A new energy vehicle charging and discharging control method and system

Abstract

The application relates to the new energy technology field, in particular to a new energy vehicle charging and discharging control method and system. The method comprises the following steps: obtaining an effective reference value corresponding to a current charging sequence in current data; if the effective reference value is greater than a preset reference threshold value, a target current charging sequence is obtained as a reference current sequence; based on the reference current sequence, the current charging current is abnormally judged, all abnormal current time points corresponding to the current charging current are marked as an abnormal current sequence, an abnormal temperature sequence corresponding to the abnormal current sequence in temperature data is obtained; a temperature prediction network is constructed and continuous prediction is carried out; a predicted temperature rise speed corresponding to a predicted temperature is calculated; and if the predicted temperature rise speed is greater than a preset temperature rise speed threshold value, the charging current is proportionally reduced. The application can effectively and timely prevent battery overheating, reduce the risk of thermal runaway, and ensure battery safety.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of new energy, in particular to a new energy vehicle charging and discharging control method and system. BACKGROUND

[0002] The power battery thermal management is the core of the safe operation of new energy vehicles. The heat generated by the internal winding core needs to pass through multiple layers of material thermal resistance to the shell. In addition, the battery has large thermal capacity and the thermal conductivity decreases after aging, which causes the internal temperature to lag and be 40-50℃ higher than the surface. The BMS is difficult to find the signs of thermal runaway in time relying on the shell sensor. The sampling of the charging current is affected by the ADC quantization error and dynamic delay, and the error can reach 50% in the high-voltage small-current scene, which further weakens the reliability of the temperature rise prediction. Therefore, a high-precision prediction method based on the fusion of historical sequences and real-time data is needed.

[0003] The prior art cannot overcome the internal-surface temperature lag of the battery and the heat transfer delay caused by aging. The surface temperature measurement of the BMS leads to the lag of the thermal runaway early warning. At the same time, the sampling error and sudden change of the charging current make it difficult to identify the abnormal current in real time, and thus the subsequent temperature rise cannot be accurately predicted and the charging current cannot be adjusted in advance, which has high safety risk. SUMMARY

[0004] In order to accurately predict the subsequent temperature rise and adjust the charging current in advance, the present application provides a new energy vehicle charging and discharging control method and system.

[0005] In the first aspect, the present application provides a new energy vehicle charging and discharging control method, which adopts the following technical scheme:

[0006] Collect and obtain the current data and temperature data of the battery of the new energy vehicle;

[0007] Obtain the effective reference value corresponding to the current charging sequence in the current data. If the effective reference value is greater than the preset reference threshold value, obtain the corresponding target current charging sequence as the reference current sequence;

[0008] Based on the reference current sequence, the current charging current is judged for abnormality. If there is abnormality, mark all abnormal current time corresponding to the current charging current as an abnormal current sequence, and obtain the abnormal temperature sequence corresponding to the abnormal current sequence in the temperature data;

[0009] A temperature prediction network is constructed and continuous prediction is performed. The input sample is the abnormal current sequence and the abnormal temperature sequence, and the output is the temperature at the next time of the abnormal current sequence as the predicted temperature;

[0010] The predicted temperature corresponding to the predicted temperature rise speed is calculated. If the predicted temperature rise speed is greater than the preset temperature rise speed threshold value, the charging current is reduced in proportion.

[0011] Optionally, the effective reference value of the current charging sequence satisfies the following formula:

[0012] , wherein, is the effective reference value, is the difference sequence of the current sequence collected in the i th charging and the current sequence collected in the first charging, is the standard deviation, is a positive integer greater than 1. Optionally, the abnormality of the current charging current is judged based on the reference current sequence, including:

[0013] The abnormality degree of the current charging current at any real time satisfies the following formula:

[0014]

[0015] , wherein, is the current abnormality degree of the current charging current at the i th moment, is the reference increment of the current at the i th moment in the previous charging process of the current charging, is the current data collected in the i th charging process of the n th charging, is the current data at the i th moment in the previous charging process of the current charging, m is the total number of charging, and x is the serial number;

[0016] , wherein, is the reference current data at the corresponding moment in the reference current sequence, is the difference between the current moment and the corresponding reference current data, When the value is greater than 0, the reference increment of the current at the current moment and the previous two continuous moments appears continuous growth, and the current has abnormal risk, otherwise, the current has no abnormal risk.

[0017] Optionally, all abnormal current moments corresponding to the current charging current are marked as an abnormal current sequence, including: marking all abnormal current moments, and taking continuous abnormal current moments as an abnormal current sequence, the abnormal current sequence contains the current at the first two moments when the current abnormality is judged.

[0018] Optionally, a temperature prediction network is constructed, including:

[0019] In response to the loss value of the loss function being less than a preset value or reaching a preset training number, a trained temperature prediction network is obtained. The loss function satisfies the following formula:

[0020] , wherein, is the loss function value, c is the total number of current charging,​​ the temperature at the a-th moment, to predict the temperature at the a-th moment, s is the temperature source at the a-th moment, b is the total number of training, as a time weight participates in the calculation of the loss function.

[0021] Optionally, continuous prediction is performed, including:

[0022] The abnormal current sequence is input into the temperature prediction network for continuous prediction to obtain temperatures corresponding to multiple future moments. The abnormal current sequence corresponding to each current abnormal moment is input into the temperature prediction network for continuous prediction to obtain the predicted temperature.

[0023] Optionally, the charging current is reduced in equal proportion, including: reducing the current current value by 0.2 times every moment until the predicted temperature rise speed is less than a preset temperature rise speed threshold.

[0024] In a second aspect, the application provides a new energy vehicle charging and discharging control system, which adopts the following technical scheme:

[0025] The acquisition module is configured to acquire and obtain current data and temperature data of the new energy vehicle battery.

[0026] The reference current acquisition module is configured to obtain an effective reference value corresponding to a current charging sequence in the current data, and if the effective reference value is greater than a preset reference threshold, a target current charging sequence corresponding to the effective reference value is obtained as a reference current sequence.

[0027] The abnormality judgment module is configured to construct a temperature prediction network and perform continuous prediction, the input sample is the abnormal current sequence and the abnormal temperature sequence, and the output is a temperature at a next moment of the abnormal current sequence as a predicted temperature.

[0028] The charging current control module is configured to calculate a predicted temperature rise speed corresponding to the predicted temperature, and if the predicted temperature rise speed is greater than a preset temperature rise speed threshold, the charging current is reduced in equal proportion.

[0029] The application has the following technical effects: by extracting a historical current charging sequence similar to the current charging environment and a corresponding effective reference value, obtaining a target current charging sequence meeting a preset reference threshold as a reference current sequence, and calculating an effective reference value and screening a benchmark, an abnormal current sequence can be more accurately identified, so that possible thermal runaway signs inside the battery can be discovered in advance, the problem of thermal runaway early warning lag caused by temperature lag can be avoided, the LSTM network trained by the abnormal current sequence is used to predict temperature rise, and when the predicted temperature rise speed exceeds the standard, the charging current is immediately reduced in equal proportion, the battery overheating is effectively prevented, the thermal runaway risk is reduced, and the battery safety is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 is a flowchart of steps S1-S5 in a new energy vehicle charging and discharging control method of the application. DETAILED DESCRIPTION

[0031] It should be clear that the application scenario of the application is as follows: due to the physical blockage of the internal heat transfer path of the battery of the new energy vehicle and the inherent limitation of the external monitoring system, the battery temperature rise is lagging. The core of the battery heat generation is located in the internal winding core (electrode and electrolyte interface), and the heat needs to overcome the thermal resistance of multiple materials such as electrodes, separators, and current collectors to be transmitted to the surface, and the battery itself has a large thermal capacity (such as the thermal lag time constant of an 80mm thick battery is as high as 1134 seconds), which causes the temperature response to be significantly delayed. At the same time, the temperature sensor of the BMS (Battery Management System, battery management system) can only capture the surface temperature of the battery shell, but there is a large difference between the internal actual temperature and the surface value (the internal temperature can reach 120℃ before thermal runaway, and the surface temperature is only 70~80℃, the temperature difference is 40~50℃), which is due to the loss in the heat transfer process (such as heat exchange between the shell and the environment) and the non-uniformity of the internal temperature distribution. In addition, after the battery ages, the electrolyte is consumed and the byproduct gas forms a thermal insulation layer between the electrodes (the thermal conductivity coefficient decreases by 29%), further prolonging the heat transfer time. These factors together cause the surface monitoring to be unable to timely reflect the internal real temperature rise, threatening the timeliness of thermal runaway protection, and the temperature has risen, but the current has not been reduced to protect the battery in advance. Therefore, the battery temperature needs to be predicted.

[0032] Further, limited by the dynamic response lag of the sampling system (>100ms), the current mutation will be missed, and the ADC (Analog-to-Digital Converter, analog-to-digital converter) quantization error will be up to 50% in the high-voltage small-current (such as DC1000V / 1A) scene due to insufficient signal-to-noise ratio. First, the reference current sequence of the current data at the current time is obtained, which is used for subsequent abnormality judgment process of the current time, to reduce the influence of the above error.

[0033] In order to complete the charging and discharging control of the new energy vehicle, the time point that needs to be controlled needs to be obtained.

[0034] The embodiment of the application discloses a new energy vehicle charging and discharging control method, referring to Figure 1 , comprising the following steps:

[0035] S1: collecting and obtaining current data and temperature data of the battery of the new energy vehicle.

[0036] It should be noted that the current data refers to the current value measured by the current sensor during the battery charging and discharging process. These data are used to evaluate the charging and discharging rate of the battery, the change of the electric quantity, and whether there is an abnormal current (such as overcurrent, sudden change, etc.). Specifically, during the battery charging process, the current flows from the charging device to the battery. A high-precision current sensor (such as a Hall effect sensor, a shunt, etc.) can be used to measure the current, and the analog current signal measured by the sensor is converted into a digital signal for subsequent processing and analysis, which is not specifically limited in the present application.

[0037] In addition, the temperature data refers to the temperature value of the inside and outside of the battery measured by the temperature sensor during the battery charging and discharging process. These data are used to evaluate the thermal state of the battery and to discover the risk of thermal runaway in time. Among them, the temperature of the battery internal winding core is usually measured by installing a temperature sensor inside the battery, the temperature of the surface of the battery shell is usually measured by installing a temperature sensor on the battery shell, and the temperature of the environment where the battery is located is usually measured by installing a temperature sensor around the battery. Furthermore, a high-precision temperature sensor (such as a thermistor, a thermocouple, etc.) can be used to measure the temperature, and the analog temperature signal measured by the sensor is converted into a digital signal for subsequent processing and analysis, which is not specifically limited in the present application.

[0038] S2: Obtain the effective reference value corresponding to the current charging sequence in the current data, and if the effective reference value is greater than the preset reference threshold, obtain the corresponding target current charging sequence as the reference current sequence.

[0039] It should be noted that in order to avoid the abnormal change of the current caused by the aging of the battery, it is necessary to obtain the effective reference current sequence. The current charging sequence refers to the sequence of current data recorded in time sequence during the battery charging process, which reflects the current change of the battery during the charging process, including the size, fluctuation, and sudden change of the charging current. The effective reference value is a value obtained by analyzing and calculating the historical charging data, which is used to evaluate the reliability and reference value of the current charging sequence. It reflects the similarity between the current charging sequence and the historical charging sequence, and the stability of the current charging sequence in a statistical sense. Furthermore, the current sequence of the previous (historical) charging is taken as the reference current sequence, and the current data corresponding to the charging times similar to the reference current sequence is obtained.

[0040] In one embodiment, the current charging times is the 100th time, and the current sequence of the 99th charging is the reference current sequence. The effective reference value of one current charging sequence satisfies the following polynomial: wherein, is the effective reference value, is the current sequence collected in the 99th time, and is the current sequence collected in the 100th time. a difference sequence of the current sequence collected in the kth time, is a standard deviation. is a positive integer greater than 1. Wherein, is the current sequence collected in the kth time, is the current sequence collected in the k-1th time and the current sequence collected in the kth time. is the fluctuation degree of the difference sequence of the current sequence collected in the kth time, is the lower the correlation degree between the current sequence collected in the kth time and the current sequence collected in the k-1th time, that is, the worse the effect of the current sequence collected in the k-1th time as a reference current sequence participating in subsequent calculation. In addition, is the current sequence collected in the kth time, is the current sequence collected in the kth time and the current sequence collected in the k-1th time. is the ratio of the number of times of the kth collection and the k-1th collection, the smaller the value, the closer the charging times to the present, the higher the reliability of the data, and vice versa, the lower the reliability of the data.

[0041] Further, a reference value threshold is set, and the preset reference threshold is a key parameter for judging whether the current collected current sequence can be used as a reliable reference current sequence, that is, a preset value for evaluating the effectiveness and reliability of the current collected current sequence, which can be a benchmark value set based on historical data and experience, for screening out reliable reference current sequences similar to the current charging environment. The target current charging sequence is the current charging sequence corresponding to the effective reference value of the current data greater than the preset reference threshold.

[0042] As an example, when the effective reference value of the new energy vehicle is greater than the reference value threshold, it means that the current collected current sequence can be used as a reference current sequence. Thus, the reference current sequence of the current charging is obtained.

[0043] S3: Abnormal judgment is made on the current charging current based on the reference current sequence, and if there is an abnormality, the current charging current corresponding to all abnormal current time points is marked as an abnormal current sequence, and an abnormal temperature sequence corresponding to the abnormal current sequence time point in the temperature data is obtained.

[0044] In an embodiment, the current charging current is compared with the reference current sequence in real time, that is, the abnormality of the current charging current is judged based on the reference current sequence. Wherein, the reference current sequence is the reliable reference current sequence screened from the historical charging data, which is used as a reference standard for the current charging process. These sequences reflect the current change when the battery is normally charged in a similar charging environment. When the reference current sequence is obtained, it needs to be calculated based on the reference current.

[0045] As an example, the current charging process is collected in real time, the current value at the corresponding time is extracted from the reference current sequence, the ratio of the current charging current to the reference current is calculated, and the increase of the current charging current is evaluated.

[0046] Further, if there is an anomaly, the abnormal current sequence is marked and the corresponding temperature is recorded. Based on all the reference current sequences of the current charging, the current charging current real-time anomaly evaluation is obtained. These reference current sequences have similar charging environments, that is, the current time of the current charging should be consistent with the charging current at the corresponding time in the reference current sequence. The current judgment is constructed, and whether the current charging current is abnormal at the current time is obtained according to the current data sequence of each charging.

[0047] As an example, the current charging process is collected in real time, the current value at the corresponding time is extracted from the reference current sequence, the ratio of the current charging current to the reference current is calculated, and the increase of the current charging current is evaluated. , wherein, is the current abnormality degree at the i-th time, is the reference increase of the current at the i-th time in the previous charging process, is the current data at the i-th time collected in the n-th charging, is the current data at the i-th time in the previous charging process, m is the total number of charges, and x is the serial number.

[0048] , wherein, is the current reference data at the corresponding time in the reference current sequence, is the difference between the current time and the corresponding current reference data. The larger the value, the greater the sudden increase of the current at the current time. When the value is greater than 0, it proves that the reference increase of the current at the current time and the previous two consecutive times has a continuous increase, and the current has an abnormal risk. Otherwise, the current does not have an abnormal risk. Thus, the abnormality of the current time in the current charging can be judged.

[0049] Thus, the real-time current is judged for abnormality, all abnormal current times are marked, and the continuous abnormal current times are taken as an abnormal current sequence (including the current of the previous two times when the current abnormality time is first judged), and the temperature sequence at the corresponding time of the abnormal current sequence is obtained.

[0050] S4: Construct a temperature prediction network and continuously predict, the input sample is the abnormal current sequence and the abnormal temperature sequence, and the output is the temperature at the next time of the abnormal current sequence as the predicted temperature.

[0051] It should be noted that when the abnormal time occurs in the current charging process, the temperature data at the time of temperature rise needs to be monitored, and the temperature prediction network is constructed to predict the temperature caused by the abnormal current.

[0052] In an embodiment, a temperature prediction network (LSTM) is constructed, the input sample is the abnormal current sequence and the temperature at the corresponding time, and the output is the temperature at the next time of the abnormal current sequence. Wherein, in response to the loss value of the loss function being less than a preset value (0.01) or reaching a preset training number (5000), a trained temperature prediction network is obtained. The loss function satisfies a coefficient polynomial: , wherein, is the loss function value, c is the total number of the current charging (for example, the total number of the current charging is the 100th charging), is the temperature at the a time, is the predicted temperature at the a time, s is the temperature source at the a time, and b is the total training number. The value can be used as a time weight to participate in the calculation of the loss function, The greater the value, the closer the training sample is to the current charging number, and the less likely the battery is to have obvious degradation. New training samples are continuously involved in the training of the temperature prediction network, and the temperature prediction network is updated.

[0053] As an example, using the temperature prediction network: at the current charging time, the current abnormal time is determined, and the corresponding abnormal current sequence is obtained. The abnormal current sequence is input into the temperature prediction network for continuous prediction to obtain the temperature corresponding to the future multiple times. Each abnormal current sequence corresponding to the current abnormal time is input into the temperature prediction network for continuous prediction to obtain the predicted temperature.

[0054] S5: Calculate the predicted temperature rise speed corresponding to the predicted temperature, and if the predicted temperature rise speed is greater than a preset temperature rise speed threshold, then reduce the charging current by an equal proportion.

[0055] In an embodiment, the predicted temperature rise speed is calculated by subtracting the current temperature at the current time from the predicted temperature divided by the interval time. Further, a temperature rise speed threshold (for example, 0.79℃ / s for lithium iron phosphate and 10.52℃ / s for ternary lithium) is set, and when the temperature rise speed is greater than the corresponding temperature rise speed threshold, the charging current is immediately reduced at this moment. Wherein, the preset temperature rise speed threshold is a pre-set value, representing the maximum temperature rise rate allowed by the battery during charging, in units of degrees Celsius per second (℃ / s), which is not limited in the present application.

[0056] As an example, the reduction mode is to reduce the current value by a constant ratio (0.2) at each time, that is, to reduce the current value by 0.2 times at each time until the predicted temperature rise speed is less than the temperature rise speed threshold, which indicates that the temperature rise is effectively controlled at this time. Assuming that during a certain charging process, the system predicts that the temperature rise speed of the battery is 1.5°C / s, and the preset temperature rise speed threshold is 1.2°C / s. The system will immediately take measures to reduce the charging current by a constant ratio of 20% at each time until the temperature rise speed is reduced to below 1.2°C / s.

[0057] The embodiments of the present application also disclose a new energy automobile charging and discharging control system, comprising:

[0058] The acquisition module is configured to acquire and obtain current data and temperature data of the battery of the new energy automobile.

[0059] The reference current acquisition module is configured to obtain an effective reference value corresponding to the current charging sequence in the current data, and obtain a target current charging sequence corresponding to the effective reference value as a reference current sequence if the effective reference value is greater than a preset reference threshold.

[0060] The abnormality judgment module is configured to construct a temperature prediction network and perform continuous prediction, input samples are an abnormal current sequence and an abnormal temperature sequence, and output is a temperature at a next time of the abnormal current sequence as a predicted temperature.

[0061] The charging current control module is configured to calculate a predicted temperature rise speed corresponding to the predicted temperature, and reduce the charging current by a constant ratio if the predicted temperature rise speed is greater than a preset temperature rise speed threshold.

[0062] The above are preferred embodiments of the present application, and are not intended to limit the protection scope of the present application, therefore: any equivalent changes made according to the structure, shape, principle of the present application should be covered within the protection scope of the present application.

Claims

1. A charging and discharging control method for a new energy vehicle, characterized in that, Includes the following steps: Collect and acquire current and temperature data of new energy vehicle batteries; Obtain the valid reference value corresponding to the current charging sequence in the current data. If the valid reference value is greater than a preset reference threshold, then obtain the corresponding target current charging sequence as a reference current sequence. The valid reference value of the current charging sequence satisfies the following formula: ,in, For valid reference values, For the first The current sequence acquired in the second acquisition and the first acquisition The difference sequence of the current sequence from one acquisition. Standard deviation, It is a positive integer greater than 1; Based on the reference current sequence, anomalies are judged in the current charging current. If anomalies are found, all abnormal current moments corresponding to the current charging current are marked as abnormal current sequences, and the abnormal temperature sequence corresponding to the abnormal current sequence in the temperature data is obtained. A temperature prediction network is constructed and continuous prediction is performed. The input samples are the abnormal current sequence and the abnormal temperature sequence, and the output is the temperature of the abnormal current sequence at the next moment as the predicted temperature. Calculate the predicted temperature rise rate corresponding to the predicted temperature. If the predicted temperature rise rate is greater than the preset temperature rise rate threshold, then reduce the charging current proportionally.

2. The charging and discharging control method for a new energy vehicle according to claim 1, characterized in that, Based on the reference current sequence, anomaly detection of the current charging current is performed, including: The degree of any real-time current anomaly during the current charging process satisfies the following formula: , ,in, The degree of current anomaly at time i. This represents the baseline increase in current at time i during the previous charging process before the current charging. The current data collected at time i during the nth charge. The current data at time i during the previous charging process is given for the current charging, where m is the total number of charging cycles and x is the sequence number. in, To reference the current reference data at the corresponding time point in the current sequence, This represents the difference between the current charging time and the corresponding current reference data. When the value is greater than 0, the baseline increase of the current at the current charging time and the two consecutive times before that shows a continuous increase, indicating an abnormal risk in the current. Conversely, when the value is less than 0, there is no abnormal risk in the current.

3. The charging and discharging control method for a new energy vehicle according to claim 2, characterized in that, Marking all abnormal current moments corresponding to the current charging current as an abnormal current sequence includes: marking all abnormal current moments and taking consecutive abnormal current moments as an abnormal current sequence, wherein the abnormal current sequence includes the current of the two moments before the first time the current abnormality is determined.

4. The charging and discharging control method for a new energy vehicle according to claim 1, characterized in that, Constructing a temperature prediction network includes: The temperature prediction network is obtained when the loss value of the loss function is less than a preset value or when a preset number of training iterations are reached; the loss function satisfies the following formula: ,in, Here, c represents the loss function value, and c represents the total number of charges in the current cycle. The temperature at time a is To predict the temperature at time a, where s is the temperature source at time a (s-th charging), and b is the total number of training iterations. It is used as a time weight in the calculation of the loss function.

5. The charging and discharging control method for a new energy vehicle according to claim 1, characterized in that, Continuous forecasting includes: The abnormal current sequence is input into the temperature prediction network for continuous prediction to obtain the temperature corresponding to multiple future times. The abnormal current sequence corresponding to each abnormal current time is input into the temperature prediction network for continuous prediction to obtain the predicted temperature.

6. The charging and discharging control method for a new energy vehicle according to claim 1, characterized in that, The charging current is reduced proportionally, including reducing the current value by 0.2 times at each moment until the predicted temperature rise rate is less than the preset temperature rise rate threshold.

7. A charging and discharging control system for a new energy vehicle, characterized in that, include: The data acquisition module is used to collect and acquire current and temperature data of new energy vehicle batteries; The reference current acquisition module is used to acquire a valid reference value corresponding to the current charging sequence in the current data. If the valid reference value is greater than a preset reference threshold, the corresponding target current charging sequence is acquired as the reference current sequence. The valid reference value of the current charging sequence satisfies the following formula: ,in, For valid reference values, For the first The current sequence acquired in the second acquisition and the first acquisition The difference sequence of the current sequence from one acquisition. Standard deviation, It is a positive integer greater than 1; An anomaly detection module is used to construct a temperature prediction network and make continuous predictions. The input samples are an abnormal current sequence and an abnormal temperature sequence, and the output is the temperature of the abnormal current sequence at the next moment as the predicted temperature. The charging current control module is used to calculate the predicted temperature rise rate corresponding to the predicted temperature. If the predicted temperature rise rate is greater than the preset temperature rise rate threshold, the charging current is reduced proportionally.

Images