each stage is iniized The value function starting with state x 0 is as follow: N 1 J ( x ) = L( x, u ) (16) u 0 k =0 k k Figure 4 Power flow of on-line operating ode In off-line operating ode, the lithiu battery is the ain energy source, the super capacitor is the auxiliary energy source, and the power flow is shown in Fig 5 The power flow is the sae as that of the pure electric vehicle The power relationship is shown in Eq (13), where P batt is the power of battery When the vehicle is running in the traction ode, the power battery and the super capacitor cooperate to provide energy for driving the otor; when the vehicle decelerates while braking, the otor is in the power generation state, the regenerative braking energy is given to the super capacitor through the DC/DC converter, the excess part of the energy is used to charge the lithiu battery P t) = P ( t) + P ( t) (13) req ( batt Figure 5 Power flow of off-line operating ode 32 The principle of dynaic prograing Dynaic prograing ethod is one of the tools to solve the global optiization proble, which belongs to Aong the, L(x k, u k ) is the state between x k and x k+1, and the optial value function obtained by optiizing the calculation is the function that iniizes the total value, and the corresponding control vector is the best control vector The central idea of the Bellan principle is that if the optial solution path of the proble passes through an interediate state, then the optial solution fro the state to the end of the process is the continuation of the optial solution path of the whole process [18] A A B C D8 B 16 C17 D 17 E F G 6 a) E 13 F 14 G 14 H 9 I 8 J 8 H 9 I 8 b) Figure 6 Shortest path optiization proble Fig 6 is the diagra of shortest path optiization proble, which calculates the shortest path fro point A to point K, the nuber of stages is N = 5, the figure J 8 K K Tehnički vjesnik 24, 5(2017),

6 Energy anageent strategy based on dynaic prograing for dual source trolleybus between the two lines of the edge of the line is the distance fro the previous point to the next point If the state points of the process are not too any, it is very easy to find the iniu tie path by considering all the possible paths, but it is not possible to use this ethod if there are too any state points And the use of optiization algorith can reduce the need of considering the nuber of paths The first step in applying the dynaic prograing algorith is to calculate the consuption value of the state changes between two consecutive steps As shown in Fig 6(a), all the values between any two states are shown Once all values between any two states are deterined, the iniu value of all state points is reversed to the end point At stage N = 4, the iniu value of each state point to end point K is the corresponding inter-state value In the case of N = 3, there are three paths to the end point K for point F, which are F H K, F I K, F J K, and the corresponding distances are 15, 14 and 16, respectively The best path fro the state point F to the state point K is the path of the state point I, with the shortest path length 14 Siilarly, the best path for point E to point K is E H K, the iniu distance is 13, and the best path fro point G to point K is G J K, and the iniu distance is 14 After calculating these figures, the algorith oves to N = 2 to calculate the iniu value of state points B, C, and D According to the Bellan optiization principle, the optial path fro the state points E, F and G to K is not affected by the previous tie step, so the iniu value fro B to K is the state fro B to E, F and G The value plus the iniu value fro these points to K, for exaple, the iniu value of path B E K is equal to the value of state between B and E plus the iniu value of E to K In the sae way, we can calculate the iniu value of all the state points as shown in Fig 6(b) and the optial path with the shortest distance of 20 fro A to K The path passes through the state points B, F and I 33 Dynaic planning process The goal of energy anageent of dual-source trolleybus is to realize the reasonable allocation of power between the energy sources This process is diretized by tie and can be regarded as a ulti-stage decision proble The dynaic prograing ethod can be used [19, 20] When the dual-source trolleybus is in on-line operating ode, the power is supplied fro the grid to the vehicle The super capacitor replenishes the peak power and recovers the braking energy The dynaic prograing ethod is used to optiize the power distribution between the grid and the super capacitor The goal is to reduce the peak of the grid power P grid ax ; in the off-line operating ode, the dynaic prograing ethod is used to reasonably allocate the power between the battery and the super capacitor, the evaluation index is the vehicle energy consuption ECR, the forula is as follows E _ J ECR = (17) L t Liwei Zhang et al E _ J = ( P + P )dt (18) 0 batt Where ECR is the vehicle energy consuption per 100 k; E_J is the energy consued in the vehicle travelling process; L is the travel distance of the vehicle, is the unit conversion factor The SOC of the super capacitor is the state variable in both odes, and the control variable is the electric power P of the super capacitor, as shown in the following equation: { SOC t) }, u { P } x = = (19) ( The relationship between the syste state variable and the control variable is: SOC t P ( t)dt 0 ( t) = SOC0 + E (20) Where E is the axiu capacity of the super capacitor, and the charge state transition equation is obtained by deriving and diretizing both sides of the above equation: P SOC k + 1 = SOCk + ts, k = 1,, Nt 1 (21) E Where t s is the sapling tie and N t is the nuber of sapling points For the syste state variable SOC, it can be diretized into N s values between the axiu and iniu values, as shown in the following equation: SOCax SOCin SOC i = SOCin + ( i 1), i = 1,, Ns (22) N 1 For the battery and the super capacitor, in order to extend their service life, their SOC should work within a certain allowable range during working process, naely: SOC SOC s ( t) SOCbatt ( t) SOCbatt ax ( t) (23) t) SOC ( t) SOC ( t) (24) batt in in ( ax If the SOC value is too high, the trolleybus is hard to recover renewable energy, which will produce energy waste; on the other hand, if the SOC value is too low for a long tie, it will greatly reduce the battery life, and affect the vehicle acceleration perforance In actual operation, we set the SOC of the battery and super capacitor in the range of [03, 09] 34 Dynaic Prograing reverse order ethod Based on the Bellan principle, we can see that if the interediate state to the end state of the optial path of the whole process is regarded as a subroutine, then the 1444 Technical Gazette 24, 5(2017),

7 Liwei Zhang i dr Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja optial path of the subroutine is the latter part of the optial process of the whole process The dynaic prograing algorith is an optiization ethod of reverse calculation The recursive equation of the reverse order ethod is: J J SOCi ) ( SOC in ) = 0 ( L ( x, u ) + J ( SOC )) k ( = k k k k + 1 N t Nt j, (25) Where J k (SOC) represents the iniu value consued fro tie k to the final tie along the optial path, and L k (x k, u k ) is the value consued by state x k to x k+1 Fig 7 is the flow chart of dynaic prograing reverse order ethod, roughly divided into the following three steps: 1) Initialize, the vehicle driving process is divided into N stages, and fro the N th stage to calculate the iniu value fro this stage to the final state; (2) For the kth stage, the optial decision of the kth stage is obtained for the value of Eq (15) u j under the state variable x i of this stage (3) If k = 1, then go to (4), otherwise, k = k 1, go to (2); (4) When k = 1, the optial decision of the whole cycle is obtained axiu output power of the grid is 272 kw, and the axiu output power of the power grid is 217 kw under the dynaic prograing ethod Figure 8 Rule based control siulation wavefor of on-line operating ode Initialization For k=n:1 no no For x=x in:x ax j=1 While u j<u ax no Figure 9 Siulation results of off-line situation with the Dynaic Prograing based control strategy Cost Function Min(Cost Function) j=j+1 End Figure 7 Backward induction based flow chart for dynaic prograing 4 Siulation analysis In order to verify the validity of the proposed strategy, the siulation odel of dual source trolleybus was developed twice in the ADVISOR software platfor The dynaic prograing algorith was written on the online and off-line operating odes with CYC_NYCC cycle condition as road load, and then copared with siple rule based control siulation Aong the, the axiu value of the SOC of the battery and the super capacitor is set to 09, the iniu value is set to 03, the initial value is 09, and the tie diretization step is set to 1 s Fig 8 and Fig 9 are the siulation wavefors of CYC_NYCC operating conditions under the on-line operating ode under the rule based control and dynaic prograing control Under the rule based control, the Figure 10 Rule based control siulation wavefor of off-line operating ode Fig 10 and Fig 11 are the siulation wavefors with CYC_NYCC operating conditions under the off-line operating ode of rule based control and dynaic prograing based control, Fig 10 is the wavefor under the cut-off frequency control, fro 545 s to 552 s, the power of lithiu battery is higher than 80 kw, and under the dynaic prograing based control, the power of lithiu battery can be controlled below 50 kw Tehnički vjesnik 24, 5(2017),

8 Energy anageent strategy based on dynaic prograing for dual source trolleybus Liwei Zhang et al 6 References Figure 11 Dynaic prograing based control siulation wavefor of off-line operating ode Tab 5 shows the siulation results under two control ethods In the on-line operating ode, the peak power of the dynaic prograing control is reduced by 202% copared with that of rule control In the off-line operating ode, the power consuption per 100 k of the vehicle under the dynaic prograing based control is reduced by 5% copared with the rule based control Table 5 Coparison of siulation result of CYC_ NYCC driving cycle Line operation Offline operation 100 Control Maxiu power of k of power Strategy the grid consuption Dynaic planning 217 kw 1437 kw h/100 k Rule control 272 kw 1512 kw h/100 k Coparing results Reduced by 202% Reduced by 5% 5 Conclusion This paper studies the structure of the battery - super capacitor power syste for dual source trolleybus The paraeters of the otor, lithiu battery and super capacitor were designed on the basis of theory deduction, the siulation odel was re-developed in ADVISOR software and finally established the siulation odel for dual source trolleybus In order to reduce the peak power of the grid and the energy consuption during the off-line operating ode, the global optiization energy anageent strategy based on the dynaic prograing ethod was designed, and copared with the siple rule based control strategy The results showed that the energy anageent strategy based on dynaic prograing can ake better use of supercapacitor to reduce the peak power of the power grid and the battery, and effectively recover the braking capacity and iprove its econoic perforance It proves that the dynaic prograing ethod used to carry out energy optiization anageent is feasible Acknowledgeents This work was financially supported by National Key Research and Developent Progra (2016YFB B22) and Fundaental Research Funds for the Central Universities (2017JBM060) [1] Ling, Tianjun Energy saving and new energy vehicle technology developent status and infrastructure construction // Auto Electric Parts 10( 2011), pp 1-10 [2] Liu, Kaijun; Zhao, Ruliang The Application and Prospect of Dual Source Trolleybus // Urban Public Transport 12(2012), pp 9 [3] Sun, Fengchun; Bin, Liu; Wang, Zhenpo Analysis of energy consuption characteristics of dual-source trolleybus // IEEE Transportation Electrification Conference and Expo, 2014, pp1-5 [4] Devie, A; Venet, P; Pelissier, S; Trigui, R Battery duty profile of a heavy-duty trolleybus // IEEE Vehicle Power and Propulsion Conference, 2012, pp [5] Cao, J; Eadi, A A new battery/ultracapacitor hybrid energy storage syste for electric, hybrid, and plug-in hybrid electric vehicles // IEEE Transactions on Power Electronics, 2012, pp [6] Naxin, Cui; Juanjuan, Fan; Chenghui, Zhang; Jian, Wu Research on predictive control based energy anageent strategy for Hybrid Electric Vehicle // IEEE International Syposiu on Power Electronics for Distributed Generation Systes, 2012, pp [7] Cui, Naxin; Bu, Gang; Wu, Jian; Fu, Xiaoling; Zhang, Chenghui Real-TieOptiization of Energy Manageent Strategy forplug-in Parallel Hybrid Electric Vehicles // Transactions of China Electrotechnical Society 26, 11(2011), pp [8] Wang, Qi; Sun, Yunkun; Huang, Yonghong Research on a Distribution Strategy of Braking Force Used in Hybrid Electric Vehicles with Battery-Ultracapacitor Hybrid Energy Storage Syste // Transactions of China Electrotechnical Society 29, 1(2014), pp [9] Xie, Shixiao; Yang, Li; Li, Lina A ChanceConstrained Prograing Based Optial Configuration Method of Hybrid Energy Storage Syste // Power Syste Technology 5(2012), pp [10] Zhang, Chunjiang; Dong, Jie; Liu, Jun; Ben, Bing A Control Strategy for Battery-Ultracapacitor Hybrid Energy Storage Syste // Transactions of China Electrotechnical Society 29, 4(2014), pp [11] Zhu, Xiangfen Capacity Optiization of Hybrid Energy Storage Syste Based on Particle Swar Optiization // Ningxia University, 2014 [12] Chang, Fengqi; Zheng, Zedong; Li, Yongdong A Novel Hybrid Energy Storage Topology and its Power Sharing Algorith // Transactions of China Electrotechnical Society 30, 12(2015), pp [13] Tang, M; Gong, D; Liu, S; Zhang, H Applying ultiphase particle swar optiization to solve bulk cargo port heduling Proble // Advances in Production Engineering & Manageent 11, 4(2016), pp [14] Gong, D; Tang, M; Liu, S Reconsidering Production Coordination: A Principal-Agent Theory Based Analysis // Advances in Production Engineering & Manageent 12, 1(2017), pp [15] Lu, Ping; Deng, Shuo; Li, Weihua Multiple Object Tracking Algorith Based on Deep Learning // ZTE Technology Journal 23, 4(2017), pp [16] Zhang, Hao; He, Jinghan; Bo, Zhiqian; Hu, Wei; Li, Ke; Zhou, Wen Service Restoration Based on Dynaic Prograing // Transactions of China Electrotechnical Society 26, 12(2011), pp [17] Bellan, R Bottleneck probles and dynaic prograing // In Proceedings of the National Acadey 1446 Technical Gazette 24, 5(2017),

9 Liwei Zhang i dr Strategija upravljanja energijo zasnovana na dinaičko prograiranju za trolejbus s dvojni izvoro napajanja of Sciences of the United States of Aerica 39(1953), pp [18] Bellan, R Dynaic Prograing // Princeton University Press, 1957 [19] Li, Wen; Zhang, Cheng-ning Power Manageent of Parallel Hybrid Electric Power Train // Journal of Shanghai Jiaotong University (Science) 1(2013), pp [20] Perez, L V; Bossio, G R; Moitre, D et al Optiization of power anageent in an hybrid electric vehicle using dynaic prograing // Matheatics and Coputers in Siulation 73(2006), pp Authors addresses Liwei Zhang, Associate Professor (Corresponding Author) Beijing Jiaotong University No 3 Shangyuancun, Haidian District, Beijing , China E-ail: lwzhang@bjtueducn Yang Yang, Master Candidate Beijing Jiaotong University No 3 Shangyuancun, Haidian District, Beijing , China E-ail: @bjtueducn Minghe Sun, Doctoral Candidate Beijing Jiaotong University No 3 Shangyuancun, Haidian District, Beijing , China E-ail: hsun@bjtueducn Hui Liu, Engineer ZTE Corporation Nanshan District, Science and Technology South Road, Shenzhen, China E-ail: liuhui10@ztecocn Tehnički vjesnik 24, 5(2017),