TWI572879B - Electronic apparatus and method to estimate state of charge for battery - Google Patents

Description

Electronic device and method for evaluating battery residual state

The present invention relates to an electronic device and method for evaluating the state of a residual charge of a battery. More specifically, the present invention relates to an electronic device and method for evaluating the state of residual state of a battery by a neural network.

With the continuous advancement of electronic technology, modern people often use devices or products with various forms of batteries in their daily lives. In general, when the battery is discharged, its residual power exhibits a state of non-linear characteristics, so it is difficult to accurately detect it. Therefore, the residual state of the battery is one of the estimated values obtained by evaluation, rather than one of the values obtained by accurate detection. For example, batteries used in hybrid electric vehicles (HEVs) or electric vehicles (EVs) have a high discharge rate (C-rate), which exhibits a relatively strong nonlinear characteristic. The state of charge, so it is almost impossible to accurately assess the state of residual power of such batteries.

There are many methods for evaluating the state of residual state of a battery, such as an ampere-hour counting method, an open circuit voltage (OCV) measuring method, or a battery impedance measuring method.

The amp-hour counting method evaluates the state of the residual charge by detecting the actual amount of power in the battery. In the case of using this method, the state of the residual power of the obtained battery is evaluated in relation to the quality of the sensor used to detect the actual amount of power of the battery. Therefore, the state of the residual power of the battery evaluated by the ampere-hour counting method varies depending on the accuracy and error of the sensor for detecting the actual amount of power of the battery.

The open circuit voltage measurement method is based on the open circuit voltage of the battery as a reference to evaluate the state of the residual power. In the case of using this method, it can only be used in electricity. The battery's residual state is evaluated when the pool is not in use. In addition, the open circuit voltage measurement method is quite susceptible to the external environment (such as temperature) and reduces the accuracy of evaluating the residual state of the battery.

The battery impedance measurement method is based on the impedance value of the battery. An assessment of the state of its residual capacity. However, the battery impedance measurement method is quite susceptible to the temperature and the degree of battery aging, which reduces the accuracy of evaluating the residual state of the battery.

In view of this, how to provide a review that is not affected by the external environment Estimating the state of the battery's residual state allows the user to quickly and accurately obtain the state of the battery's residual charge, which is an urgent problem to be solved in the industry.

It is an object of the present invention to provide an electronic device for evaluating the state of residual charge of a battery.

To achieve the above objective, the electronic device of the present invention comprises a sensing module, a storage unit and a processor. The sensing module is electrically connected to the battery. The storage unit is configured to store a training database. The processor is electrically connected to the battery, the sensing module and the storage unit and has a neural network. The sensing module is configured to capture a temperature, a voltage of one end, and a discharge current of the battery; generate a temperature signal, a voltage signal, and a current signal corresponding to the temperature, the terminal voltage, and the discharge current of the battery; and output The temperature signal, the voltage signal, and the current signal. After receiving the temperature signal, the voltage signal and the current signal, the processor calculates the residual state of the battery according to the training database, the temperature signal, the voltage signal and the current signal by using the neural network. Finally, the processor outputs a display signal with one of the remaining states of the battery.

Another object of the present invention is to provide a method for evaluating the state of residual power of a battery for the electronic device described in the preceding paragraph. The method includes the following steps: causing the sensing module to capture a temperature, a voltage of one end, and a discharge current of the battery; and causing the sensing module to generate a temperature signal corresponding to a temperature, a terminal voltage, and a discharge current of the battery a voltage signal and a current signal; causing the sensing module to output the temperature signal, the voltage signal, and the current signal; causing the processor to receive the signal a temperature signal, the voltage signal, and the current signal; causing the processor to calculate a state of residual power of the battery according to the training database, the temperature signal, the voltage signal, and the current signal by using the neural network; and The processor outputs a display signal having a state of residual state of the battery.

In summary, the present invention is used to evaluate the state of residual power of a battery. The electronic device and method can calculate the residual state of the battery by using the neural network of the processor and the pre-stored training database after the sensing module draws the temperature, the terminal voltage and the discharge current of the battery. Finally, a display signal having the state of the residual state of the battery calculated above is output. Accordingly, the electronic device and method for evaluating the state of the residual state of the battery of the present invention can effectively overcome the prior art, and the accuracy of the residual state of the battery of the battery is lowered due to the influence of the external environment. problem. At the same time, through the pre-stored training database, the speed of calculating the residual state of the battery by the neural network can also be improved.

Other objects, advantages, and technical means and embodiments of the present invention will become apparent to those skilled in the <RTIgt;

1‧‧‧Electronic device

11‧‧‧Sensor module

13‧‧‧ Processor

15‧‧‧ storage unit

17‧‧‧Battery

19‧‧‧Display unit

110‧‧‧temperature signal

111‧‧‧Temperature sensing circuit

112‧‧‧Voltage signal

113‧‧‧ voltage sensing circuit

114‧‧‧current signal

115‧‧‧ Current sensing circuit

130‧‧‧Display signal

131‧‧‧Input layer

133‧‧‧ hidden layer

135‧‧‧ Output layer

The nodes of the input layer 1311, 1312, 1313‧‧

1331, 1332, 1333, 1334, ..., 133 (n-2), 133 (n-1), 133n‧‧‧ hidden layer nodes

1350‧‧‧ Battery residual status

1351‧‧‧ nodes of the output layer

150‧‧‧Target Training File

151‧‧‧ Training database

170‧‧‧temperature

172‧‧‧ terminal voltage

174‧‧‧Discharge current

176‧‧‧Model of battery

1 is a schematic view of an electronic device according to a first embodiment of the present invention.

Fig. 2 is a schematic view showing a neural network such as an electronic device according to a first embodiment of the present invention.

Figure 3 is a flow chart for evaluating the state of residual charge of a battery in accordance with a second embodiment of the present invention.

Figure 4 is a flow chart showing the temperature, terminal voltage and discharge current of the battery taken in the second embodiment of the present invention.

Figure 5 is a flow chart showing the generation of temperature signals, voltage signals, and current signals corresponding to the temperature, terminal voltage, and discharge current of the battery in accordance with the second embodiment of the present invention.

Figure 6 is a flow chart showing the output temperature signal, voltage signal and current signal of the second embodiment of the present invention.

Figure 7 is an enhanced particle swarm optimization for the third embodiment of the present invention. (enhanced particle swarm optimization; EPSO) algorithm to calculate the flow chart of the training data file.

The present invention is not limited by the embodiment, and the embodiment of the present invention is not intended to limit the invention to any specific environment, application or special mode as described in the embodiments. Therefore, the description of the embodiments is merely illustrative of the invention and is not intended to limit the invention. It should be noted that, in the following embodiments and drawings, components that are not directly related to the present invention have been omitted and are not shown; and the dimensional relationships between the components in the drawings are merely for ease of understanding and are not intended to limit the actual ratio.

The first embodiment of the present invention is shown in Fig. 1. 1 is a schematic diagram of an electronic device 1. The electronic device 1 includes a sensing module 11, a processor 13, a storage unit 15, a battery 17, and a display unit 19. The sensing module 11 includes a temperature sensing circuit 111, a voltage sensing circuit 113, and a current sensing circuit 115. The temperature sensing circuit 111, the voltage sensing circuit 113, and the current sensing circuit 115 are electrically connected to the processor 13 and the battery 17, respectively. In other words, the sensing module 11 is electrically connected to the processor 13 and the battery 17.

The processor 13 is electrically connected to the storage unit 15, the battery 17, and the display unit 19, which has a radial basis function (RBF)-like neural network as shown in FIG. The storage unit 15 stores a training database 151. The training database 151 contains a plurality of training data files calculated by an EPSO algorithm. The training data files of the training database 151 correspond to batteries of various models. In this embodiment, the processor 13 and the storage unit 15 are two separate components and are electrically connected to each other. However, in other embodiments, the storage unit 15 can be embedded in the processor 13 by an integrated circuit process, and is not limited to the two separate components described in this embodiment.

The calculation of the aforementioned training data file by the EPSO algorithm includes the following steps. Step (1): setting the number of particles; step (2): defining each dimension component of the position vector of each particle; step (3): initializing the position vector of each particle in a random manner; step (4): Calculate one of the evaluation functions for each particle with a mean squared error; Step (5): Record a region of each particle The best solution (personal best; pbest) and one of the best global solutions (global best; gbest); step (6): correct one of the particle motion speed and position vector; step (7): according to each The evaluation function of the particle selects the particle to be mutated; step (8): corrects the velocity of the particle to be mutated and the position vector; step (9): determines whether the end condition is met, and if so, the position vector of each particle The dimensions of each dimension are recorded as a training data file; if not, proceed to step (4).

In detail, the foregoing steps (4) to (8) are completed. The calculation of the second iteration. Wherein, in step (1), the larger the number of particles to be set, the more time is needed to calculate the training data file by the EPSO algorithm to obtain the overall optimal solution; otherwise, the smaller the number of particles to be set, the less Less time to calculate the training data file by EPSO algorithm, but it is more difficult to obtain the overall optimal solution. In step (2), each dimension component of the position vector of each particle is a central position, a width, and a node weight of one of the RBF-like neural networks. In step (5), in each iteration of each particle, one of the least evaluation functions of each of the particles is selected as the optimal solution for the region; and among the optimal solutions for all regions, the smallest one is selected. The evaluation function is the overall optimal solution. In step (6), the motion velocity and position vector of each particle are corrected according to the regional optimal solution and the overall optimal solution. In step (9), the center position, width and node weight of the Gaussian function of the RBF-like neural network are recorded as training data files.

The RBF-like neural network shown in Figure 2 contains an input layer 131, a hidden A layer 133 and an output layer 135. The input layer 131 includes three nodes 1311, 1312, 1313. The hidden layer 133 includes a plurality of nodes 1331, 1332, 1333, 1334, ..., 133(n-2), 133(n-1), 133n. The output layer 135 includes a node 1351. In the present embodiment, the number of nodes of the hidden layer 133 is set by an orthogonal least squares (OLS) method. However, in other embodiments, the number of nodes of the hidden layer 133 can be set in different manners, and is not limited to the OLS method described in this embodiment.

The present embodiment will be further described below with reference to FIG. 1 and FIG. The electronic device 1 evaluates the operational flow of the state of the residual state of the battery 17. When the electronic device 1 In operation, the temperature sensing circuit 111 draws a temperature 170 from the battery 17; the voltage sensing circuit 113 draws a voltage 172 from the battery 17; and the current sensing circuit 115 draws a discharge current 174 from the battery 17. Subsequently, the temperature sensing circuit 111 generates a temperature signal 110 corresponding to the temperature 170; the voltage sensing circuit 113 generates a voltage signal 112 corresponding to the terminal voltage 172; and the current sensing circuit 115 generates a current signal corresponding to the discharging current 174. 114. Finally, the temperature sensing circuit 111, the voltage sensing circuit 113, and the current sensing circuit 115 output the temperature signal 110, the voltage signal 112, and the current signal 114 to the processor 13, respectively.

When the processor 13 receives the temperature signal 110, the voltage signal 112, and the power After the stream signal 114, one of the models 176 of the battery 17 is retrieved, and the target training file 150 corresponding to the model 176 of the battery 17 is retrieved from the training data file of the training database 151 of the storage unit 15. Then, the processor 13 calculates a residual state 1350 of the battery 17 according to the target training file 150, the temperature signal 110, the voltage signal 112, and the current signal 114 by using a neural network such as the second figure. In other words, the processor 13 calculates the residual state 1350 of the battery 17 based on the training database 151, the temperature signal 110, the voltage signal 112, and the current signal 114 by using a neural network such as the second diagram. Finally, the processor 13 outputs a display signal 130 having a residual state 1350 of the battery 17.

In detail, when the processor 13 receives the temperature signal 110, the voltage signal After 112 and the current signal 114, the temperature signal 110 is input to the node 1311 of the input layer 131 of the neural network, the voltage signal 112 is input to the node 1312 of the input layer 131 of the neural network, and the current signal 114 is input to A node 1313 of the input layer 131 of the neural network. Then, each node 1331, 1332, 1333, 1334, ..., 133(n-2), 133(n-1), 133n of the hidden layer 133 of the neural network receives the temperature signal 110 and the voltage signal 112, respectively. And the current signal 114 is calculated according to the center position, the width, and the node weight record of the Gaussian function of the RBF-like neural network of the target training file 150. Finally, node 1351 of output layer 135 of the neural network is outputting the residual state 1350 of battery 17.

When the display unit 19 receives the battery 17 output by the processor 13 After the display signal 130 of the residual state 1350, the battery is displayed according to the display signal 130. 17 residual power status 1350.

A second embodiment of the present invention is shown in FIG. 3, which is for evaluating an electric A flowchart of a method for determining the state of the residual state of the battery, the method for estimating the state of the residual state of the battery described in this embodiment can be applied to an electronic device, such as the electronic device 1 described in the first embodiment. The electronic device includes a sensing module, a storage unit, and a processor. The sensing module is electrically connected to the battery. The processor is electrically connected to the battery, the sensing module and the storage unit and has a type of neural network. The storage unit is configured to store a training database containing a plurality of training data files calculated by an enhanced particle swarm optimization algorithm.

Method for evaluating the state of residual state of a battery as described in the second embodiment The following steps are included. First, in step 301, the sensing module captures one of the temperature of the battery, the voltage of one end, and a discharge current. In step 303, the sensing module generates a temperature signal, a voltage signal, and a current signal corresponding to the temperature, the terminal voltage, and the discharge current of the battery. In step 305, the sensing module outputs a temperature signal, a voltage signal, and a current signal. In step 307, the processor receives the temperature signal, the voltage signal, and the current signal. In step 309, the microprocessor captures one of the battery models. In step 311, the microprocessor retrieves the target training file corresponding to one of the battery models from the training data file of the training database stored in the storage unit according to the model of the battery. Next, in step 313, the processor calculates the residual state of the battery according to the target training file, the temperature signal, the voltage signal, and the current signal by using a neural network. Next, in step 315, the processor outputs a display signal having a state of residual state of the battery. In step 317, the display unit receives the display signal. Finally, in step 319, the display unit displays the state of the residual power of the battery according to the display signal.

It should be noted that in the method for evaluating the residual state of the battery according to the second embodiment, the sensing module of the electronic device includes a temperature sensing circuit, a voltage sensing circuit, and a current sensing circuit. The temperature sensing circuit, the voltage sensing circuit, and the current sensing circuit are electrically connected to the battery and the processor, respectively. The step 301 of sensing the module to extract one of the temperature of the battery, the voltage of one end, and a discharge current further includes the steps as shown in FIG. First, in step 401, the temperature sensing circuit draws the temperature of the battery. Next, in step 403, the voltage sensing circuit draws the terminal voltage of the battery, and finally, In step 405, the current sensing circuit draws the discharge current of the battery.

The sensing module generates a temperature, a terminal voltage, and a discharge corresponding to the battery The step 303 of one of the current temperature signal, a voltage signal, and a current signal further includes the steps as shown in FIG. First, in step 501, the temperature sensing circuit generates a temperature signal corresponding to the temperature of the battery. Next, in step 503, the voltage sensing circuit generates a voltage signal corresponding to the terminal voltage of the battery. Finally, in step 505, the current sensing circuit generates a current signal corresponding to the discharge current of the battery.

The sensing module outputs temperature signals, voltage signals, and current signals. Step 305 further includes the steps as shown in FIG. First, in step 601, the temperature sensing circuit outputs a temperature signal. Next, in step 603, the voltage sensing circuit outputs a voltage signal. Finally, in step 605, the current sensing circuit outputs a current signal.

In addition to the above steps, the second embodiment evaluates the residual state of the battery The method of the present invention can also perform all the operations described in the electronic device 1 of the first embodiment and have all the functions corresponding thereto, and those skilled in the art can directly understand the state of the residual battery of the evaluation battery of the second embodiment. The method of performing the operations and having such functions based on the disclosed content of the electronic device 1 of the first embodiment is not described herein.

Specifically, the residual power of the battery is evaluated in the second embodiment. The method can be executed by a computer program product. When an electronic device loads the computer program product and executes a plurality of instructions included in the computer program product, the state of the residual battery of the evaluation battery described in the second embodiment can be completed. The method. The aforementioned computer program product can be stored in a computer readable recording medium, such as read only memory (ROM), flash memory, floppy disk, hard disk, optical disk, flash drive, tape, network available Access to the database or any other storage medium known to those skilled in the art and having the same function.

A third embodiment of the present invention is shown in Fig. 7, which is an EPSO The algorithm calculates a flow chart of the training data file, which includes the following steps. First, in step 701, the number of particles is set. In step 703, each dimension component of a position vector of each particle is defined, wherein each dimension component of the position vector of each particle is a center position, a width, and a node of a Gaussian function of the neural network. Weight. In step 705, the position vector of each particle is initialized in a random manner. In step 707, one of the evaluation functions of each particle is calculated with the mean square error. In step 709, an optimal solution for one of the regions of each particle and an overall optimal solution for all of the particles are recorded, wherein, in each iteration of each particle, a minimum evaluation of one of the previous iterations of each particle is selected. The function is the best solution for its region; and in the best solution for all regions, a minimum evaluation function is chosen as the overall optimal solution.

Then, step 711 is performed to correct one of the moving speeds of each particle. And a position vector, wherein the velocity of each particle and the position vector are corrected according to the optimal solution of the region and the overall optimal solution. In step 713, the particles to be mutated are selected according to the evaluation function of each particle. In step 715, the motion velocity and the position vector of the particles to be abrupt are corrected. Next, step 717 is executed to determine whether the end condition is met.

In step 717, if the end condition has been met, step 719 is performed. Each dimension component of the position vector of each particle is recorded as a training data file. In other words, the center position, width, and node weight of the Gaussian function of the neural network are recorded as training data files. In step 717, if the end condition has not been met, step 707 is continued.

Specifically, the EPSO algorithm is used to calculate the training described in the third embodiment. The training data file can be executed by a computer program product. After a device loads the computer program product and executes a plurality of instructions included in the computer program product, the training data calculated by the EPSO algorithm described in the third embodiment can be completed. file. The aforementioned computer program product can be stored in a computer readable recording medium such as a read only memory, a flash memory, a floppy disk, a hard disk, a compact disk, a flash drive, a magnetic tape, a database accessible by the network, or familiar with Any other storage medium known to the skilled artisan and having the same function.

In summary, the present invention is used to evaluate the state of residual power of a battery. The electronic device and method can calculate the residual state of the battery by using the neural network of the processor and the pre-stored training database after the sensing module draws the temperature, the terminal voltage and the discharge current of the battery. Finally, a display signal having the state of the residual state of the battery calculated above is output. Accordingly, the present invention is used to evaluate the residual current of the battery. The electronic device and method of the quantity state can effectively overcome the problem that the accuracy of the estimated battery residual state is lowered due to the influence of the external environment in the prior art. At the same time, through the pre-stored training database, the speed of calculating the residual state of the battery by the neural network can also be improved.

The embodiments described above are only intended to illustrate the embodiments of the present invention, and to explain the technical features of the present invention, and are not intended to limit the scope of protection of the present invention. Any changes or equivalents that can be easily made by those skilled in the art are within the scope of the invention. The scope of the invention should be determined by the scope of the claims.

1‧‧‧Electronic device

11‧‧‧Sensor module

13‧‧‧ Processor

15‧‧‧ storage unit

17‧‧‧Battery

19‧‧‧Display unit

110‧‧‧temperature signal

111‧‧‧Temperature sensing circuit

112‧‧‧Voltage signal

113‧‧‧ voltage sensing circuit

114‧‧‧current signal

115‧‧‧ Current sensing circuit

130‧‧‧Display signal

150‧‧‧Target Training File

151‧‧‧ Training database

170‧‧‧temperature

172‧‧‧ terminal voltage

174‧‧‧Discharge current

176‧‧‧Model of battery

Claims (8)

An electronic device for evaluating a residual state of a battery, comprising: a sensing module electrically connected to the battery for extracting a temperature, a voltage of one end, and a discharge current of the battery, corresponding to the a temperature signal, a voltage signal, and a current signal of the battery, and outputting the temperature signal, the voltage signal, and the current signal; and a storage unit for storing a training database, the training The data library includes a plurality of training data files calculated by an enhanced particle swarm optimization (EPSO) algorithm; and a processor electrically connected to the battery, the sensing module, and the storage unit, a type of neural network for receiving the temperature signal, the voltage signal, and the current signal, and extracting a model of the battery, and selecting, according to the model of the battery, a model corresponding to the battery from the training data file. a target training file, through which the neural network, the training file, the temperature signal, the voltage signal, and Calculating the residual current signal of the state of charge of the battery, and outputs a residue of one of the battery charge status display signal. The electronic device of claim 1, wherein the sensing module comprises: a temperature sensing circuit electrically connected to the battery and the processor for capturing the temperature of the battery to generate a battery corresponding to the battery The temperature signal of the temperature, and outputting the temperature signal; a voltage sensing circuit electrically connecting the battery and the processor for capturing the voltage of the terminal of the battery to generate the voltage signal corresponding to the terminal voltage of the battery, and Outputting the voltage signal; and a current sensing circuit electrically connected to the battery and the processor for capturing a discharge current of the battery, generating the current signal corresponding to a discharge current of the battery, and outputting the current signal . The electronic device of claim 1, further comprising a display unit electrically connected to the processor for receiving the display signal and displaying the residual state of the battery according to the display signal. An electronic device for evaluating a residual state of a battery, comprising: a sensing module electrically connected to the battery for capturing temperature, one end voltage and a discharging current of the battery, generating a temperature signal and a voltage signal corresponding to the temperature, the terminal voltage and the discharging current of the battery And a current signal, and outputting the temperature signal, the voltage signal and the current signal; a storage unit for storing a training database; and a processor electrically connecting the battery, the sensing module, and the storage a unit having a radial basis function (RBF)-like neural network for receiving the temperature signal, the voltage signal, and the current signal, by the web-based base function neural network, according to the training The data base, the temperature signal, the voltage signal, and the current signal calculate a state of residual power of the battery, and output a display signal having a state of residual power of the battery. A method for evaluating a residual state of a battery, the method is used for an electronic device, the electronic device includes a sensing module, a storage unit, and a processor, the sensing module is electrically connected to the battery, The processor is electrically connected to the battery, the sensing module and the storage unit and has a neural network. The storage unit stores a training database, and the training database includes a complex number calculated by an enhanced particle swarm optimization algorithm. a training data file, the method comprising the steps of: causing the sensing module to capture a temperature, a voltage of one end, and a discharge current of the battery; causing the sensing module to generate a temperature, a terminal voltage, and a discharge corresponding to the battery a temperature signal, a voltage signal, and a current signal; the sensing module outputs the temperature signal, the voltage signal, and the current signal; and the processor receives the temperature signal, the voltage signal, and the current signal; Having the processor capture a model of the battery; causing the processor to retrieve the corresponding data from the training data files according to the model of the battery One of the model target training files; the processor is configured to calculate the residual state of the battery according to the target training file, the temperature signal, the voltage signal, and the current signal by using the neural network; and The output has one of the display states of the residual state of the battery. The method of claim 5, the sensing module of the electronic device includes a temperature sensing a circuit, a voltage sensing circuit, and a current sensing circuit, the temperature sensing circuit, the voltage sensing circuit, and the current sensing circuit are electrically connected to the battery and the processor respectively, so that the sensing module captures The step of generating a temperature signal, a voltage at one end, and a discharge current, generating and outputting a temperature signal, a voltage signal, and a current signal corresponding to the temperature, the terminal voltage, and the discharge current of the battery further comprises the steps of: The sensing circuit captures the temperature of the battery; the temperature sensing circuit generates the temperature signal corresponding to the temperature of the battery; causes the temperature sensing circuit to output the temperature signal; and causes the voltage sensing circuit to capture the end of the battery The voltage sensing circuit generates the voltage signal corresponding to the terminal voltage of the battery; causing the voltage sensing circuit to output the voltage signal; causing the current sensing circuit to capture the discharge current of the battery; and causing the current sensing The circuit generates the current signal corresponding to the discharge current of the battery; and causes the current sensing circuit to output the current signal. The method of claim 5, the electronic device further comprising a display unit electrically connected to the processor, the method further comprising the steps of: causing the display unit to receive the display signal; and causing the display unit to display the signal according to the display signal The status of the battery's residual capacity is displayed. A method for evaluating a residual state of a battery, the method is used for an electronic device, the electronic device includes a sensing module, a storage unit, and a processor, the sensing module is electrically connected to the battery, The processor is electrically connected to the battery, the sensing module and the storage unit and has a web-based function-like neural network, the storage unit stores a training database, and the method comprises the following steps: Taking a temperature, a voltage of one end, and a discharge current of the battery; causing the sensing module to generate a temperature signal, a voltage signal, and a current signal corresponding to the temperature, the terminal voltage, and the discharge current of the battery; The measuring module outputs the temperature signal, the voltage signal and the current signal; and the processor receives the temperature signal, the voltage signal and the current signal; Having the processor calculate the residual state of the battery according to the training database, the temperature signal, the voltage signal, and the current signal by using the web-based function-like neural network; and causing the processor to output the battery One of the residual power states displays a signal.