Process for preparing electric battery cells for electric or hybrid vehicles

Abstract

The invention relates to a method for preparing battery cells (1) following mechanical manufacturing, comprising an electrical treatment (F1 to F5) followed by a qualification phase after which the cells are brought to a predetermined final charge level (C0). According to the invention, a statistical model of the cells predicts their capacity by means of measurements of electrical parameters (V, Q, I) of their treatment, or of their physical parameters (°T, weight, etc.), without the need to directly measure this capacity by exploiting a complete discharge. The model is trained beforehand on analogous cells of known capacity. (Figure 2)

Description

Title of the invention: Process for preparing electric battery cells for electric or hybrid vehicles. Technical field of the invention

[0001] The invention relates to rechargeable electric batteries, including those with high electrical capacity, for example of several hundred ampere-hours, suitable for the propulsion of electric or hybrid vehicles.

[0002] It relates more particularly to a process for preparing cells that make up such electric batteries. Here, "preparation" refers to all the operations following the mechanical manufacturing of the cells. These operations consist of an electrical processing phase designed to develop the electrical discharge capacity of the cells, then a qualification phase, called "grading," designed to verify that a required capacity has been achieved, and finally, after rejecting non-conforming cells, bringing the qualified cells to a final state of charge before their removal from the factory. This final state is generally a partial but well-defined state of charge, meeting safety and other criteria. Prior art

[0003] Electrical treatment consists of a series of preliminary applications of electric currents to the cells to induce a chemical transformation of the electrodes, notably by forming a passivation layer, enabling them to store electrical energy at the desired capacity. The applied electric currents may be of constant intensity, constant voltage, or other characteristics. Each application often lasts several hours and may be separated by significant periods of cell rest. The treatment phase can extend over several days.

[0004] When the final capacity of the cells is assumed to have been reached, ordinary preparation processes include, in the qualification phase, a step consisting of fully charging the cells and then fully discharging them. The discharge capacity is estimated based on the current intensity and the time required for this complete discharge, known in English as "grading discharge capacity (GDC)," from the maximum charge state. Finally, a partial recharge must be performed to reach the final charge state mentioned above.

[0005] After the long and energy-intensive electrical treatment, this final group of steps is itself expensive and relatively time-consuming. It involves first using a full charge of the cells just to measure their capacity, and then carrying out the final, at least partial, recharge immediately afterwards. The discharge must be carried out at a moderate rate, for example, C / 3, where, according to a common convention in the field, a discharge rate C corresponds to a complete discharge of the capacity (also denoted C) in one hour. A complete discharge at a rate of C / 3 therefore takes three hours, and the final partial recharge takes approximately one hour, assuming the final state of charge is to be 30% as required by certain standards.

[0006] There is therefore a great economic and ecological interest in shortening and simplifying such preparation processes. Description of the invention

[0007] The invention was designed for this purpose. In its general form, it relates to a method for preparing electric battery cells to power electric or hybrid vehicles, comprising an electrical treatment phase by applying electric currents according to a determined time function, followed by a final qualification phase which includes bringing the cells to a determined final charge level after discharge, characterized in that:

[0008] -a statistical model of cells is used in order to obtain a prediction of a cell discharge capacity (the GDC mentioned above) from input variables, the input variables being derived from parameter measurements on the cells, which are made before or during the treatment phase, and from corresponding input variables and capacity measurements, obtained previously, on similar training cells of the model;

[0009] -said discharge is stopped as soon as the determined final charge rate is reached, and it is a final step in the qualification phase.

[0010] The model therefore uses measurements taken on the cells during the preliminary electrical processing phase to estimate their capacity without the need to physically measure it by imposing a complete discharge from a full charge state: the cells are instead brought directly to their final charge state.

[0011] The savings achieved are substantial: the process can be shortened, by almost two hours if we use the figures from the previous example, by replacing the complete discharge with a partial discharge stopped at the final state, with a corresponding saving of electrical energy. The logistics of cell processing are also improved.

[0012] Since the capacity of the cells must be predicted accurately, and the tolerance on the final state of charge is also narrow, a model that was both accurate and reliable had to be developed. It appeared that a statistical model was satisfactory in this respect, despite the difficulty of correlating the capacity acquired by the cells during the electrical treatment, and the electrical and physical parameters that can then be measured.

[0013] These parameters may include at least one of the following: an applied electric current intensity, a voltage between cell terminals, and a value for the quantities of electricity applied to or withdrawn from the cells during certain stages of the processing phase. They may also include at least one of the following: a temperature, an internal pressure, a solid weight, and an electrolyte weight of the cells.

[0014] Furthermore, it is advantageous for the statistical model to use input variables comprising, in addition to direct measures of said parameters, indirect measures comprising extremal, mean, integrated or time-derived values ​​of said parameters during certain stages of the processing phase.

[0015] However, it did not appear useful to use a very large number of input variables in the model to predict capacity, even though there is no disadvantage in making a large number of measurements which will undergo significant selection to give the input variables of the model, and most of which will not be used.

[0016] Thus, in advantageous embodiments of the process, the parameters are measured periodically during at least part of the processing phase, and the statistical model then performs such a selection of measurements; a number ranging from 10 to 40 input variables of the model is generally sufficient to give accurate predictions of capacity.

[0017] The model can be developed, by artificial intelligence or otherwise, during a training or calibration step. Advantageously, it then includes a selection of the model's input variables by determining, among the measured parameters, those that are best correlated with the capacity values ​​of the training cells.

[0018] The model can be constructed in many different ways. For example, some implementations may use a decision tree, deep learning, or regression algorithm to exploit the variables. Brief description of the figures

[0019] The present invention will be better understood upon reading the description of non-limiting examples of embodiments, with reference to the accompanying figures, which illustrate: • [Fig.l]: a diagram of a cell subjected to electrical processing and measurement; • [Fig. 2]: the evolution of certain characteristic parameters of the process in accordance with prior art, in the form of temporal functions; • [Fig.3]: the modifications to these parameters that the invention implies; • [Fig.4]: the complete installation that carries out the process. Detailed description of an implementation method

[0020] For clarity, identical elements are identified by identical reference numerals from one figure to another. The examples given are purely illustrative; variations are within the grasp of a person skilled in the art without departing from the invention.

[0021] Figure 1 schematically illustrates the installation required for cell production during the electrical processing and qualification phases. A cell 1 has its terminals 2 connected to an electrical circuit 3 equipped, in particular, with a DC generator 4 capable of delivering current in either direction, and means 5 for measuring the electrical parameters in the circuit 3, notably the current intensity and the voltage between the terminals 2. Other measuring means 6 are fitted to the cell 1; these may relate, in particular, to its temperature or internal pressure. Properties of the electrolyte may also be measured. Separate and unshown measuring means may be added to determine, in particular, the dry weight of the cell 1, or that of the electrolyte.

[0022] Figure 2 illustrates the execution of a concrete process of a known kind, often referred to as a "recipe" in the prior art. It is a current time profile illustrated in the form of four diagrams representing, from top to bottom of the figure, three of its parameters: the voltage V measured across terminals 2 by circuit 3, the quantity of electricity Q applied to or drawn from cell 1 during each charging or discharging step (Q = ΣIdt during the step, where I is the applied current intensity), and then the current intensity I itself. The fourth diagram in the figure represents the temperature °T of cell 1, which is also measured. The quantity of electricity Q is measured indirectly, by being calculated proportionally to the current intensity I and the duration of its application according to the formula above.The process lasts several days, denoted J1 to J5, and essentially consists of applying four constant-intensity currents during as many training stages F2, F3, and F4 on day J2, as well as applying a constant-voltage, variable-intensity current during another stage F'; then applying a constant-voltage, variable-intensity current during a stage F' and three final constant-intensity currents during a last training stage F5; and finally, two qualification stages G6 and G7 on the last day J5. Many other methods are known in the prior art, each of which could be improved with this invention. The current applications can generally be consecutive or, conversely, separated by rest periods, with currents flowing in either direction, possibly with different intensities and durations.Non-uniform currents can also be considered, at voltage. constant for example. Resting times can be motivated by the rise in temperature or the swelling of cells, which in turn depend on the chemistry and electrical processing of the cells.

[0023] The first qualification step G6, carried out under negative direction current, corresponds to the complete discharge of cell 1, and the second qualification step G7 to the final partial recharge step, up to the imposed final state, at the charge rate Co. These steps are carried out to completion in ordinary processes.

[0024] But in accordance with the invention, the process is interrupted at the said stopping instant tb where cell 1 already reaches this imposed state during the first qualification step G6, which is therefore abbreviated, while the second qualification step G7 is omitted entirely.

[0025] Figure 3 illustrates the qualification phase when the invention is used, with diagrams of the quantities of electricity Q applied to or drawn from cell 1, and of the current intensity I. It consists of a single step, denoted G'6, where the cells 1 are directly brought to the final charge state Co after the electrical treatment phase, i.e., without going through complete discharge or final recharging.

[0026] The overall installation, characteristic of the invention, is illustrated in [Fig. 4]. The cells 1 undergo the preparation process on a production line 8 up to time tp. A computer server 9 contains a statistical model that receives the values ​​of the input variables from measurements taken by means 5 and 6 on the production line 8 up to time ti, or by other means, elsewhere or earlier. The model then predicts the capacity of each of the cells 1. When a batch 10 of the cells 1 has been processed, their capacity predictions and possibly other characteristics are provided to a first control station 11. The cells 1 deemed non-compliant are identified and withdrawn from the market.The computer server 9 also communicates with a second control station 12, for example to obtain an evaluation of the model, which can periodically be retrained using new calibration cells 1, particularly if its predictions become inaccurate. For this purpose, a training module 13 is used, already used for the preliminary calibration of the model.

[0027] Measurements on cells 1 can be taken periodically or not between the initial time t0 of the process and the end of the treatment phase, and preferably even up to the stopping time tb. However, while these measurements can be very numerous, i.e., up to several tens of thousands for a cell 1, it is necessary for the model to use only a limited number of input variables due to its limited computing power and an expected performance loss. Learning is hampered if there are too many variables. Furthermore, most measures are either of little use or redundant in many cases. A selection process is undertaken to choose as input variables only those best suited to predicting capacity, starting with those most strongly correlated with it. However, the desire to avoid using redundant measures leads to the elimination of variables that are too highly correlated with others.

[0028] The choice of input variables for the model, and of the measurements used to form these input variables, can be made by artificial intelligence during the initial calibration phase of the model, which is performed on cells 1 whose capacity has also been measured according to the ordinary process. The measurements most suitable for predicting capacity are indeed difficult to predict, and they may also differ for each model of cell 1 and each recipe. However, it has been noted that measurements more particularly likely to be included in the model include, at least for some, steps of the electrochemical treatment which are denoted Fi to F5 in the example in [Fig. 2]:

[0029] -quantities of electricity Q: the total value applied or withdrawn during the step, the average value during the step, or the time derivative;

[0030] -voltage V: the maximum during the step, the average value during the step, or the time derivative;

[0031] -internal pressure: the average during the step;

[0032] -temperature °T: the maximum or average during the step;

[0033] -current I: the average value or the time derivative during the step;

[0034] -the dry weight of the cell;

[0035] -the weight, or other physical or chemical characteristics, of the electrolyte.

[0036] This list is not exhaustive. It includes both direct measurements and indirect measurements obtained by processing the direct measurements, for example through integration or numerical differentiation. The input variables may be associated with measurements taken at specific times, time intervals, or variable times (for extrema, for example). It should also be noted that while many of these measurements are taken during the electrical processing, others are taken, or may be taken, before this processing, such as weighings.

[0037] The variables resulting from the selection of measures can number from 10 to approximately 40 in plausible embodiments. These variables are provided as input to the model, which directly determines the capacity of the cell 1 under consideration, the duration of the single qualification step, and the stopping time ti of the preparation process.

[0038] It has been found that such a small number of model input variables is sufficient to predict the capacity of a cell with high accuracy, approximately ±0.2%, including for non-conforming cells, which have capacities much lower than standards, considered as industrial production waste and commonly referred to as "scraps".

[0039] The statistical model can be of various known kinds: decision tree, deep learning, or regression in particular.

[0040] If a decision tree is chosen, the algorithm will consist of successively comparing the input variables, or only some of them depending on the branch of the tree, to respective thresholds, so as to evaluate the capacity of the corresponding cell with increasing precision. The thresholds are estimated during the calibration of the model, as are the variables used in each branch. Typically, the variables used first will be the most discriminating, that is, those most strongly correlated with capacity.

[0041] Among the many variants of decision trees, the random forest seems very interesting here, the decision being made on several trees at the same time.

[0042] Deep learning can consist of using a neural network, for example, a multilayer perceptron. The input variables of each layer are transformed according to the model y = o(Ax+b), where x is the input vector, y the output vector of the layer, o an activation function (a sigmoid, for example), A a weighting matrix, and b a bias matrix. The input variables of the model constitute the input vector of the first layer, and the capability evaluated by the model is found in the output vector of the last layer.

[0043] Regression algorithms are based in particular on classical linear regression or the use of support vectors. They consist of assigning weighting coefficients to the input variables, enabling the direct calculation of capacity.

[0044] The model can also be hybrid or composite, that is to say, made up of several algorithms of different kinds working simultaneously, or of which on the contrary only one is chosen as the best to evaluate the capacity, according to the results of the preliminary calibration, according to the characteristics of the cell.

[0045] The envisaged application for the invention relates to batteries for electric or hybrid vehicles, composed of an assembly of the cells mentioned above.

Claims

Demands

1. A method for preparing electric battery cells (1) for powering electric or hybrid vehicles, comprising an electrical treatment phase (Fi to F5) by applying electric currents according to a determined time function, and then a final qualification phase which includes bringing the cells to a determined final charge rate (Co) after a discharge, characterized in that: -a statistical model (9) of the cells is used in order to obtain a prediction of a discharge capacity of the cells from input variables, the input variables being derived from parameter measurements (V, Q, I, °T) on the cells (1), which are made before or during the treatment phase, and from corresponding input variables and capacity measurements, obtained previously, on similar training cells of the model;-said discharge is stopped (ti) as soon as the determined final charge rate is reached, and it is a final step in the qualification phase.;

2. A method for preparing electric battery cells according to claim 1, characterized in that the parameters include at least one of an applied electric current intensity (I), a voltage (V) between terminals (2) of the cells (1), and a value of quantities of electricity (Q) applied to or withdrawn from the cells, during certain steps of the processing phase.

3. A method for preparing electric battery cells according to any one of claims 1 or 2, characterized in that the parameters include at least one of a temperature (°T), an internal pressure, a solid weight, and an electrolyte weight of the cells.

4. A method for preparing electric battery cells according to any one of claims 2 or 3, characterized in that the statistical model uses input variables comprising, in addition to direct measurements of said parameters, indirect measurements comprising extremal, mean, integrated or time-derived values ​​of said parameters during certain steps of the processing phase.

5. A method for preparing electric battery cells according to any one of claims 1 to 4, characterized in that the measurements are repeated periodically during at least part of the processing phase, and the statistical model performs a selection of the measurements.

6. Method for preparing electric battery cells according to claim 5, characterized in that the statistical model uses 10 to 40 input variables.

7. A method for preparing electric battery cells according to any one of claims 1 to 6, characterized in that it comprises, in a step of training the statistical model (9) made by means of the training cells, a selection of the input variables of the model by determining, among the measured parameters, those which are best correlated with the capacity values ​​of the training batteries.

8. A method for preparing electric battery cells according to any one of claims 1 to 7, characterized in that the statistical model includes a decision tree algorithm for exploiting the input variables.

9. A method for preparing electric battery cells according to any one of claims 1 to 7, characterized in that the statistical model includes a deep learning algorithm for exploiting the input variables.

10. A method for preparing electric battery cells according to any one of claims 1 to 7, characterized in that the statistical model includes a regression algorithm for exploiting the input variables.

Images