TW200411201A - Apparatus and method for searching most suitable charging curve of battery - Google Patents

Abstract

The present invention provides an apparatus and method for searching most suitable charging curve of battery. A programmable battery charging/discharging module is used to perform multiple multi-stage charging tests to a charging battery. The test uses the battery charging/discharging module to execute operation for a combinational optimal problem searching rule, so as to establish a programmable charging curve based on the operated result, and perform multi-stage charging to the battery based on the charging curve, thereby searching the maximum charging capacity based on the searching rule to find the most suitable charging curve of the battery. Accordingly, the battery manufacturer can quickly determine a most suitable charging curve when the battery is produced, and the charging curve can be loaded into a charger to effectively increase the charging efficiency and lifetime of the battery.

Description

200411201 V. Description of the invention (1) [Technical field to which the invention belongs] The present invention relates to a battery optimum charging curve searching device and method 'especially a battery optimum charging curve searching device and method capable of effectively improving battery charging efficiency and service life. [Previous technology] With the design and manufacture of electronic products shrinking in size, various portable electronic products are greatly improving consumers' convenience in daily life-the trend of jujube use also makes Rechargeable batteries for power supplies are becoming more and more important. For example, conventional mobile phone batteries such as lithium-based batteries, nickel-cadmium batteries, and hydrogen batteries. How to complete high-capacity charging in a short period of time without reducing this type after multiple charging? The battery life is the bottleneck of whether the relevant portable electronic products can be more popular in the market or meet the needs of consumers. In addition, for electric vehicles that have been developed but are rarely heard, the aforementioned charging efficiency and Battery life issues are also the main reason for its difficulty to be extended. The battery charging method used in the conventional charger is a certain current / constant voltage method, which first charges the battery with a certain current and waits for it to reach the upper limit of the battery pack ^ (for example, the lithium battery is about 4 · 1 to 4.2 volts) at the same time, then the voltage is charged again. However, this conventional method cannot meet the requirements of fast charging and extended service life, because the constant-voltage charging stage will greatly extend the charging time, and will also Lead to reduced battery life; as for other conventional multi-stage charging methods, such as multi-step constant current charging (Multi-Step) or earth-phase pulse current (Multi-step Pul sing) charging, High voltage charging efficiency, but optional when charging

200411201 V. Description of the invention (2) The number of combinations of the selected charging curves is extremely amazing, and this value will gradually increase with the increase of the number of battery charging stages and the complexity of its electrochemical (E 1 ectr chemica 1) characteristics. With a spectrum of hundreds of thousands of groups, it is impossible for battery manufacturers to test the optimal charging combination and mark it on the battery before the battery leaves the factory. This test process not only consumes a lot of manpower and resources, but also delays the time to market of the battery. In addition, for the charger manufacturer, because it is also unable to perform this test to find the optimal charging curve for various types of batteries, it is difficult to optimize even if the charger designed by it has programmable functions Control efficiency; therefore, although the charging efficiency and battery life of the multi-stage charging method are better than the constant current / constant voltage charging method, the optimal charging curve that can exert its maximum efficiency is difficult to obtain by quick search, and With the increase of battery types and battery manufacturing formulas, the search for this optimal charging curve is bound to be more complicated. , Resulting in the multi-stage charging efficiency greatly reduced, difficult to apply needed. Therefore, how to establish a battery optimal charging curve search device and method, and apply it to the known multi-stage charging to improve the search speed, so that battery manufacturers and charger manufacturers can design for the same type of battery, and then effectively improve The charging efficiency and service life of batteries are indeed the issues that need to be solved in this related research and development field. [Summary of the Invention] —— Therefore, an object of the present invention is to provide a device and method that can quickly search for the optimal battery charging curve. Another object of the present invention is to provide a battery optimum charging curve searching device and method to effectively improve the charging efficiency of the battery.

17030. ptd Page 7 200411201 V. Description of the invention (3) Another object of the present invention is to provide a device and method for searching for the optimal charging curve of a battery, so as to effectively improve the battery life. In order to achieve the foregoing and other objectives, the battery optimum charging curve search device provided by the present invention is a programmable battery charge-discharge module, including: a charging unit, a discharging unit, and a charging unit and a charging unit respectively. The central processing unit connected to the discharge unit. The central processing unit can execute the "Combinatorial Optimization -Problem" search rule, and the programmable battery charge and discharge module can be used for external charging. The battery pack uses the central processing unit to calculate the combination of the most __ dagger problem search rules, enters a stylized charging curve into the charging unit, and performs more than one charging battery pack according to the charging curve. Phase charging. The method for searching the battery optimal charging curve provided by the present invention includes: inputting a stylized combined optimization problem search rule into a charging unit, and setting the search target of the search rule to the capacity of the battery after charging Maximize; charge a battery pack in a multi-stage charging mode, and measure the charging capacity of the battery pack after charging is completed; repeat the battery. Repeat multiple multi-stage charging of the battery pack, and measure the amount after each charge The measured battery pack capacity is combined with the combined optimization search algorithm to determine the charging curve of the jfc battery pack in the next multi-stage charge; when the multi-stage charge is performed, the charging result has reached a preset When the optimal charging target is reached, you can stop searching and use the charging curve of the most batteries in the battery pack at that time as the optimal charging curve for the multi-stage charging. By using the foregoing battery optimal charging curve searching device and method,

17030.ptd Page 8 200411201 V. Description of the invention (4) The manufacturer can pre-enter the device with a stylized combinatorial optimization problem search rule before the battery leaves the factory. Search the charging curve, and quickly search for the most suitable charging curve by using the mathematical principle of the search algorithm, to solve the problem that it requires a lot of manpower and material resources to test in practice; it can be used in the invention for rapid Combined optimization problem search rules for optimized search include coded ant search algorithm (Ant Colony System Algorithm), evolutionary strategy rules (E v ο 1 u ΐ i ο n S trategies), and genetic algorithm (Genet ic Algorithms), neural-like network (Neutral

Networks), and Simulated Annealing, etc. In addition, any conventional multi-stage combined charging method can be searched and tested using the present invention, including multi-stage constant-current charging, multi-stage pulsed current charging, and multiple Stage constant current / constant voltage charging, etc., can be tested in conjunction with the present invention to search for individual optimal charging curves. In summary, the present invention can mark the optimal charging curve obtained on the battery before the battery leaves the factory, or provide the search result to the charger manufacturer 'to load the charger and set the Optimal charging cable 'so' can quickly find the optimal charging curve without delaying production. The time to market can also effectively improve the battery charging efficiency and service life, and it can meet the needs of users. [Embodiment] The battery optimal charging curve searching device 1 proposed by the present invention is as shown in the figure, and its basic configuration diagram is a programmable battery charge-discharge module 10 'and is connected with a plurality of battery cells. The composition of the test battery pack 2 is connected,

17030.ptd Page 9 200411201 V. Description of the invention (5) A stylized multi-stage charging and discharging process, wherein the programmable charging and discharging module 10 includes a central processing unit 11 and a charging unit 1 2 And a discharging unit 13, the central processing unit 11 includes at least a computing unit and a measuring unit (not shown), wherein the computing unit can input a stylized charging curve (that is, the charging parameters of multi-stage charging) into the The charging unit 12 enables the charging unit 12 to perform multi-stage charging of the plurality of batteries in the test battery pack 20 according to the charging curve, and the measurement unit of the central processing unit will The discharge results of the discharge unit 13 measure the charged capacity of the individual batteries in the test battery pack 20 and input them to the calculation unit of the central unit to calculate the new charge for the next multi-stage charge. Curve, and then input the stylized new charging curve into the charging unit 12 to repeat the foregoing charging test until an acceptable optimal charging curve is obtained. FIG. 2 is a schematic diagram of a conventional five-phase constant current charging method. A preferred embodiment of the present invention adopts this charging method as an example of obtaining an optimal battery charging curve (optimal charging current combination). Of course, the present invention The proposed method can also be applied to the conventional multi-phase pulse current charging or multi-phase constant current / constant voltage charging methods, and the optimal charging curve can also be obtained; as shown in Figure 2, we can first target the five phases. A charge # current is set for charging, and the set current value will decrease as the number of stages increases. This design is based on the principle of conventional multi-stage constant current charging: when the voltage of each stage reaches the rechargeable battery At the upper limit of the voltage, the charging current value is reduced to a preset value and enters the next stage of charging to avoid the battery terminal voltage being too high and causing damage to the battery; as can be seen from the figure, each stage

17030.ptd Page 10 200411201 V. The charging time end of the description of the invention (6) depends on the current value set at this stage. Therefore, the optimal battery charging curve sought by the present invention is one that enables the final charging efficiency. The best five-phase current combination. In addition, in order to meet the fast charging requirements on actual charging, taking this embodiment as an example, the total time of the five-phase charging needs to be limited before testing to obtain the set time. The five-stage constant-current charging curve that achieves the highest charging efficiency; Table 1 shows the charging current values that can be selected in the five-stage charging used in this embodiment. The unit of each current value is the standard of the rechargeable battery. The corresponding charging current value of the Rated Apacity. This current selection combination takes into account the upper voltage limit (about 4.1 · 4.2 V) of the lithium battery for the mobile phone used in this embodiment and the set value of 30. Depending on the total charging time in minutes, the five-stage combination has a total of 1 71, 0 7 2 (1 2x 1 2x 1 2x 1 lx 9) charging curves, and the optimal charging curve required by the present invention is Under the limitation of the upper limit of the voltage of the rechargeable battery, the charging curve that can reach the maximum charging capacity in 30 minutes of charging, that is, the optimal current combination in this multi-stage constant current charging. The aforementioned 171,07 two kinds of charging curve combination, if it is based on the characteristics of multi-stage. Constant current charging, the current value of each stage of charging must be less than the current value of the previous stage to avoid the battery voltage exceeding its upper voltage limit, so Actually, the combination that can be applied to the five-phase constant current charging should be slightly less than the 17 1, 0 7 2 groups, but this is still a very large number. It is not possible for battery manufacturers to address this large number of current combinations one by one before leaving the factory. Tests, the search method of the battery optimal charging curve proposed by the present invention can be applied to this search process, so as to quickly find the optimal charging curve in the large number of combinations; the method proposed by the present invention uses a combination of Optimization problem (C 〇mbinat 〇ria 1

17030.ptd Page 11 200411201 V. Explanation of the invention (7) 0 ptimizati ο η P r 〇b 1 em), in order to quickly find the optimal solution of the charging curve, such search rules are led by artificial intelligence 〇 ^ 8A has been developed for a long time, this preferred embodiment will use the sincere warfare proposed in the past ten years ^ ^ w code Ant search algorithm (Ant Colony System A Igor it hm) to solve this _ day day Figures 3A to 3D are the development principles of this search rule, which is designed to imitate the feeding principles of ants in the exploded self-deficient boundary. Since the ants are foraging, the second mine is on the path that they pass. Leaving a chemical called feromon (Pheromone), f, and the nature of the ant will make each sacrifice follow the previous ant $ rate is directly proportional to the amount of pheromone left by the previous ant, that is, The more the number of pheromone on the path chosen by Xi ', the greater the probability that the next two yards will choose this path. As shown in the figure, when an obstacle is placed between ants, initially Ants choose two ends of the obstacle: the probability of walking will be the same, but by Guijiang, who chooses to walk on the short end (point M), will reach the food faster, so in the unit time, the number of ants passing and the amount of pheromone remaining will be longer. ^^ 1 And the subsequent multi-digits will follow this: 2: Second ,: If the time is long enough, the most bumpy & 1 will follow, and all the ants will follow the path, this is the ant f > η Wang 4 chooses the pheromone on the path of the button to meet y λ…, the ants passing through will accelerate the ant colony 5 Vap〇rat 1〇η), this natural phenomenon is based on this self == -The shortest path to feeding. The ant search method loop enables the virtual sacrifice ant characteristics to establish a virtual ant colony operation (FUness starts searching for a preset optimal value for the shortest path. The optimal value is used to simulate the real ant colony foraging. The object to be searched according to the search rule

17030.ptd Page 12 200411201 V. Description of the invention (8) Depending on the simulation method of this search rule, the probability model of a virtual pheromone is established. The probability value will be updated along with the update of a virtual pheromone function. Taking the probability that the virtual ant colony chooses the path, the path selection probability function assumes that the number of virtual pheromone on each path is directly proportional to the number of virtual ants that select the path. Therefore, at each specific time point, each The probability that the path is selected will be proportional to the total number of virtual ants that passed the path from the beginning to this point in time. This is the design principle of the search law. Among them, the path selection probability function that determines the selection of the virtual ant colony is shown below. :

: [rfma Σ i ^ ma household phase y where τ (N) is the amount of pheromone remaining in path (i, j) in the Nth search, which is updated after each search Virtual pheromone function, and α indicates the degree of non-linearity of this probability function. If α is larger, if there is more pheromone on one path than on the other, the probability that it will be selected by the ant will be The more it is zoomed in; the path selection probability function p (Ν) will be updated after each search is completed. This is because the pheromone function 7: (Ν) will be set after each search in this simulated search. Update once, and its pheromone update law is as follows:

17030. ptd page 13 200411201 V. Description of the invention (9) where p is a parameter, so that (1-P) represents the evaporation ratio of pheromones, and △ 剣 is the kth ant in the Nth time The amount of pheromone remaining in the path (i, j) in the search, its value can be expressed as: when the Kth ant passes the path Lk (when the force is-4 = 0 other sounds, where L is the kth ant The optimal value of is used to simulate the length of the path searched by the kth real ant, and Q is a preset parameter. The aforementioned a, p, and Q parameter values can be simulated by the tester according to the reality they want to simulate. In addition, in order to simulate the situation where the real ant colony has the largest amount of pheromone in the shortest path per unit time, we can give a shortest path among the paths found by all virtual ants in the Nth search. An additional amount of virtual pheromone is added to make the path easier to be selected by the ant colony in subsequent searches to simulate the search characteristics of the real ant colony. The virtual pheromone increase rule of the optimal path is shown below. Among them, suppose that in the Nth search The path taken by an ant is the shortest one in the search, the amount of pheromones remaining in the Λ shortest path will be added by an additional amount: τ ^ (Ν + ΐ) = τ ^ (Ν + 1) + —S-

17030.ptd Page 14 200411201 V. Description of the Invention (ίο) F (I 〇 is the path length searched by the first ant, that is, its optimal function value, Q is the same parameter as above; therefore, as The foraging characteristics of real ant colonies. When the number of searches performed by this search rule is greater, the probability value P (N) of the virtual ant colony choosing a better path will be greater, and the virtual ants making the same better search results. The ratio of the colony to the total ant colony will also be higher. The search path selected by this most ant colony is an approximate optimal path. From a theoretical point of view, when the number of search tests is extremely large, all virtual ant colonies are The same best path will be selected, but in practice when using this search rule to solve the problem, if the time cost of the test is considered, the test can be ended only when a certain number of virtual ant colonies have selected the same path. That is, the Near-Optimum Solution of this proposition, and the judgment condition for determining whether the test can be ended can be set by the tester in advance. Therefore, we can use the battery of this embodiment. The optimum charging curve search of the five-stage charging is simulated as an optimization problem according to the ant-type search rule:. Maximizing the optimal function F (I), where F is the battery charging capacity I-[Ιι 12 13 14 15] is limited to I, ^ I] If 1 < j, i, j two l · ·· 5 can set other charging parameters before performing this optimization search, for example

17030.ptd Page 15 200411201 V. Description of the invention (11) For example, the number of batteries tested, the total time for five stages of charging, the test conditions to be suspended, etc., the embodiment of the present invention tests 15 lithium-based batteries for mobile phones. The total time of five-stage charging is set to 30 minutes. The test equipment shown in Figure 1 and the five-phase constant-current charging method shown in Figure 2 are used to test. The design principle of this test can be shown in Figure 4, 1 5 Only a virtual ant (meaning 15 lithium batteries to be charged) will start from the starting nest, randomly select the optional charging current value shown in Table 1, and then perform five-phase current selection and-complete the charging to the end point, that is, see In order to complete a five-stage charging test, at this time, the test system will calculate its Fermon function according to the charging path selected by each virtual ant and establish its path selection probability function as a reference for the virtual ant to select the charging path during the second test. , Repeat the pheromone update, the probability function, and the test repeatedly until a certain proportion of the 15 virtual ants (that is, the test is stopped before being set) Conditions) all select the same charging path. At this time, the charging path is an approximately optimal charging current combination, that is, the optimal charging curve of the type of test battery under the charging condition. This embodiment is set to have 6 0% of the virtual ants found the same path. The test was ended when 9 of the 15 test batteries were charged the same, and the test was suspended when the curve was charged; the following will further explain the invention Implementation method of the embodiment. 'Refer to Fig. 5 shows the actual test equipment used in this embodiment, which is extended from the configuration diagram shown in Fig. 1. Among them, an MCN-type battery life tester developed by Bitrod e company in the United States is used to test The battery is used for charging and discharging. In this embodiment, 15 lithium-based batteries are tested at the same time, and the charging results of the test are recorded after each five-stage charging test.

17030.ptd Page 16 200411201 V. Description of the invention (12) The calculation of the ant search rule by an external computer is used to update the pheromone function r (N) and path selection probability of the next five-stage charging according to the charging result. Function p (N), input this stylized path selection result (charging curve) into the Bi trode MCN tester, and repeat the five-stage charging test until an acceptable charging curve is obtained; the sixth The flowchart shown in the figure is the test flow of ant search in this embodiment. In step S 10, 15 lithium-based batteries to be charged are tested first. The test items include acceptable open circuit voltage (0CV). , Impedance value, nominal capacity value, weight, etc. The test results are listed in Table 2. Based on this result, you can set the charging parameters such as the battery's upper voltage limit and the unit charge current C value. Step S 1 5 will perform an ant search The setting of the law, the parameters set in this embodiment are the number of charging stages = 5, and the current value that can be selected is shown in the above table 1. As for the parameters of the search law, they are set separately. a = 1. 0, p = 0.7, Q = 8 0 0; after the setting is completed, the process proceeds to step S 2 0, and the 15 lithium-based batteries are subjected to multi-phase constant current charging, which is performed by Bitr 〇de The MC Ν tester is charged according to the programmed charging curve and discharged with a small current after charging is completed. According to the discharge result, the charging capacity of each lithium battery is recorded. Among them, there is no virtual fee for the first test. The Lomon function is generated, so its charging curve is determined by a stylized random selection. The selected value is based on the selectable current values listed in Table 1. However, the random selection should still be charged at each stage. The current value is less than the limit of the current value of the previous stage. This limit function can be implemented by stylized design. Step S 2 5 is to evaluate the optimal value. The optimal value is set to the capacity of the lithium battery after charging. This search The search setting of the rule is to find a charge that maximizes the optimal value

17030. ptd Page 17 200411201 V. Description of the invention (13) Curve, step S 2 5 will evaluate the optimal value according to the battery charging capacity recorded in step S 2 0 and calculate it in step 30; step 30 will The virtual pheromone function is updated according to the evaluation of the optimal value of step 25. The calculation method is the same as the previous mathematical model. The result of this update will be determined in step S35. The 15 virtual ants will be selected in the next test. The probability function of the path of the charging path is selected, and new charging curves will be established for the 15 lithium-based batteries according to the new probability function. Step S 40 adds 1 to the number of tests recorded in the program, and then The discriminating block in step S45 determines whether to suspend the test. This discriminating block is determined by the abort test condition set in advance. In this example, the abort test condition is set when 60% of the virtual ants find the same path. That is, when 9 lithium-based batteries are charged with the same charging curve, the test is suspended, and if the result of step S 35 does not meet the conditions of the suspension test, it returns to step S 2 0, The new re-charging curve 15 for charging the test battery, until you find a near-optimal charge profile is reached. The results listed in Table 3 are the charging test results of this example. From the data listed in the table, it can be found that when the number of searches of the ant-type search rule increases, the charging efficiency of each battery does indeed increase, reaching within 30 minutes. The number of batteries with a charging capacity of more than 70% obviously increases with the increase in the number of tests. This junction also shows that this embodiment will reach the preset test suspension conditions at the 20th test: at least 9 After testing the battery with the same charging curve, the charging curve used in the fifth stage of charging during the 20th test is listed in Table 4, among which the numbers are 1, 2, 4, 5, 6, 7, 8, 10, 13, 14, 4, and 15 lithium-ion batteries all have the same charging curve, that is, there are

17030.ptd Page 18 200411201 V. Description of the invention (14) 1 All 1 virtual ants have searched for the same charging path. This result obviously has reached the pre-set termination test conditions. At the same time, the charging curve selected by the majority is It can be regarded as the approximate optimal charging curve under the charging conditions. From the data in Table 4, it can be seen that the charging curve is a five-step process in the order of 2.1C, 1.7C, 1.5C, 1.3 C, and 1.0 C. Stage 3 constant current charging, as shown in Table 3, it can reach a charging capacity of more than 70% in 30 minutes, which is consistent with the charging efficiency of lithium batteries used in mobile phones. In other words, the upper limit of the capacity that can be charged in 30 minutes is about 70% of the original nominal capacity. Figures 7 and 8 are the data collation of the test results of this example. Figure 7 shows that the charging capacity of the 15 test batteries will gradually increase with the increase of the number of tests. The figure shows that the standard deviation of the charging capacity between the 15 batteries will gradually decrease as the number of tests increases. The trend of the two figures is in line with the search principle of the ant search rule, that is, by the update of the pheromone function. , The probability of each virtual ant searching for a better charging path will gradually increase, and the probability of finding the same better charging path between different virtual ants will also increase. This trend is reflected in the 15 as shown in the figure. Changes in the capacity of a lithium battery after charging and its standard deviation. In addition, since the battery's chargeable capacity value will gradually decrease and its service life gradually decrease after long-term use and repeated charging, in order to verify the efficacy of this embodiment, it is necessary to obtain a high charge. After the efficient charging curve, the multi-stage charging of the charging curve is compared with the conventional constant current / constant voltage charging method to verify the impact of the two charging methods on the battery life. Figure 9 shows the results of the comparison between the two Which is used to

17030. ptd Page 19 200411201 V. Description of the invention (15) The comparison of the charging curve obtained in this example is a constant current / constant voltage charging method with a constant current of 2 · 1 C. The test method used is After 20 charging cycles, change to a constant current of C / 3 for 30 minutes and measure its capacity value, and use the measured capacity as a percentage of the capacity after the first charge as the capacity indicator. As shown in the figure, it can be found that the battery charged by the charging curve obtained in this embodiment has a longer service life than the conventional battery with constant current / constant voltage. If the capacity is reduced to 8 times the original charge after charging 0% is a comparison benchmark. It can be found that the charging method of this embodiment still has a capacity of 80% when the charging cycle reaches 9 27. On the contrary, the capacity of the conventional method has dropped to 80% after only charging once. The charging curve obtained in this embodiment can indeed slow down the battery life. For some battery use fields that require full charging (Fu 1 1 y C harged), a simulation test can also be used to verify the effectiveness of this embodiment. , Figure 10 shows the results of this test. It is the same as the test conditions in Figure 9, but it is no longer changed to charge with a small current every 20 cycles. Instead, it is charged and measured according to the setting method in each cycle. By comparison, it can be seen from the figure that the battery life of the charging method proposed in this embodiment is still longer than that of the conventional charging method, and it can also be found that the charging method of this embodiment has a charging cycle of 3 7 7 It still has a capacity of 80%, which is much higher than the conventional method of reducing the capacity to 80% after only 300 times of charging. Therefore, compared with the conventional technology, the device and method for searching for the optimal charging curve of the battery of the present invention have the effect of effectively improving the charging efficiency and service life of the battery. The battery manufacturer can optimize the problem with this device and related combinations. The search rule (such as the ant search rule of this embodiment) searches for possible charging curves, and charges the optimal charge obtained before leaving the factory.

17030. ptd Page 20 200411201 V. Description of the invention (16) The curve is marked on the battery, and the search results can also be provided to the charger manufacturer to load the charger and set Yuan Shi charging curves for different types of batteries. In this way, it can quickly find the optimal charging curve without delaying the time to market, and can effectively improve the battery charging efficiency and service life, which fully meets the needs of users. However, the above are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention. For example, all equivalent changes or modifications made by those skilled in the art under the spirit and principles disclosed by the present invention For example, the use of other relevant programmable charge and discharge equipment, other combined optimization search problems for the search of the optimal charging curve, or the application of the method proposed by the present invention to other various multi-stage charging methods, etc., should still be used. All are covered by the patent scope described below.

17030. ptd Page 21 200411201 V. Description of the invention (17) Table 1: Selectable charging current stage (C) 1 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1-5 1.4 2 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 3 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 4 1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6-5 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5---Table 2 : Test result of charging battery: maximum, minimum, average, weight (g) 29.95 29.47 29.80 Open circuit voltage (V) 4.17 4.13 4.15 Impedance 値 (ιηΩ) 113.6 89.4 94 · 21. Nominal capacity (mAh) 938 918 930

17030. ptd Page 22 200411201 V. Description of the invention (18) _ J to 寮 s —α 1 1 H- ^ i% c 1 8 1 GI 1—- · Η " · s | FREE Bromine Ό CO ii CK « «-Λ 1 1 w ϊ to 3 I ~ · 'Mr 嬅 to-^ J ►— * FREE TO $ --J GN CO $ c ^ o, Os Ό 3 3 · 〇o, ^ Ί > t § G > > * ·· 〇, sf—Λ- --〇 $ a " —J 1 ~ · $ O " · --J § § --0 1— * 〇-_ VO $ 〇 o, o, Os s 2 o, > "-CN. > ^ l CN CN 3 s to © O uo ^ J to -—JS $ LO 5- LO FREE 5- ΐά --〇1 --- $ 3 $ CN Cl ·, Oo 〇. O. ≫ " g ls LO P 〇 ^ -J σ, o $ 茭 © -0 § r, 〇 · *. Α ·, «νΛ σ, 彡 $ σ ·. 〇 · ,, 00 σ, oo § ¥-Office $ ￥ $ ijj $ ilo-«〇LO 5 " rf free $ -J-^ 0 p s-% 〇 · Ό Try § S3 $ c ^ LO • 3s: s σ ·, σ \ to .name -o to $ -0 to 〇1 ~ >;--j iI〇r,, name—J 1 »· $ s s- 〇. ΟΟ ο S On σ ·· S 〇 ·, $ On C? S FREE PLUS $ w • S --0 PLUS —j LO § PLUS --J a% σ, Ό $% 'S Os Ό σ, LO σ'. 〇. --J [〇--J to J §5 FREE -ο -0 § 〇 · σ · S. This 0% § C ^ CO > " σ, --〇OO -ο • s ίο, s to S- -O --〇-ο r, --ja -ο § § s Os窆 C? S C5 \ -O 〇oo Ό --j § o. '-Λ ΪΟ --J »~ >» --J to © © $-~%! I Ον • Ο Os 、, s σ » Si, s C ^ try 2 C5 \ s- Ch oo 1 " " > · o $ 5 " 'Name% sg --0 I— * -J a _ gi ίο $ σ, 〇 · 〇 «οο § a $ σ, Os 'Name On G ^. 5-Si 1 ~ > · 1 · " · CO Os Si s $% 突 οο gpa' Name 8¾, S C5S try ss; N 〇LO Os CD σ, LO $ 1 ~ to S " σ,% o 5- O, Ό ©%;-© 3 © Os σσ, οο oo Si a σ; 00 5- ON LO 1 ·· LO —j --J -〇 III§ ο. 'ν ^ ^ -J ρ-ύ § Os Ό $ c ^ § S; s $ o, 〇, Os 窆 E Ψ- -0 § §-" 〇% id ίο a $ Cl ·, Ό 5 ^ OS ν〇, s ° 5 Si $ s CN 〇. 〇 · I—4 · iiS fltD

HI η > ϊτ η m Τώί < =. _ Xue _

17030.ptd Page 23 200411201 V. Description of Invention (19)

Table 4: Charging curves used by lithium batteries in the 20th test

Lithium-based battery numbering Phase 1 Phase 2 Phase 3 Phase 4 Phase 5 1 2.1C 1.7C 1.5C 1.3C 1C 2 2.1C 1.7C 1.5C 1.3C 1C 3 2C 1.4C 1.2C 1C 0.8C 4 2.1C 1.7C 1.5C 1.3C 1C 5 2.1C 1.7C 1.5C 1.3C 1C 6 2.1C 1.7C 1.5C 1.3C 1C 7 2.1C 1.7C 1.5C 1.3C 1C 8 2.1C 1.7C 1.5C 1.3C 1C 9 2.1C 1.6C 1.5C 1.3C 1C 10 2.1C 1.7C 1.5C 1.3C 1C 11 2.5C 1.8C 1.4C 1.3C 1C 12 1.7C 1.6C 1.5C 1.3C 1C 13 2.1C 1.7C 1.5C 1.3C 1C · 14 2.1C 1.7C 1.5 C 1.3C 1C 15 2.1C 1.7C 1.5C 1.3C 1C Page 24 17030.ptd 200411201 Simple illustration of the drawing [Simplified illustration of the drawing] Fig. 1 is the configuration diagram of the device for searching the optimum charging curve of the battery of the present invention; Fig. 2 It is a schematic diagram of a conventional five-phase constant-current charging method; Figures 3A to 3D are not intended to be the development principle of the ant search rule used in the embodiment of the present invention, and Figure 4 is a schematic diagram of the design principle of the embodiment of the present invention; FIG. 5 is a schematic diagram of a search device used in an embodiment of the present invention; FIG. 6 is a flowchart of a search method used in an embodiment of the present invention; and FIG. 7 is a flowchart of a test process of an embodiment of the present invention Trend chart of battery charging capacity; Figure 8 is a trend chart of the standard deviation of battery charging capacity during the test process of the embodiment of the present invention; Figure 9 is a comparison chart of the service life of batteries charged with the results of the embodiment of the present invention and conventional technology; And FIG. 10 is another comparison diagram of the service life of the battery charged with the result of the embodiment of the present invention and the conventional technology. I Battery Optimal Charging Curve Search Device 10 Programmable Charge and Discharge Module II Central Processing Unit 12 Charging Unit 1 3 Discharging Unit 2 0 Test Battery Pack

17030.ptd Page 25

Claims (1)

200411201 6. Scope of patent application 1. A battery optimal charging curve searching device is used to search the optimal charging curve of a set of multi-stage charging parameters at each stage, so that the battery after multi-stage charging according to the optimal charging curve has a close The largest charging capacity, the search device includes: a battery test section, which can charge multiple batteries in multiple stages and measure the charging capacity of each battery after charging; and-an arithmetic processing section, which can be Search for the Combinatorial Optimization Problem: Bay 1J, before obtaining the optimal charging curve, continuously receive 0 from the battery test section to measure the charging capacity of each battery after the previous charge, in order to The measured charging capacity after charging determines the charging curve of each battery at the next charge, and provides the next charging curve of each battery to the battery test section. 2. For example, the battery optimal charging curve searching device of the first patent application scope, wherein the search rule applicable to the combination optimization problem is an ant search rule (Ant C ο 1 ο ny S ystem A 1 g 〇rithm). 3. —A kind of battery optimal charging curve searching device, which is used to search the optimal charging curve of a set of charging parameters for each stage of multi-stage charging, so that the battery after multi-stage charging according to the optimal charging curve has a near maximum charging capacity. The search device includes: a battery test section, which can charge multiple batteries in multiple stages and measure the charging capacity of each battery after charging; and a calculation processing section, which is based on the Ant Colony System Algorithm. Before finding the optimal charging curve, 17030. ptd page 26 200411201 VI. Patent application scope Receive the measured charging capacity of each battery after the previous charge from the battery test department to determine the next charge of each battery based on the measured charge capacity of each battery after the previous charge The charging curve of each battery is provided to the battery test section. 4. For the battery optimal charging curve searching device of the first or third item of the patent application scope, wherein the battery test section is a programmable battery test device, and the arithmetic processing section is a computer. 5. For the battery optimum charging curve searching device of the scope of patent application No. 4, wherein the battery testing device is a person who can charge the battery and discharge the battery. 6. For the battery optimal charging curve searching device of the scope of patent application item 4, wherein the battery test device and the computer are independent of each other and connected by a signal cable. 7. For the battery optimal charging curve search device of the first or third item of the scope of patent application, the charging parameters of each stage of the multi-stage charging are the charging currents of the stages of the multi-stage charging. 8. If the battery optimal charging curve searching device of the first or third item of the scope of patent application is applied, the calculation processing section ends when the determined charging curve of each battery in the next charging is the same, then the process ends. Search and use the same charging curve as the optimal charging curve. 9. For the battery optimal charging curve search device of the first or third item of the patent application scope, the battery is selected from one of the group consisting of a lithium-based battery, a nickel-metal hydride battery, and a nickel-cadmium battery. 1 0. —A method for searching the optimal charging curve of a battery, which is used to search a group of multiple 17030.ptd Page 27 200411201 Six. The optimal charging curve of the charging parameters in each stage of the patent application phase, so that the battery after multi-stage charging according to the optimal charging curve has a maximum charging capacity, including the following steps: ( 1) Perform multiple multi-stage charging of multiple batteries and measure the charging capacity of each battery after charging. Among them, the first charging is based on the initial charging curve of each battery, and after the second charging, the following steps are followed (2 ) • The charging curve of each battery determined on the next charge; _ (2) According to the search applicable to the Combined Optimization Problem (Combinatorial Optimization Problem), go to φ, from the previous charge of each battery Measure the charging capacity to determine the charging curve of each battery for the next charge; and (3) Repeat the above steps (1) and (2) after discharging each battery until the batteries determined in step (2) are charged next time. When there is a certain proportion in the charging curves of, the search ends, and the same charging curve is used as the optimal charging curve. 11. The search method for the optimal charging curve of the battery, such as the item 10 in the scope of patent application, wherein the search rule applicable to the combined optimization problem is the Ant Colony System Algorithm. 1 2. A method for searching the optimal charging curve of a battery, which is used to search the optimal charging curve for a set of charging parameters of each stage of multi-stage charging, so that the battery after multi-stage charging according to the optimal charging curve has a near-maximum charge The capacity includes the following steps: (1) Perform multiple multi-stage charging on multiple batteries and measure the charging capacity of each battery after charging. The first charging is based on each battery. 17030.ptd Page 28 200411201 VI. Initial charge curve of the patent application pool. After the second charge, follow the charge curve of each battery determined in the following step (2); (2) According to the ant formula The search rule (Ant Colony System A 1 gorithm) determines the charging curve of each battery for the next charge from the measured charging capacity of each battery after the previous charge; and (3) repeats the above steps (1) after discharging each battery, (2), until a certain proportion of the charging curves for the next charging of the batteries determined in step (2) is the same, the search is ended, and the same charging curve is used as the optimal charging curve. 13. If the method of searching for the optimum charging curve of a battery in item 10 or item 12 of the scope of patent application, wherein the charging parameters of each stage of the multi-stage charging are the charging currents of the stages of the multi-stage charging. 14. If the method of searching for the optimal charging curve of a battery is in item 10 or item 12 of the scope of patent application, wherein the initial charging curve of each battery is randomly generated. 15. If the battery of item 10 or item 12 in the scope of the patent application is the most suitable for charging. The method of searching the electric curve, wherein the battery is selected from the group consisting of a lithium-based battery, a nickel-metal hydride battery, and a recording battery. One of them. 17030.ptd p. 29