Design, analysis, and optimization of PMSM for ultra-deep water pile hammer application

Deepwater anchor piles are the foundation for ensuring the development of global offshore oil, gas, and wind energy resources, and an underwater pile hammer is the key to ensuring the installation of deepwater anchor piles. Currently, the development and production technology of pile hammers are basically mature, including the MHU series developed by MENCK Company in Germany, the S series manufactured by IHC Company in the Netherlands, and the MODEL series produced by HPSI Company in the United States.¹ However, most of them are small and medium-sized hammers that are used for driving piles above water or on land. Only MENCK and IHC are capable of manufacturing hydraulic piling hammers suitable for underwater applications.^2–4 The motors, which are the power source of the hammer driving system, are generally induction motors for both MENCK and IHC. For example, the motors used in underwater hydraulic piling hammers produced by MENCK are three-phase squirrel cage asynchronous motors produced by Foss Company in Germany.^5,6 The three-phase asynchronous motor has strong constant speed performance, but this performance also determines the poor speed regulation of the asynchronous motor. Due to the disadvantages of the asynchronous motor, such as its low power factor, large volume, and low efficiency, it is not ideal to apply this motor to driving hydraulic pumps of underwater pile hammers in practical. In addition, the three-phase asynchronous motor used for UDPH is filled with pure water for cooling. Although the cooling effect is good, due to the strong conductivity of water, the insulation requirements of the motor are very strict, which increases the manufacturing cost and the risk of an internal short circuit. As the permanent magnet synchronous motor (PMSM) has the advantages of high power factor, small size, lightweight, high efficiency, low noise, and good dynamic performance, etc., it can replace the traditional three-phase asynchronous motor in the UDPH.^7–9 Its characteristics are highly matched with the motor operation requirements under deepwater conditions, and it is the current research focus of high-performance underwater motors all over the world.¹⁰ Therefore, it is necessary to study the underwater PMSM applied for UDPH.

Recent advancements in offshore oil and gas exploration have heightened the application of deepwater pile hammer technology in marine engineering. Since the 1990s, Menck has introduced a deepwater pile hammer capable of operating at water depths of up to 300 m. In parallel, IHC has been innovating deepwater pile hammers that utilize water as a medium, replacing hydraulic oil, though there’s no documented usage yet. Research from MENCK highlights the significant role of their MHU-S and -T series of underwater hydraulic hammers in oil and gas exploration. These hammers, equipped with dual-action hydraulic technology and a modular design, achieve an energy transfer efficiency above 95%. This efficiency positions the MHU series as lightweight yet powerful tools for diverse pile configurations.¹¹ Moreover, Ref. 12 offers a comprehensive analysis of subsea structures in deepwater projects, discussing the functionality and depth limitations of hydraulic hammers. The study also notes that, when combined with MENCK's deepwater girdle-type power pack, these hammers can facilitate installations in depths reaching 3000 m. On another note, Pop et al. highlighted the surge in offshore pile driving activities due to the growing number of offshore wind farms in European waters. Such activities emit high-pressure acoustic waves in the surrounding environment, raising concerns about marine life.¹³ Addressing this, Jiang et al. explored the acoustic response of column vibration, focusing on the noise produced by pile-driving hammers. They proposed a double-shell pipe pile structure, showcasing its efficacy in noise reduction.¹⁴ 

At present, PMSM is widely used in industrial production, digital control centers, industrial machinery drives, and other fields. It is well known that electric vehicles have been studied since the 1980s, and the PMSM is currently used as the driving motor. Since the release of the Prius I in 1997, Toyota has newly developed the Prius IV multilayer magnetic barrier with built-in PMSM with an output power of 53 kW and a maximum efficiency of 97%. The 2016 BMW i3 high-speed PMSM developed by BMW has an output power of 126 kW and an efficiency of 94%.¹⁵ In 2017, Tesla developed the Model 3, whose driving motor is also a PMSM, with a maximum output power of 192 kW and a maximum torque of 430 Nm.¹⁶ In order to enhance the power density of high-power electric motorcycle drive motors, a novel PMSM with an optimized asymmetric eccentric air-gap structure is proposed.¹⁷ In rail transit, the Japan Railway Comprehensive Technology Research Institute produced the first RMT1 direct drive PMSM in the 1990s, whose weight was reduced by one-third and efficiency was enhanced by 2.7%¹⁸ compared with asynchronous motors of the same specification. The direct drive PMSM produced jointly by CRRC Zhuzhou and Siemens has been put into use in the subways in Shenyang, Changsha, and Shenzhen. In the field of the military with its strict requirements, PMSM is still applicable.

Deepwater PMSMs, known for their high efficiency, reliability, and compact size, are increasingly adopted in deep-sea endeavors like offshore oil and gas extraction, submarine operations, and subsea exploration. Table I gives the application of radial PMSM in submarine and destroyer. Gambhire et al. delved into underwater robotic systems, particularly highlighting the potential of sliding mode control in PMSM speed servo systems, offering fresh insights into deepwater motor control.¹⁹ Reference 20 discussed the role of PMSM in oil production for deep-well artificial lifts, with a focus on the temperature analysis of a submersible PMSM driving a progressing cavity pump (PCP). Tian et al. explored the integration of PMSM in submarines, emphasizing the dual-position feedback control and drive system during navigation.²¹ Furthermore, Ref. 22 detailed the use of a PMSM in designing a controlled power system for underwater transmission mechanisms, reflecting humanity's deepening engagement with marine exploration.

However, despite the widespread adoption of PMSM in sectors, such as automotive, rail transport, and submarines, there is a noticeable gap in research and practical applications concerning pile hammers, especially in the realm of UDPH.

Based on the above, a novel PMSM for UDPH with an underwater depth of 2500 m and an impact energy of 450 kJ is proposed in this paper. The method of combining field calculation and magnetic circuit calculation is adopted to complete the design of the motor. Then, electromagnetic field finite element analysis and temperature field analysis are conducted to verify the rationality of the new motor.^23,24 In addition, aiming at improving efficiency and reducing cogging, parameter optimization is carried out by a robust parameter optimization design method, and the optimal parameter scheme of the PMSM is obtained. Furthermore, in order to verify that the temperature of this motor can meet the requirements, loss analysis and temperature field coupling calculations of the optimized PMSM are conducted. The calculation results of the electromagnetic field and temperature field prove the rationality of the motor designed in this paper, and it meets the application requirements of UPDH. Therefore, this study can provide a new theoretical basis for improving the traditional UDPH and further assist in the development of deepwater oil and gas resources and offshore wind energy. Table II gives the comparison between traditional and new motors.

Currently, only three-phase asynchronous induction motors are utilized for UDPH, with PMSM yet to be implemented. There has not been any significant research on the use of submersible PMSM in UDPH. It is important to note that three-phase asynchronous induction motors and PMSM possess distinct rotor magnetic poles. The former uses induction coils for the rotor’s magnetic poles, while the latter employs permanent magnets. This distinction leads to fundamental differences in their electromagnetic and structural designs. PMSM offers clear benefits, including high magnetic energy, reduced size and weight, and superior efficiency and output torque.

In this study, PMSMs are used to power hydraulic pumps, which, in turn, drive the hammer. Figure 1 depicts the UDPH’s layout.

According to the requirements of water depth (2500 m), impact energy(450 kJ), and frequency of the UDPH, the rated parameters of the motor are calculated and determined, as given in Table III.^25,26

Compared to conventional ground motors, the PMSM examined in this research faces notable challenges due to its operation at a daunting depth of 2500 m underwater.

1. Sealing performance: Given the deepwater conditions and seawater’s conductivity, it is crucial to implement reliable sealing measures. This prevents seawater and micro-organisms from infiltrating the motor, which could lead to short circuits or even significant motor damage.

2. Pressure resistance: Seawater pressure escalates with increasing depth, rising by ∼1 MPa for every 100 m. At 2500 m, the motor encounters a pressure of around 25 MPa. Such immense pressure can compromise delicate structures, making it essential to factor in this pressure during motor design.

3. Specification requirements: The motor’s capacity and weight are vital considerations, especially concerning energy transmission during UDPH’s hammering process. Enhancing UDPH efficiency entails reducing the motor’s weight and volume without compromising its power.

4. Temperature management: PMSMs operating in deep-water settings grapple with limited heat dissipation compared to their counterparts in the air. Coupled with their higher power density, this results in more significant losses per unit volume. Consequently, various motor components undergo a pronounced temperature increase. Such elevated temperatures not only affect the motor’s performance and longevity but also threaten its insulation, risking potential motor damage.

5. Corrosion resistance: The electrochemical nature of seawater, combined with material erosion and marine microbial activity, makes metals susceptible to corrosion. Therefore, it is imperative that the motor’s casing be crafted from corrosion-resistant materials, given its direct exposure to seawater.

Compared to ground motors, the structure of deep-water PMSM may vary based on their application in different underwater situations, primarily to address the specific requirements and challenges posed by the marine environment. These include factors such as pressure and cooling demands. The deep-water permanent magnet motors studied in this article exhibit the following specific structures. Fig. 2 shows the layout of PMSM for UDPH.

1. Pressure compensator. In order to withstand the high pressure, a pressure compensator is installed at the end of the motor. It is a positive pressure type that maintains the internal pressure of the motor slightly higher than the external seawater pressure to prevent the strong corrosive seawater from pouring into the motor.

2. Oil-filled systems. In contrast to typical water-filled motors, the pressure protector’s interior and the motor assembly are entirely filled with insulating motor oil. Due to water’s poor insulating properties, higher conductivity, and the potential for increased instability under high pressure, oil is filled instead of water to offer better protection for motor insulation and also lubricate rotating components.

3. Oil circulation systems. In addition to featuring a pressure protector, at the other end of the motor, there is an impeller installed. The impeller, protector, and oil passages together form an oil circulation system. During the circulation process, the motor oil dissipates heat while also lubricating bearings and other components. The hollow motor shaft and the gap between the stator and rotor are the main passageways for the motor oil.

4. Temperature rise estimation. Considering the harsh operating environment and limited heat dissipation conditions of deepwater motors, accurate temperature rise estimation is crucial for the construction of the PMSM studied in this study. In the calculation process, the flow of internal motor oil and external seawater should both be carefully taken into account during the calculation process. As described in more detail in the following parts, this study will employ a multi-physics two-way coupled method.

According to the requirements of water depth (2500 m), the pressure compensator needs to provide a pressure of no less than 25 MPa and the volume is 20 l. As the vibration-absorption structures are part of UDPH, a detailed design will not be presented in this paper.

First, the basic size parameters of the motor according to the empirical formula, including the pole-slot fit, the inner diameter and length of the stator core, the structure size of the stator and the rotor, etc., are calculated according to the empirical formula.^27,28

Then, according to the basic structural parameters calculated above, further calculations are conducted to get the magnetic circuit parameters, including stator tooth width, stator tooth calculated length, stator yoke calculated length and height, rotor yoke calculated length and height, etc. After obtaining the magnetic circuit parameters of the motor, the no-load magnetic circuit of the motor can be calculated.²⁹ The process is shown in Fig. 3.

After the previous calculation and the continuous adjustment of parameters, the basic design scheme of the PMSM is determined as given in Table IV.

The finite element simulation model is established to verify the rationality of the electromagnetic design of the motor. The basic guiding ideology of the finite element method is to combine variational principle and subdivision interpolation to solve partial differential equations. The specific implementation process is to divide the region to be solved into finite elements, discretize the continuous problem, and establish and solve the differential equations with each element node as the unknown quantity.³⁰ The finite element analysis model and meshing model of the motor are established by ANSYS Maxwell, as shown in Figs. 4 and 5.

In Maxwell 2D, the no-load state boundary conditions and current excitation are set, and the no-load state finite element analysis is carried out to obtain the no-load back EMF, magnetic density cloud map, and radial airgap magnetic density.

It is shown in Fig. 6 that the no-load back EMF waveform of the PMSM is an ideal sinusoidal flat-topped wave,³¹ and the voltage peak value is lower than the rated voltage of the motor by 3000 V, indicating that the motor can start normally. From the Fourier spectrum analysis diagram of the no-load back EMF, it can be seen that there are fewer harmonics, which meet the design requirements.³² 

Figure 7 shows that the airgap flux density of the PMSM designed is an ideal flat-top wave with small fluctuation. It can be seen from the Fourier spectrum decomposition that the harmonic content is less, indicating less fluctuation. The maximum airgap magnetic density is 1.1 T, which meets the design requirements.

The no-load magnetic flux density and magnetic density vector are shown in Fig. 8, which indicate that the maximum magnetic flux density in the stator teeth is 1.65 T and the maximum magnetic flux density in the stator yoke is 1.43 T, both within a reasonable range.

Through the finite element analysis of the load state of the PMSM, the changes in the output torque, current, magnetic flux density, and magnetic flux density vector under the loading state can be observed to further verify the rationality of the design.³³ 

From the loading torque curve shown in Fig. 9(a) and the loading three-phase current curve in Fig. 9(b), it can be seen that the operation of the motor is gradually stable and the fluctuation value is small after 60 ms, which meets the design requirements.

Figures 10(a) and 10(b), respectively, show the magnetic flux density and the magnetic flux density vector at 0.02 s under load. As shown in Fig. 10, when the motor is running under load, the overall distribution of the magnetic flux of the motor is relatively uniform, and there is no magnetic saturation, indicating that there is less magnetic flux leakage. Thus, the motor can maintain good operating performance under load.

The Taguchi robust parameter optimization method can obtain the optimal parameter combination by establishing a reasonable test scheme and reducing test times. The sensitivity of product manufacturing processes to various types of noise factors is reduced through optimal design parameter combinations, and as a result, the quality stability of the product is improved. The idea of this method is to improve product quality through design rather than testing, which can reduce manufacturing costs and achieve ideal products. With considering noise factors, improving efficiency and reducing cogging torque of this motor are set as the optimization objections,³⁴ and ultimately operating performance and stability of this motor are improved.

Controllable factors affecting motor efficiency and cogging torque are selected as follows: stator notch width B_s0, stator slot depth H_s2, polar arc coefficient α_p, airgap length δ, and permanent magnet thickness h_m, as shown in Fig. 11. Three levels of each controllable factor are taken: initial value, low-level value, and high-level value. The processing tolerance of stator inner diameter ΔD_il, airgap Δδ, and permanent magnet thickness Δh_m that are prone to errors in the motor manufacturing process are selected as noise factors, and two levels of each noise factor are selected.

Each controllable factor takes three levels, namely, level 1 (initial value), level 2 (low-level value), and level 3 (high-level value). Then, the controllable factor level table is established as given in Table V.

A robust parameter optimization orthogonal table with different levels of controllable factors and noise factors was established, and all calculations were completed according to this table. Table VI presents the efficiency optimization orthogonal table and calculation results, and Table VII demonstrates the cogging torque optimization orthogonal table and calculation results, respectively.

Taguchi regards the signal-to-noise ratio (S/N) as the evaluation index of robust parameter design, and the higher the S/N, the higher the stability of the product. The optimization objective of this time is to improve the motor efficiency and reduce the cogging torque. Obviously, efficiency is usually expected to be higher, while the cogging torque is desired to be lower. Therefore, the optimization objective of efficiency is the larger-the-better characteristic, and conversely, it is the smaller-the-better characteristic for the optimization objective of cogging torque.

The average value of the total efficiency S/N and total cogging torque S/N can be obtained by the S/N results given in Tables VI and VII. Table VIII illustrates the average value of total S/N for both efficiency and cogging torque.

The mean value of efficiency S/N and efficiency for each controllable factor at different levels are given in Tables IX and X, respectively. The main effect diagrams of efficiency S/N and efficiency are demonstrated in Figs. 12 and 13.

The mean value of cogging torque S/N and cogging torque for each controllable factor at different levels are given in Tables XI and XII, respectively. The main effect diagrams of the cogging torque S/N and cogging torque are demonstrated in Figs. 14 and 15.

From Figs. 12 and 13, it can be seen that the controllable factor combinations meeting with the maximum efficiency S/N and efficiency are B_s0(1.5), H_s2(17), α_p(0.7), δ(1.1), and h_m(5). Figures 14 and 15 show that the controllable factor combinations meeting with the maximum cogging torque S/N and minimum cogging torque are B_s0(1.5), H_s2(17), α_p(0.8), δ(1), and h_m(5).

Variance analysis can quantitatively analyze the relative importance of each controllable factor to the S/N, thereby determining the optimal combination. Table XIII gives the relative importance that each controllable factor to both efficiency S/N and cogging torque S/N.

From Table XIII, it can be seen that the relative importance of H_s2, δ, and h_m to the efficiency S/N is larger than that to cogging torque S/N, while the relative importance of B_s0, α_p to cogging torque S/N is larger than that to efficiency S/N. The optimal Pareto factor combination for this design is determined as B_s0(1.5), H_s2(17), α_p(0.8), δ(1.1), and h_m(5), as given in Table XIV. Efficiency and cogging torque are recalculated according to the optimized parameter, and the comparison of results before and after optimization is given in Table XV.

From Table XV, it can be seen that after this robust parameter optimization, the motor efficiency has been increased by 1.1%, the cogging torque has been reduced by 99.9%, and the motor performance has been greatly improved.

The winding loss, which is also called copper loss, is caused by winding resistance under the action of the current; thus, it is related to the resistance and the current. Different from traditional induction motors, the permanent magnet synchronous motor used for the ultra-deepwater pile hammer only has stator winding loss but no rotor winding loss. Winding loss accounts for a large proportion of losses and is also one of the main sources of motor heating and temperature rise.

The wingding loss can be calculated by Maxwell, and the loss curve is shown in Fig. 14.

It can be seen from Fig. 16 that the winding loss of the motor tends to be stable after 50 ms, and the loss value is about 3.75 kW.

As is known, the resistivity increases with temperature, and consequently, the winding loss will increase with temperature. The influence of temperature on winding loss is considered by the two-way coupled method introduced in detail in our early study,²⁰ and the calculation process is iterative.

The core loss curve, hysteresis loss curve, and eddy current loss curve of the motor can be obtained through Maxwell simulation, as shown in Fig. 17. As the additional loss accounts for a small portion of core loss, it can usually be ignored.

Figure 17 indicates that the core loss of the motor tends to be stable after 50 ms and is about 3.1 kW under stable operation, of which the hysteresis loss is about 2.5 kW, accounting for about 80%, and the eddy current loss is about 560 W, accounting for about 20% of the core loss.

Mechanical loss is usually caused by the friction inside the bearing and the friction between rotor rotation and oil. Referring to the experience, 2% of the rated power is selected as the loss value and without a detailed analysis of it. There is eddy current loss generated inside the permanent magnet due to harmonic magnetic fields in the airgap. However, it is very small compared to copper loss and stator core loss so that it has almost no impact on motor temperature.

In order to verify that the temperature rise and output performance of the optimized motor meet the design requirements in a deep water environment of 2500 m, temperature field and electromagnetic field calculations in deep water environments are conducted, respectively, in this section.

Deepwater PMSM, unlike those operating in air, is filled with oil to balance the external pressure of seawater. This method, however, leads to additional fluid friction losses when the rotor rotates. Moreover, the need for deepwater PMSM to be compact and have a high power density results in increased losses per unit volume. Consequently, these significant losses cause a rise in the temperature of the motor. High temperature rise not only affects the performance and service life of the motor but also damages the insulation layer of the motor. Therefore, the accurate temperature rise calculation for these deepwater PMSMs is crucial. Due to the oil inside and seawater outside, the flow of the fluid field has a significant impact on heat dissipation of the motor. It is necessary to analyze the internal fluid field and calculate losses of the motor to employ a two-way coupled method for the accurate temperature rise calculation. Such comprehensive analysis ensures the motor's safe and prolonged operation, while also adhering to temperature rise constraints.

Considering the skin effect of the motor, the influence of temperature on the winding loss,^35–37 and the effect of the fundamental and harmonics on the iron loss, the motor loss is calculated.^38–41 Through the calculated winding loss and iron loss, the heat generation rate of the motor heat source is further calculated, as given in Table XVI. The thermal conductivity of each thermally conductive component is obtained by consulting relevant materials, as given in Table XVII.^15,42,43

We establish the finite element analysis model of the motor and draw the mesh with ICEM, as shown in Fig. 18.^44–46 

The 3D model for the temperature field calculation into ANSYS Fluent was imported, and then, we set the boundary conditions.⁴⁷ Assuming that the initial temperature of the motor and seawater is 3 °C, the seawater around the casing, the motor airgap oil, and the bearing lubricating oil are fluid. The heat generation rate and thermal conductivity were loaded to the corresponding components according to Tables XVI and XVII, respectively.⁴⁸ 

Lubricating oil is mainly in the gap and the shaft internal channel, and the oil circulation route is shown in Fig. 19. The flow rate affects the motor temperature; meanwhile, temperature plays a part in changing the viscosity of the fluid. As the viscosity only affects the mechanical loss that accounts for a small proportion of whole loss, the effect of temperature on fluid is not considered in this paper.

According to Sec. IV of the robust parameter optimization, those factors that affect efficiency and cogging torque are optimized. The new values are given in Table XIV. In order to confirm that the output performance of the motor meets the design requirements under the optimized parameter combination, it is necessary to carry out electromagnetic field finite element calculation. Based on the electromagnetic field analysis of non-optimization, we conduct another calculation for the optimized electromagnetic design. The relevant parameter settings in Ansys Maxwell refer to Tables IV and XIV.

The PMSM designed in this paper is internally oil filled, which is different from the existing way of internal water filling of the asynchronous motor for the deep-water pile hammer. The heat generated in the motor is mainly dissipated by the seawater outside the motor and the lubricating oil inside the motor, both of which are fluids, and their velocity will affect the heat dissipation of the motor. Underwater 2500 m, the seawater is relatively stable. According to the situation that is not conducive to heat dissipation, the seawater flow rate is set to 0.3m/s. In order to study the influence of the lubricating oil flow rate on heat dissipation of the motor, the flow rate was set to 0, 0.1, 0.3, and 0.5 m/s, respectively. The temperature contours of the motor as a whole and each component under different flow rates are obtained through simulation.

Through all the simulation calculations of the PMSM temperature field with different lubricating oil flow rates, all the results of different flow rates are summarized in Table XVIII for comparison.

As shown in Figs. 20 and 21 and Table XVIII, the highest temperature part of the motor is the stator winding. When the lubricating oil flow rate is 0 m/s, the winding temperature is 67.01 °C, which is 64 °C higher than the temperature of the seawater around the motor; when the lubricating oil has a flow rate, the temperature is about 62 °C, which decreases by about 7%. The reason for this is that the winding loss of the motor is the largest, and so, the heat generated by the winding is the highest; when the lubricating oil flows, it takes away more heat, and the temperature of winding is lowered; in addition, due to the thermal conductivity of the insulating material around the copper wire is low, heat dissipation of the winding is slow.

Figure 22 and Table XVIII show that the temperature of the magnet is much lower than that of the winding and stator core. When the lubricating oil is still, the temperature of the magnet is 22.21 °C, which is the highest than that of the lubricating oil at different flow rates. When the lubricating oil has a flow rate, the temperature of the magnet decreases significantly; it is 4.56 °C with oil flow rate 0.5 m/s, which is about 80% lower than that of still. This is because the lubricating oil in the airgap is direct contact with the magnet, which can take away more heat, and when the lubricating oil has a flow rate, the heat taken away is greatly increased. It has a similar trend for rotor core temperature as shown in Fig. 23.

Figure 24 and Table XVIII indicate that the stator core has the highest temperature except for the windings. When the lubricating oil flow rate is 0 m/s, its maximum temperature is 43.07 °C, which is about 24 °C lower than that of winding. That is because the core loss of the motor is smaller than the copper loss when the motor is running under the rated state. In addition, the stator core is next to the seawater through the housing that has high thermal conductivity and rapid heat dissipation ability.

When the lubricating oil inside the motor has a flow rate, the motor temperature decreases significantly compared with that with oil still. However, when the flow rate of lubricating oil continues to increase, the temperature of each part does not decrease significantly. Therefore, when designing the motor cooling system, it is only necessary to give a certain flow rate to the lubricating oil, and the requirements for its flow speed are not high.

In addition to the temperature field analysis of the underwater 2500 m, the electromagnetic field analysis of the optimized motor is also calculated.

Figures 25(a) and 25(b), respectively, show the magnetic flux density and the magnetic flux density vector after optimization under load. As shown in Fig. 25, when the motor is running under load, the overall distribution of magnetic flux of the motor is relatively uniform, and there is no magnetic saturation, indicating that there is less magnetic flux leakage. Therefore, the motor after optimization has reasonable electromagnetic distribution.

From the loading torque curve shown in Fig. 26(a) and the loading three-phase current curve in Fig. 26(b), it can be seen that the stable output power is about 1.375 K N/m and the motor is gradually stable after 50 ms, which meets the design requirements.

Through the temperature field calculation, it can be concluded that when the motor is running under rated conditions, the operating temperature distribution of the optimized motor meets the design requirements with different oil flow rates, and the magnet will not be demagnetized at high temperature. Meanwhile, the electromagnetic analysis indicates that the electromagnetic field distribution and output power are all rational and meet the design requirements. Therefore, the PMSM designed in this paper is reasonable.

Figure 27 presents the prototype of the UDPH, which has undergone validation through both land-based simulation tests and sea trials.⁴⁹ As illustrated in Fig. 28, the driving system of this pile hammer employs a traditional three-phase asynchronous motor. Unlike PMSM, the rotor of this motor is composed of excitation windings, and its interior is filled with pure water instead of insulating oil. The motor has an external diameter of ∼370 mm, and when filled with water, it weighs around 1.2 tons. Furthermore, the hydraulic system of this UDPH is powered by four three-phase asynchronous motors. In all tests, including the land-based experiments and sea trials, all systems demonstrated stable performance.

In order to test the output performance of the induction motor used by the UDPH, a motor is employed as loading as shown in Fig. 29. The results are basically consistent with those of the sea trial experiments.

To demonstrate the benefits of the PMSM for UDPH, a performance comparison between the traditional three-phase asynchronous motor used in the prototype and the PMSM introduced in this study is conducted. The results are detailed in Table XIX.

According to Table XIX, the PMSM is noticeably lighter than the asynchronous motor, with a weight reduction of 0.58 t per unit and a total decrease of 2.32 t across four units. When using the PMSM, this reduced weight results in less gravitational potential energy during the hammering process, which leads to energy conservation. When operating under stable conditions, the PMSM registers a temperature of ∼61.8 °C, in contrast to the 75.6 °C of the three-phase asynchronous motor. A significant factor to consider is that both the stator and rotor of the three-phase asynchronous induction motor are equipped with copper windings, leading to more copper loss. The PMSM, on the other hand, employs permanent magnets instead of rotor windings, effectively eliminating rotor copper loss. This design not only ensures a lower temperature rise but also enhances motor insulation. In addition, the PMSM’s capacity to adjust speed in real-time is superior to that of the three-phase asynchronous motor in terms of accuracy.

The comparison of outer diameter and axial length is given in Table XX. It is obvious that PMSM has a smaller volume compared to the traditional induction motor, which is about 90 mm smaller in outer diameter and about 250 mm shorter in axial length. This is of great help in reducing the weight of UDPH and improving energy efficiency.

In this paper, based on the overall design parameters of the ultra-deepwater pile hammer, the underwater 2500 m depth, and the 450 kJ impact energy requirements, the calculation of rated parameters and the electromagnetic design of the PMSM are completed. Then, the finite element analysis of PMSM is carried out by Maxwell. The no-load and load analysis results show that the back EMF, magnetic density, and output torque of the motor are all reasonable. Finally, the various losses of the motor are analyzed, the temperature field simulation model is established, and the temperature field analysis under different lubricating oil flow rates is completed. The results show that the temperature distribution of the motor is reasonable. To further verify the superiority of the PMSM, we compared it with the traditional three-phase asynchronous motor based on the sea trial results of the 450 kJ and 2500 m UDPH prototype. From the analysis, it is evident that the PMSM demonstrates significant advantages in terms of weight, volume, and temperature rise. Through the simulation analysis of the electromagnetic field and temperature field, combined with the comparison of sea trial results with the traditional three-phase asynchronous motor, we further confirmed the rationality and efficiency of the PMSM.

However, the research and application of PMSM for UDPH are still in the early stages, and there are several issues that require further investigation. The current cooling structure is relatively simple, leading to a high temperature rise in the windings, which, in turn, demands higher insulation requirements. In addition, in terms of optimization, multi-objective optimization design has not been pursued, and only manufacturing errors have been considered. In the future, the motor’s cooling structure should be thoughtfully designed to reduce the temperature rise, thereby enhancing the motor’s efficiency. Furthermore, continued research in optimization is essential. Emphasis should be placed on multi-objective optimization design and the application of global optimization algorithms to identify the best solutions.

The authors would like to thank the editor and the reviewers for their helpful comments and suggestions. This research was funded by the National Key R&D Program of China (Grant Nos. 2021YFB3401400 and 2021YFB3401402), the Fundamental Research Funds for the Central Universities—the Opening Fund of National Engineering Laboratory of Offshore Geophysical and Exploration Equipment, China University of Petroleum, Qingdao Grant No. 266580, China (Grant No. 20CX02303A), and the Ministry of Industry and Information Technology of the P R China (Grant No. CH02N20) under Grant No. (2020)313.

The authors declare no conflict of interest.

Liping Tan: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Writing – original draft (equal); Writing – review & editing (equal). Yucong Wang: Data curation (equal); Formal analysis (equal); Software (equal). Teng Wang: Formal analysis (equal); Validation (equal). Changjiang Li: Methodology (equal); Validation (equal). Fenghua Miao: Investigation (equal); Software (equal). Wensheng Xiao: Data curation (equal); Formal analysis (equal); Investigation (equal). Junguo Cui: Data curation (equal); Funding acquisition (equal); Project administration (equal). Hongyan Wang: Data curation (equal); Investigation (equal).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.