3D simulation of orbit loss and heat load on limiters of ICRF-NBI synergy induced fast ions on EAST

Fast ions synergy induced by ion-cyclotron range of frequency (ICRF) and neutral beam injection (NBI) are of interest not only because of their advantage of heating the plasma and drive currents, but also because of their disadvantage of damaging plasma surface components and driving MHD instabilities. In this paper, we calculate the fast ion loss and the deposition distribution of the lost particles on the limiters in EAST under the synergistic effect of the ripple field and collisions with the full-orbit-following simulation program ISSDE for the first time. The previous models to study the NBI fast ion loss by the action of ICRF are relatively simple and consider fewer influencing factors. Most studies on fast ion loss have used toroidal uniform boundaries. In this work, we consider the distribution of ICRF-NBI synergy induced fast ions with different minority H concentrations. After setting the limiter boundary, we consider the prompt fast ion loss caused by the equilibrium field and the fast ion loss caused by the ripple field and collision. Under the action of minority-ion ion-cyclotron resonant heating, the NBI fast ion distribution function has spread in the high-energy part, especially for the minority H concentration of 1%, and the fast ions show each anisotropic distribution near the resonance band on the poloidal dimension. The synergistic loss caused by the ripple field and collision will first be greater than the loss caused by either factor, and then reach a final loss fraction of 3.8%. The heat load power density of the lost fast ions on different limiters is not uniform, as well as on each limiter, which is related to the distance from the limiter to the plasma, the relative position between the limiters and the parallel direction of most fast ions. Once the study of ICRF-NBI synergy induced fast ion loss caused by the action of ripple and collision has been done, we can do optimization in a targeted manner. Such as adding ferromagnetic inserts to reduce the ripple loss and optimizing the limiters’ position to reduce or control the generation of impurities.


Introduction
Energetic ion confinement is one of the most important topics in energetic ion research. Many energetic ions are generated in the tokamak device because of the fusion reaction, neutral beam injection (NBI), and radio frequency (RF) wave heating. To reach the ignition temperature, the auxiliary heating, such as high-energy NBI heating or resonant heating using RF waves, must supplement the intrinsic ohmic heating [1]. Ioncyclotron resonant heating (ICRH) has a variety of heating schemes [1][2][3], and it has the advantage of localized deposition of energy and the ability to heat the core plasma [4,5]. Different from the RF wave heating, NBI heating is independent on any resonance or coupling conditions [6], and its heating physical mechanism is clear and has a high plasma heating efficiency, and it can preferentially heat the electrons or ions in the background plasma as needed [7]. Under wave-particle resonance conditions, the RF wave energy leads to the interaction of the electric wave field with the NBI fast ions, which will further change the distribution function of the fast ions, an effect often referred to as the 'synergistic effect' [8].
Both NBI and ion-cyclotron range of frequency (ICRF) auxiliary heating systems are available in EAST. The energetic particles generated by ICRH and NBI have the advantage of heating the plasma and driving the current, but at the same time the lost energetic particles will damage the plasma surface components, the first wall (FW) and the limiters. The impurity particles generated by sputtering during ablation process limit the plasma current, plasma density and wave injection power to a narrow range, thus reducing the device confinement performance or even lead to the plasma disruption [9,10]. Therefore, studying fast ion loss and heat load deposition is important for fast ion confinement studies. Loss channels for fast ions include drift-induced loss, ripple-induced loss, collision-induced loss, and ICRH-induced loss [4]. In the experiment, fast ion loss diagnostics (FILDs) combined with other diagnostics can yield mechanisms of fast ion transport and loss. Although the FILDs cannot identify which ion type is present, they can provide information on fast ion energy and pitch angle, thus providing the mechanism of the fast ion energy source. Since FILDs detect the fast ion loss in localized regions, the measurement information used to demonstrate the underlying physics is insufficient, and thus a reliable theoretical simulation model is needed to describe the energetic ions and the corresponding effects [11]. The simulation of the realistic FILDs signal, the specific geometry of the machine, and the background magnetic field can provide us with a useful tool to interpret the measured data and optimize the detector position, and thus gain a better understanding of the fast ion generation mechanisms and transport processes [12].
A simplified model of fast ions produced by minority ion heating in EAST has been investigated, and a simplified resonance band was used in this study, which shows that the fast ion loss occurs mainly at the ends of the resonance band, with less in the intermediate region. The fast ions are lost uniformly in the toroidal dimension and increase with increasing ICRF energy [13]. However, the synergistic production of fast ions by ICRF and NBI in EAST has been less studied, and the changes of NBI fast ion orbital types during this process have not been studied. The synergistic loss of these fast ions of the ripple field and collision is also less studied. Not only that, most of the previous studies on the fast ion loss in EAST have used toroidal uniform boundaries, while the practical situation in the mid-plane needs to consider the effect of the limiters, i.e. the 3D asymmetric wall needs to be considered. In this paper, we will use the TRANSP which integrated the calculation of NUBEAM and TORIC modules, to get the distribution of ICRF-NBI synergy induced fast ions, and use the full-orbit-following Monte-Carlo simulation program ISSDE to track the fast ion trajectory. We further study the change of fast ion orbital types, stochastic loss of fast ions and the heat load power density of fast ions on the limiters. This work considers not only the effect of prompt orbit loss but also the effect of fast ion loss under the ripple field and collisional effects.
This paper is structured as follows. In section 2, we present the initial equilibrium, the initial distribution of ICRF-NBI synergy induced fast ions, the treatment of ripple field and collision, the distribution of the limiter, and the full orbital simulation program ISSDE; in section 3, we present the variation of fast ion orbital type under the synergistic effect of ICRF and NBI, the fast ion loss and power density deposition distribution on the limiter by the action of ripple field and collision; in section 4, we make some summary. Figure 1 shows the flowchart of our simulation. We use EFIT to reconstruct 2D tokamak equilibrium at the moment of t 1 = 0.03 s in EAST and use the TRANSP which integrated calculation of NUBEAM and TORIC, to get the slowing down distribution of ICRF-NBI synergy induced fast ions at the moment t 1 = 0.03 s, and set up the ripple field and the background plasma temperature and density profiles. The equilibrium profile, the ICRF-NBI synergy induced fast ion distribution, and the ripple field and plasma density and temperature profiles are taken as input to the ISSDE program, from which the fast ion trajectories are calculated. As the limiter boundary is set, the particle distributions of fast ions under the equilibrium field, ripple field, and collision are counted, as well as the fast ion loss fraction and heat deposition distribution.

ICRF-NBI synergy induced fast ion distribution
In our simulations, we use EFIT to obtain the equilibrium configuration of EAST at the initial moment, and in cylindrical coordinates, which would also be used in TRANSP to get the initial distribution of the fast ions. The relation between the expression of the EAST magnetic field strength B and the poloidal magnetic flux function ψ is [14] where g(ψ) is a function related to the poloidal current density, for a given ψ and g(ψ), we can calculate the values of each  According to EFIT, we get the EAST discharge #101 735 poloidal magnetic flux function with the toroidal and poloidal magnetic field distribution as shown in figure 2. The corresponding magnetic field configuration is close to the doublenull equilibria. According to the equation (1), we can calculate the corresponding magnetic field intensity distribution as shown in figures 2(b) and (c). The toroidal magnetic field B ϕ is decreasing along the major radius, i.e. there is an inward gradient of the toroidal magnetic field toward the major radius. Within the last closed flux surface (LCFS), the poloidal magnetic field is rotating counterclockwise. Because of the gradient and curvature of the magnetic field, the fast ions experience gradient drift and curvature drift, thus producing motion across the magnetic lines.
The neutral beam is injected into the plasma and changed to fast ions by collisional ionization and charge exchange with background electrons and ions. The NBI system in EAST uses two deuterium beamlines with a power of 2-4 MW and beam energy of 50-80 keV [15]. At the current stage, EAST is equipped with a 12 MW/25-70 MHz ICRF system [16,17]. And EAST obtained the first results from H-mode plasmas generated by ICRF heating in 2012 [18]. The NUBEAM is a Monte-Carlo package developed by PPPL for tokamak fast ion deposition evolution, slowing down, and thermalization [15]. NUBEAM can calculate both the NBI fast ions and fast ions produced by fusion reactions [19,23]. The full-wave simulation code TORIC [20] solves the linear sixth-order reduced wave equation for ICRF in the toroidal geometry with a Fourier representation for the poloidal dimension and a finite element method on the flux dimension [21]. Using TRANSP which integrated calculation of NUBEAM and TORIC modules, we can get the slowing down distribution of ICRF-NBI synergy induced fast ions, which is shown in figures 3 and 4. The heating effect of ICRF on NBI fast ions is different for different H minority ion concentrations. As shown in figure 3, the density distributions of NBI fast ions are calculated in the projection of the poloidal cross-section without ICRF heating and as the H concentrations of 1% and 5%, respectively, and the red line indicates the LCFS. When there is ICRF heating, the density distribution of fast ions in the poloidal crosssection is anisotropic, and the fast ions are denser near the resonance band. It shows the distribution of ICRF-NBI synergy induced fast ions in the pitch-angle and energy space as in figure 4. The particle fraction of NBI fast ions in the high energy band becomes larger by the action of ICRF, and the fraction of fast ions in the high energy band is higher when the minority H concentration is 1% than the case when the minority H concentration is 5%. This is because the heating mechanism of ICRF contains both fundamental frequency heating of H and high harmonic heating of NBI D ions. The high harmonic heating of NBI D ions by ICRF dominates when the minority H concentration is 1%, while when the minority H concentration is 5%, more of the ICRF injected power will be absorbed by H ions through fundamental frequency heating, so that the fraction of high energy ions corresponding to   NBI D ion is less than the case of the minority H concentration of 1%. More details can be found in the [22]. Change of the particle fraction in the high energy band and the change of the pitch-angle distribution will change the fraction of different particle orbit types and thus affect the fraction of fast ion loss. It worths noticing that ,we only considered the distribution of fast ions synergy induced by ICRF and NBI in our calculations. In fact, ICRF pure heating of bulk D ions is also important. Calculations in [22] show that 25% of the ICRF injection power will be absorbed by bulk D ions and 10% by NBI D ions, so bulk D ions also contribute to the fast ion loss. We will present the specific effects in out future work.

Treatment of ripple fields, collisions and boundaries
After obtaining the tokamak equilibrium magnetic field and the position, and velocity distribution of the fast ions at the initial moment, we can calculate the evolution of each fast ion by ISSDE. Under the influence of ripple fields and collisions, the fast ions experience ripple loss caused by the ripple field, pitch-angle scattering loss, and random diffusion loss caused by collisions [24][25][26]. Assuming that the ripple field generated by a finite toroidal field coilB φ ∝ cos(Nφ), where φ = φ p − φ 0 , φ p is the toroidal angle at any point in space, φ 0 is the toroidal phase of the first toroidal field coil, and φ is the toroidal angle from any point in space to the first toroidal field coil. AndB φ is the toroidal component of ripple perturbation field. According to Yushmanov's approach, we obtain [27,28] where δ(R, Z) is the ripple amplitude scales and B 0 , R 0 is a constant, according to The R-component and Z-component ofB should satisfy the following formB Bring the equations (4) and (5) into we can derive For EAST, N = 16, φ 0 = 11.25 • , δ(R, Z) has an analytic expression as [29][30][31] where δ 0 is the minimum of the ripple field and R 0 is the radius corresponding to the location where the minimum occurs at Z = 0, b r is the ellipticity and w r is the ripple scalar length. According to the equation (9), we obtain the profile of the ripple and thus the distribution of perturbation field and total field on the LCFS and the midplane, as shown in figures 5-7. The ripple amplitude in the mid-plane of LCFS is about 0.7%, and the closer to the LCFS and the FW, the greater the ripple amplitude. For the equilibrium field B 0 , it is toroidally symmetric independent of the toroidal angle φ as shown in equation (1). While for the ripple field as shown in equations (2), (4) and (5), it is a function of the toroidal angle, so the presence of the ripple field breaks toroidal symmetry of the total magnetic field. The toroidal component of ripple perturbation field δB φ is distributed asymmetrically but periodically in the toroidal dimension, as shown in figure 6(a), and the closer to the LCFS and the FW, the larger the corresponding ripple field is, as shown in figure 6(b). Since 2φ 0 = 2π/N, when φ p = 2nφ 0 , the δB ϕ is at the trough position, where the corresponding total magnetic field B φ is minimum, and when φ p = (2n + 1)φ 0 , the δB φ is at the peak position, where the corresponding total magnetic field B φ is maximum. The black line indicates the position of the windows and also the position of the trough of δB φ . The results are qualitatively consistent with those in the literature [28]. Compared to the ripple amplitude on the high field side, the ripple on the low field side is an order of magnitude higher, so the superposition of the ripple field has a greater effect on the low field side, as shown in figure 7(b).
Since the collision parameters are a function of the background plasma density and temperature, it is necessary to get the background density and temperature values at the local position of the particles and then bring them in to solve for the corresponding collision coefficients. The experiment data and fit data of density and temperature distribution profiles of EAST discharge #101 735 are shown in figure 8. We can then get the background density and temperature values of the local position of the fast ion by using the interpolation method. In our simulation, fast ions with energy less than 3/2 times 2.5 keV will be defined as thermal particles.
The boundary used in most fast ion loss simulations in EAST is a toroidal uniform FW boundary, as shown in figure 2(a), while in practice, the FW in the mid-plane is often composed of limiters to protect the antennas and the windows. To simulate the fast ion loss in a more practical case, we consider the 3D boundary containing the limiters. The limiters in EAST are composed of four low-hybrid (LH) limiters (in B and E ports), four ion cyclotron (IC) limiters (in I and N ports,) and a main limiter (between G and H ports). Their coordinates are shown in table 2. From the table, we can see that the main   The experiment data and fit data of density and temperature distribution profiles of EAST discharge #101 735. The data of the experiment is from [22]. Te data is from TS (Thomson scattering) diagnostic. T i data is from XCS (X-ray imaging crystal spectrometer). ne is from reflectometry, then the core profile is reshaped to match the line average density from the 11-channel POINT (POLarimeter-INTerferometer) diagnostic.  limiter is closer to the major plasma, followed by the IC limiters and the LH E limiters, while the LH B limiters are the farthest from the major plasma. Based on the actual engineering data, we get the limiters in the poloidal cross-section, the top view, and the three-dimensional distribution, as shown in figures 9-11.

ISSDE numerical discretization method
ISSDE is an implicit midpoint collision code based on the theory of stochastic equivalence of the Stratonovich stochastic differential equation (SDE) and the Fokker-Planck (FP) equation. First, the FP equation under the Rosenbluth Mac-Donald Juddy potential representation is used to get the Stratonovich SDE of the cases as the background plasma is Lorentzian and Maxwellian plasma, or the background particles with arbitrary distribution according to the stochastic equivalence between FP equation and Stratonovich SDE. Second, we discretize the Stratonovich SDE according to the definition of the Stratonovich integral and the requirements of the Euler-Maruyama method, then we get the implicit midpoint format algorithm. Third, the discretized Stratonovich SDE is processed by the splitting method, and the collision part is solved by the Newtonian and quasi-Newtonian methods. Finally, the kinetic trajectory of each particle is obtained by solving the Stratonovich SDE, and the evolution of the ensemble-averaged quantity of the particle system is obtained by the statistical method [32]. The ISSDE uses an implicit midpoint format and a splitting method to ensure that the error in the deterministic part of the equation do not accumulate. Compared with the stochastic Runge-Kutta method, the numerical method got by combining the implicit midpoint and the splitting method has higher numerical stability, and it shows the results in figure 5 in [32]. For the correctness of the program, the particle orbits calculated for the ISSDE full orbit without collisions and ripple fields are consistent with those got by the guiding-centre orbit algorithm provided by the PTC code [33], as shown in figure 1 in [34]. We have given the  benchmark of the Coulomb collision part with the analysis solution in [32]. ISSDE has calculated the prompt loss and slowing down process of NBI-generated fast ions because of Coulomb collisions in the equilibrium configuration of EAST [34]. More details can be found in [32,34].

Fast ion loss and heat load power density distribution
In our simulations, we consider the loss fraction of ICRF-NBI synergy induced fast ions at different minority ion concentrations under the equilibrium field, ripple field, and collisional effects. The physical problems we investigate include (a) the effect of the presence or absence of ICRF on the fast ion loss without collision and ripple field, (b) the effect of ripple field on the fast ion loss without collision and ICRF, (c) the synergy effect of ripple field and ICRF without collision, (d) the synergy effect of collision and ICRF without ripple field, and (e) the synergy effect of ripple field, collision, and ICRF. Besides, we also consider the effect of the limiter, the loss distribution of NBI fast ions on the limiter.

The NBI fast ion orbital type change under the ICRF second harmonic
First, we consider the orbit types variation of the ICRF-NBI synergy induced fast ions without collision and ripple fields. We classify the fast ion orbit into six types: 1. counter passing-confined (P − − C), 2. counter passing-lost (P − − L), 3. trapped-confined (T − C), 4. trapped-lost (T − L), 5. copassing-confined (P + − C), and 6. co-passing-lost (P + − L) types. Since these particle energies are different, they cannot be drawn on one orbit type map. We use the full-orbitfollowing program ISSDE to count the particle types. Let the total time be 10 ms, and the loss boundary is LCFS, and it shows the statistic results in figure 12. Without ICRF second harmonic, co-passing-confined particles dominate in NBI fast ions, followed by trapped-confined particles and counterpassing-confined particles. The fraction of loss particles is small and dominated by trapped-lost particles. Under the ICRF second harmonic, the number of passing NBI fast ions decreases and the number of trapped particles increases for the H concentration of 1%. The fraction of both trapped-lost particles and co-passing-lost particles increases. For the H concentration of 5%, the fraction of trapped-lost particles is also increased. For the case of minority H concentration of 5%, the heating of the minority H is dominant through fundamental frequency heating, and the heating of the NBI D ion is not dominant. So compared with the NBI only case, the difference of particle orbital type is tiny for minority H concentration of 5% case. However, the anisotropy of the distribution of NBI D ions in the poloidal dimension due to ICRF results in more particles in the ripple loss region, which is responsible for more NBI fast ion loss for minority H concentration of 5% case compared to the case of NBI only. The results are shown in next subsection.

Fast ion loss ratio and initial distribution of lost particles
We calculate the NBI fast ion orbitals for different minority concentrations with collision-only, ripple field-only, and both collision and ripple fields when the boundary is the limiters, as shown in figure 13. The blue line represents the banana orbit, and the red line represents the passing orbit. The additional drift of the particle trajectory will occur under the effect of collision and ripple fields. We then make the total time 501 ms and count the fast ion loss fraction with time for these three cases, as shown in figure 14. Comparing the loss of fast ions with or without ICRF and with different minority H concentrations, we can find that the fraction of fast ion loss is larger with ICRF than without ICRF, while the largest fraction of fast ion loss is for the minority H concentration of 1%. Ripple fields and collisions also increase the fraction of fast ion loss, and the fraction of lost fast ions brought about by ripples is larger than that caused by collisions. The loss caused by the collision includes    the loss of pitch-angle scattering loss and random diffusion loss, and the loss caused by the ripple field includes the ripple magnetic-well trapped loss and ripple diffusion loss. The synergistic loss caused by ripple field and collision will first be greater than the loss caused by either factor, but because of the collision, some fast ions will become thermal particles and the percentage of fast ion loss eventually reaches a fixed value for simulation time greater than 200 ms. With NBI only, the steady value for fast ion loss is 3%. The steady value for fast ion loss is 3.8% for the minority H concentration of 1%. It worth noticing that for core plasma, the influence of excess loss fraction of 0.8% is small. But for the heat load on limiters, the change from 3% to 3.8% means 26.7% change of heat load.
Its enhancement of the influence of the heat load on limiters should be considered.
To investigate the loss position distribution of the fast ions, we count the 3D distribution of the fast ion loss positions when the minority ion concentration is 1% for including both the ripple field and the collision case. Compared with the general FW boundary, the limiter is not symmetrical in the toroidal dimension, and the 3D images need to be drawn for the fast ion loss distribution. The resulting loss distribution of ICRF-NBI synergy induced fast ions under the ripple field and collision with the minority H concentration of 1% is shown in figures 15 and 16. From figure 15, we can see that the fast ions are lost in the upper half-plane, and the fast ions lost near the mid-plane are on limiters. Figure 16 shows the top view of the lost fast ion's distribution and the initial positions distribution of the lost ions in the poloidal cross-section. Since most of the lost ions are caused by the ripple field, the initial position distribution of the lost ions is in the ripple loss region, as shown in the [28]. And we can see from figure 9 that the LH B limiters are farthest from the major plasma, while the main limiter is closest to the major plasma. The fraction of lost fast ions on FW and each limiter is shown in figure 17. It shows in figure 17(a) that compared with the fraction of lost fast ions on FW, more particles will be lost on limiters. The fraction of fast ions lost on each limiter is related to the distance from the limiter to the major plasma and the relative position of the limiters, as well as the parallel velocity direction of most of the fast ions. For example, since the main limiter is closest to the major plasma, the corresponding loss fraction is the largest. For both the LH limiter and the IC limiter, the loss fraction on the left limiter is larger than that on the right limiter because most of the fast ion parallel velocity directions are in the current direction, i.e. counterclockwise, and the lost fast ions will touch the left limiter first. Although the IC I limiters are at the same distance from the major plasma as the IC N limiters, most of the lost fast ions contact the main limiter first because the main limiter is closer to the IC I limiters, resulting in a smaller fraction of loss fast ions on the IC I limiters compared to the IC N limiters.

Heat load distribution of fast ions on the limiter
After obtaining the lost fast ions' distribution on the FW and the limiters, we counted the heat load power density of fast ions on the FW and the limiters along the toroidal and poloidal angles, as shown in figure 18. The lost fast ion distribution on the FW of the upper half-plane shows a periodicity because of the ripple field. While the lost fast ion distribution on the limiters is nonperiodic as the limiters distributed nonperiodic. In order to get the heat load power density distribution of the lost particles on each limiter, we need to do a more accurate partitioning of the FW grid and integrate the time translation for a single sample. We take the sampling time as 1 ms and the total time as 501 ms. It derives the heat load power density distribution on the limiter in the actual case from the scale factor of the actual injected power and the sampled power. We assume the actual power of ICRF-NBI synergy induced fast ions is 2 MW. For minority H concentration of 1%, we get the heat load     power density distribution on a single limiter as in figures 19-21. Combining the results of figure 17, we can get the following four characteristics of the heat load power density distribution on the limiters: 1. The heat load power density on the left limiters is larger than that on the right limiters; 2. The heat load power density distribution on the left limiters is more concentrated compared to that on the right limiters; 3. The heat load power density on the left side of each limiter is larger.
Combined with the fast ion orbit types distribution in figure 12, we can see that most of the particles have a counterclockwise parallel velocity direction, corresponding to the current direction, which is one reason for most of the particles deposited on the left side of the limiters; 4. The heat load power density on the main limiter is 0.2 MW · m −2 , which is about twice as large as that on the LH limiter and the IC limiter, so the main limiter plays a major role in protecting the device.

Conclusion
In this paper, the orbit loss and the heat load power deposition distribution of ICRF-NBI synergy induced fast ions on the limiters in EAST are calculated by the full-orbit-following program ISSDE. The dynamic behavior of these fast ions are under the synergistic effect of ripple field and collision. Plasma parameters and equilibrium are provided by EFIT and the initial distribution of ICRF-NBI synergy induced fast ions is provided by the TRANSP which integrated calculation of NUBEAM and TORIC modules. Compared with the ISSDE version provided in [34], we add the 3D limiter boundary and the ripple field for the first time in the new version. The simulation results show the ICRF makes the NBI fast ions have more energetic particle components and changes the orbital type distribution of NBI fast ions, which increases the fraction of trapped particles and thus enhances the loss of fast ions. The ripple field and collisions have a persistent effect on the fast ion loss, which further enhances the fast ion loss. Compared with the effect of collisions, the effect of the ripple field is greater, and the ripple field makes the lost fast ion in the upper half-plane show a periodicity. The ripple field and collision have a synergistic effect on the fast ion loss. The limiters make the loss of NBI fast ions in the mid-plane not periodic but in the limiters. Because of the different distances of each limiter from the major plasma, the fraction of fast ion loss on each limiter varies, with the largest fraction of fast ion loss on the major limiter, followed by the IC limiters. It related the lost fast ion distribution deposition on the limiter to the parallel direction of most fast ions, and the lost fast ion heat load power density is greater on the left side of each limiter.
Although this article only addresses on fast ion loss and heat load physics on EAST, the approach applied in the paper can be extended to more devices, such as ITER and DAMO, which can support higher parameters of operation and will produce a higher fraction of energetic ions, such as alpha particles from fusion [35][36][37]. These high-energy ions place higher demands on the thermal load carrying capacity of the limiters. Using the methods and procedures shown in the text, we can gain a clearer understanding of these device designs and the underlying physical mechanisms. ISSDE, as a test particle tracking parallel simulation program, can calculate the dynamical behaviour of simulated particles under various external fields and perturbations, especially for multiscale multi-degree-of-freedom physics problems. However, the effect of self-consistent fields is not considered, and the collision-containing kinetic theory particle PIC program can give more information, which we will present in our future work. In this paper, only the loss mechanisms under the influence of ripples and collisions are considered. The causes of fast ion losses in real EAST are diverse and more complex, such as losses caused by TBM fields, magnetic fluid instability modes, resonant magnetic perturbation coils, tearing mode instability, etc. The impact of collisions on each other will have different effects on these loss mechanisms, which will be presented in our future work.