© Q. Sheng et al., Published by EDP Sciences 2025

1 Introduction

Against the backdrop of increasingly severe global ecological pollution and worsening energy crisis, the new energy vehicle industry is flourishing and has become an important force in promoting the transformation and upgrading of the automotive industry [1]. According to authoritative institutions, the world will achieve a cumulative sales volume of 18.236 million new energy vehicles in 2024, an increase of 27.7% compared to 2023. The booming development of new energy vehicles has driven a rapid increase in the installed capacity of power batteries. Lithium ion power batteries are the core components of new energy vehicles, accounting for 40% to 60% of the total vehicle cost [2–4]. Lithium ion batteries are the driving force of new energy vehicles. Its performance and manufacturing quality are directly related to the vehicle's range, safety, and market competitiveness [5–8]. However, in the production process of lithium-ion battery modules, especially in the module stacking and welding links, there are a series of urgent problems that need to be solved.

Currently, the main challenge facing battery module stacking and welding processes is the continuous accumulation of cell size errors. The continuous accumulation of errors beyond the module length range will lead to a decrease in the module length qualification rate, which in turn will affect the subsequent PACK process. The changes in module length and pressure values also directly affect the quality of module welds, causing problems such as weak welding and virtual welding, seriously reducing the service life and safety of battery modules. Kouhestani [9] proposed a deterministic data-driven prediction (DDP) method for multi-physics and multi-scale analysis to detect battery anomalies and predict battery failures. Rohkohl [10] proposed a method for multi-standard and real-time control of continuous battery production steps using deep learning, which can obtain the quality, minimum cost and impact of manufacturing activities. Boeselager [11,12] proposed a new electrode stacking process with a rotating handling device that can achieve continuous high-throughput material flow. At the same time, based on the rotating handling device, the alignment mechanism is modularly designed. By introducing the alignment mechanism in the assembly system, the deposition accuracy can be improved, and a novel stacking process is established. Asif [13] proposed a new framework called reconfigurable and responsive robotic manufacturing(R3M), which can autonomously adapt to fluctuating product variants and demands in the manufacturing environment. Lei [14] used numerical simulation technology to perform expansion force simulation, overcurrent temperature rise simulation, impact simulation and vibration fatigue simulation on the module at the single module level. This study completed the design of the module and electrical connector structure while ensuring the stability and reliability of the structure. Zhang [15] developed a power battery module automation equipment through structural design and control design, completed the entire power battery automation production task and improved the production efficiency of the battery module. Yang [16] developed a set of automation equipment that can assemble 21700 cylindrical power battery modules through hardware and software design and completed the assembly and testing of the entire equipment. However, they have not yet fully addressed the issues of module length instability and weld quality caused by the accumulation of cell size errors. In addition, existing control systems still have shortcomings in terms of intelligence, automation and real-time feedback adjustment capabilities.

In response to the above issues, this study proposes an innovative lithium-ion aluminum shell power battery module stacking intelligent control system. The system adopts a dual control strategy of length and pressure, which achieves intelligent control by accurately detecting and feedback the length and pressure values of the module during the pressurization process, in order to improve the first pass rate of module stacking and the stability of weld quality.

This study combines the two key parameters of module length and pressure for collaborative control, effectively avoiding the instability of module length and weld quality problems caused by the accumulation of cell size errors. This study establishes a closed-loop feedback control system to real-time detect module length and pressure values. And intelligently adjust them according to design requirements to ensure stable and reliable module quality. This study achieves efficient automation of module stacking and welding processes by integrating advanced control systems and actuators, which can reduce labor costs and improve production efficiency. The implementation of this study will provide a new solution for the production of lithium-ion battery modules. It helps to improve the overall performance and safety of the power supply system for new energy vehicles, promoting the sustainable and healthy development of the new energy vehicle industry.

2 Module stacking process analysis

The production process of lithium-ion battery module stacking is shown in Figure 1: The production line MES system places orders for production tasks. After the PLC control system receives the order, the production line begins to scan the incoming battery cells and perform OCV tests. The battery cells that pass the OCV test are plasma cleaned, and then the cleaned qualified battery cells, end insulation plates, and end plates are pre-stacked on the stacking table. After the pre-stacking is completed, they are placed in the welding fixture with the side plates for module pressurization and welding.

During the pressurization process before welding, the module length and pressure value play a significant role in affecting the quality of the module. If the module length value exceeds the design deviation, on the one hand, the distance between the battery cell poles will also deviate, resulting in the busbar welding process not being able to proceed smoothly; on the other hand, it will cause unstable welding quality. Moreover, length deviation will also cause the module to be unable to be packed into the box smoothly. If the pressure value of the module is lower than the minimum value, the height misalignment of the battery cells may easily occur during the transfer process of the robotic arm after the module welding is completed, and the flatness of the poles cannot be guaranteed to meet the requirements, which will cause cold welding in the subsequent busbar welding process, thus affecting product quality. If the pressure is too high, the battery cell will be deformed, affecting the subsequent charging and discharging performance. To solve this problem, our research proposes a dual control measure that can effectively avoid the impact of battery cell deviation on product quality.

Fig. 1 Module stacking process flow chart.

3 Methods

The intelligent control system mainly consists of three parts: detection system, control system and execution system. By establishing a closed-loop feedback control system, as shown in Figure 2, the pressure and length values of the module during the pressurization process are accurately detected and fed back. Through the operation of the control system and the execution of the execution system, this study ultimately ensured that the module length and welding pressure met the design requirements, improved the weld quality, and reduced the module scrapping and disassembly rate.

Fig. 2 Control system block diagram.

3.1 Detection system design

The purpose of the detection system is to complete dynamic detection of module length and pressure. The closed-loop servo control of the electric cylinder is responsible for detecting the module length. The module pressure detection requires dynamic detection of the pressure value during the module pressurization process. In this design of the detection system, we installed the pressure sensor at the active end of the welding workbench, that is, at the end of the electric cylinder guide rod. It can dynamically measure the pressure value applied to the module by the end of the electric cylinder, and can also avoid excessive module pressure or other safety accidents caused by inaccurate transmission of pressure values due to the tipping of the battery cell or other factors. Its structure is shown in Figure 3.

Fig. 3 Detection system structure diagram.

3.2 Selection of control system hardware

The control system is the core of the intelligent system. When pressurizing the module, the servo cylinder and pressure sensor transmit the dynamic information of the module length and pressure to the control system. The control system uses computational analysis to control the execution elements to ensure that the accuracy and consistency of the module length and pressure meet the standards, ultimately ensuring the quality of module welding. The control system is mainly composed of Beckhoff control system and Siemens PLC.

Beckhoff drive technology represents a leading and comprehensive drive system technology. Combined with the motion control solution provided by TwinCAT automation software, it is an ideal choice for realizing high-dynamic single-axis and multi-axis positioning tasks [17,18]. TwinCAT fully supports IEC61131-3 programming language, has PLC and NC functions, provides VB\C++ and other third-party interfaces, fully supports WINDOWS standard DDE\ADS\OPC and other communications, and is easy to embed and integrate with third-party software.

Siemens S7-1500 PLC uses a new backplane bus technology, high baud rate and high transmission protocol, which makes its signal processing speed faster [19]. At the same time, the S7-1500 CPU integrates 1-3 PROFINET interfaces, which can realize low-cost and fast configuration of field-level communication and corporate network communication [20]. In addition, the S7-1500 PLC is seamlessly integrated into the TIA Portal software, and the operation is simple and fast, whether it is hardware configuration, network connection and upper-level configuration, or software programming.

3.3 Design of execution system architecture

The execution system is shown in Figure 4. It is composed of a welding table base, a servo-electric cylinder, a guide rail, a centering fixture, a pressing arm, etc. The servo-electric cylinder extends a guide rod under the control of the system's control. The end of the guide rod is equipped with a pressure sensor and a movable end pressure block. The pressure block moves along the guide rail to press the module. During the module pressing process, the centering fixture and the pressing arm need to be working to ensure the centering of the battery cell during the pressing process.

Fig. 4 Execution system structure.

3.4 Software architecture design

During the production process, operators place orders to the MES system based on production tasks, and the PLC controls the production line to automatically complete the module stacking work according to the order type of the MES system. After the stacked modules are placed in the welding room, the PLC informs the Beckhoff system of the module information, and the Beckhoff system controls the movement of the servo-electric cylinder to pre-pressurize and re-pressurize the module. During the secondary pressurization process, the Beckhoff system cooperates with the S7-1500 PLC control system to achieve real-time feedback, calculation, and control of the module length and pressure value to ensure that the module length and pressure value meet the design requirements, thereby ensuring that the module welding quality meets the standards. During the entire pressurization process, the module data information will be displayed on the HMI interface and saved to the MES system to establish a ledger for the entire life cycle of the module. Figure 5 is a schematic diagram of the system architecture.

Fig. 5 System architecture.

3.5 Software programming

In order to make program debugging and transplantation more convenient, the program designed in this paper adopts a modular programming method, dividing the entire program system into MES order module, module stacking module, data processing module, status display module and PLC and Beckhoff control module. The program flow chart of the entire system is shown in Figure 6.

After the system starts working, the first step is to initialize the system and reset all data to zero. Next, the MES system, PLC system and Beckhoff system transmit order data. After the battery modules are stacked in the stacking position, they are transported to the welding station by the handling robot and the servo system controls the servo cylinder to complete the first and second pressurization. After the second pressurization, compare the current length and pressure value of the module with the set value. If the values are within the set range, the module welding work will continue. If it is not within the range, it will be fine tuned through a dual control strategy of length and pressure. After fine-tuning, if the value meets the requirements, welding will continue. If it does not meet the requirements, the module will be removed. The removed battery cells will be returned to the previous process. After passing the quality inspection, the battery cells will enter the three-dimensional warehouse again for production and processing.

The dual control strategy of length and pressure is the core part of this control system. The control module includes the control of module pressure and length. Through the control system, the module length and pressure meet the design requirements, ensuring that the process parameters such as module welding quality meet the standards. The flowchart of this module is shown in Figure 7.

During the system operation, the PLC transmits the module information to the Beckhoff system, which controls the servo cylinder to complete the pre-pressurization and secondary pressurization of the module. During the second pressurization, the control system obtains the module length and pressure stability value, and compares the measured length and pressure stability value with the design parameters. If they are within the range of process requirements, the module welding work will be carried out. If they are not within the range of process parameter requirements, they will be processed according to different situations.

Fig. 6 Program flow chart.

Fig. 7 Automatic adjustment module process.

3.5.1 Upper limit of length, lower limit of pressure

The module battery is an aluminium shell battery. During the pressurization process, the aluminium shell will be slightly deformed by the pressure, so the real-time pressure value will slowly decrease. If during the pressurization process, the module length has not reached the design range, and the pressure value has reached the maximum value alarm, the control system will control the servo-electric cylinder to delay for 1S, and then judge the pressure stability value. If the pressure value is within the range, the pressurization can continue. If it still exceeds the upper limit, the system will judge it as an unqualified module, the system will alarm, and the unqualified module will be manually removed.

3.5.2 Upper limit of length and upper limit of pressure value

If the module length is within the qualified range and the pressure stability value is within the lower limit of the set value during the pressurization process, the control system controls the servo cylinder to move forward at a speed of 0.3 mm/s until the module pressure is qualified and stops moving. If the module length reaches the lower limit of the design value and the pressure has not yet reached the qualified range, the system determines that the module is an unqualified module and displays an alarm signal on the HMI screen. At the same time, the three-color light of the control cabinet also issues an alarm, and the worker removes the unqualified module.

3.5.3 Lower limit of length and lower limit of pressure stability value

If the module length has reached the design lower limit, but the pressure stability value has not reached the design range and is at the design value lower limit, the module is judged as an unqualified module, the system alarms, and the workers remove the unqualified module.

The battery cells removed before welding need to be returned to the battery cell production section for retesting of parameters such as size, voltage and current. After passing the tests, they will be returned to the material bin for production again.

4 Experimental results analysis

The test conditions for the stacking length and pressure control system of the aluminium shell module designed in this design are: D173F120 lithium iron phosphate aluminium shell 9-core module, the module design length is 461±1mm, the lower limit of the welding pressure value is 650Kgf, and the upper limit is 1500Kgf. The test cycle is one week, two shifts per day, and the number of stacked modules ordered by MES, the number of qualified modules, the number of unqualified modules due to module length, and the number of unqualified or repaired modules due to weld quality are recorded separately.

After the modules are stacked, the transfer robot places the stacked modules on the welding workbench. The welding table fixture completes the adjustment of the module's pole flatness and left and right centring. The electric cylinder push rod is pushed out to drive the movable end to move, completing the primary and secondary pressurization of the module. Figure 8 shows the state of the welding robot welding after the module is pressurized twice on the welding table. The photo of the module after welding is shown in Figure 9 below.

The welding information of the module can be displayed on the HMI interface. The length, stable pressure and instantaneous pressure values of the module can be monitored in real time through the HMI interface.

The data obtained from the experiment are plotted into a bar graph to obtain a module production status bar graph as shown in Table 1. The bar graph records the number of modules ordered on each day of the week, the number of modules that passed the first welding, the number of modules with unqualified weld quality, the number of modules with unqualified length and the number of modules that were cleared due to unqualified length and pressure values before welding. Table 1a is the production status diagram before the control system is implemented, and Table 1b is the production status diagram after the control system is implemented.

As can be seen from Table 1, after the implementation of the intelligent control system, the first-time qualified rate of module welding has been significantly improved because the control system performs system fine-tuning control according to different situations during the module pressurization process. At the same time, because the modules whose length and pressure values exceed the limit and cannot be corrected are removed in time before welding, the number of modules that fail to meet the requirements due to length after welding is reduced to zero, reducing the workforce and material losses caused by the disassembly of unqualified modules after welding and saving production costs.

During the entire test process, the module is welded under pressure. After the welding is completed and the servo-electric cylinder retracts, the module length value will rebound slightly, but the rebound range error is within the acceptable range of the process. It will not affect the subsequent module PACK process, which meets the design requirements.

Fig. 8 Module welding.

Fig. 9 Module welding completed.

5 Conclusion

The designed intelligent dual-control system for stacking battery modules of the power supply system of new energy vehicles can effectively solve the problem of unstable module length caused by the accumulation of cell size deviation, which in turn causes unqualified module size or unstable weld quality, by coordinating the module length value and pressure value before welding. The control system can take corresponding corrective measures according to different situations where the module deviation exceeds the design range, which can effectively reduce the module failure rate. For modules that exceed the adjustable range, the welding workbench will be cleared before welding, effectively reducing the cost loss caused by disassembly of unqualified modules. In addition, the cleared modules will adopt a one-click data clearing mode to shorten equipment downtime, reduce equipment MTTR and improve equipment OEE, making the system have a high research and application value.

Funding

This research received no external funding.

Conflicts of interest

The authors have nothing to disclose.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author contribution statement

Conceptualization, A.S. and J.K.; Methodology, A.S. and Q.S.; Software, Q.S.; Validation, Q.S.; Formal Analysis, Q.S.; Investigation, A.S.; Resources, J.K.; Data Curation, H.W.; Writing Original Draft Preparation, Q.S.; Writing Review&Editing, H.W.; Visualization, A.S.; Supervision, J.K.; Project Administration, J.K.

References

All Tables

All Figures

	Fig. 1 Module stacking process flow chart.
In the text

	Fig. 2 Control system block diagram.
In the text

	Fig. 3 Detection system structure diagram.
In the text

	Fig. 4 Execution system structure.
In the text

	Fig. 5 System architecture.
In the text

	Fig. 6 Program flow chart.
In the text

	Fig. 7 Automatic adjustment module process.
In the text

	Fig. 8 Module welding.
In the text

	Fig. 9 Module welding completed.
In the text