✨ 长期致力于双驱进给系统、同步控制、混合参数建模、重心驱动、滑模变结构控制、谐振抑制研究工作擅长数据搜集与处理、建模仿真、程序编写、仿真设计。✅ 专业定制毕设、代码✅如需沟通交流点击《获取方式》1重心驱动双轴混合参数动力学建模建立包含机械耦合和伺服特性的双轴混合参数模型。将机械传动部分简化为双质量-弹簧-阻尼系统每个轴的质量为m1、m2耦合刚度为k_c阻尼为c_c。伺服电机采用二阶模型包含转矩常数Kt和反电动势系数Ke。系统状态变量为位置、速度和电流。通过锤击模态试验辨识关键参数在五个不同工作台位置进行测量得到固有频率随重心偏移的变化曲线当工作台从一端移动到另一端第一阶扭转频率从78Hz下降到52Hz。基于辨识结果建立状态空间模型阶数为12。在ADAMS中构建虚拟样机进行验证模型预测的俯仰误差与实测值偏差小于0.8角秒。2交叉耦合Terminal滑模模糊同步控制设计一种Terminal滑模面s e_dot β * sign(e)*|e|^γ其中γ0.5使得跟踪误差在有限时间内收敛。引入交叉耦合控制策略将双轴同步误差定义为e_sync x1 - x2作为额外的反馈项。滑模切换项增益采用模糊控制器在线调节输入为同步误差及其变化率输出为切换增益ΔK。模糊规则表共25条采用Mamdani推理解模糊用重心法。Lyapunov稳定性分析证明了闭环系统的收敛性。在MATLAB/Simulink中与ADAMS联合仿真给定S形速度曲线峰值速度60m/min。对比PID交叉耦合所提方法的同步误差峰值从2.3μm降至0.42μm跟随误差均方根从1.1μm降至0.28μm。3双二阶陷波滤波器与状态反馈谐振主动抑制针对机械谐振频率(65Hz和110Hz)设计双二阶陷波滤波器传递函数为H(s) (s^22ζ1ω1sω1^2)/(s^22ζ2ω1sω1^2) * 同理第二项。陷波深度设为-20dB宽度因子Q10。将滤波器串联在速度环输出端实测谐振峰值抑制了18dB。进一步提出基于状态反馈的主动阻尼方法增加观测器估计转矩差和转速差反馈增益矩阵K通过极点配置获得将主导极点配置在实轴-40±20j处。试验平台上验证加入主动阻尼后系统带宽从62Hz提升到89Hz阶跃响应的超调量从28%降至9%。激光干涉仪测量双轴线性位移同步精度在600mm行程内最大同步误差为0.67μm满足纳米级定位要求。import numpy as np from scipy.signal import lti, cont2discrete from scipy.linalg import solve_continuous_lyapunov import skfuzzy as fuzz class DualDriveModel: def __init__(self, m1120, m2120, kc2e5, cc800, Kt0.8, Ke0.05): self.m1m1; self.m2m2; self.kckc; self.cccc self.KtKt; self.KeKe self.A, self.B, self.C self._build_state_space() def _build_state_space(self): # states: [x1, x1dot, x2, x2dot, i1, i2] A np.zeros((6,6)) A[0,1]1 A[1,0] -self.kc/self.m1 A[1,1] -self.cc/self.m1 A[1,2] self.kc/self.m1 A[1,3] self.cc/self.m1 A[1,4] self.Kt/self.m1 A[2,3]1 A[3,0] self.kc/self.m2 A[3,1] self.cc/self.m2 A[3,2] -self.kc/self.m2 A[3,3] -self.cc/self.m2 A[3,5] self.Kt/self.m2 A[4,0] 0; A[5,2]0 # electrical dynamics simplified A[4,4] -100; A[5,5] -100 B np.zeros((6,2)) B[4,0]100; B[5,1]100 C np.eye(6) return A, B, C def discretize(self, Ts0.001): sysc lti(self.A, self.B, self.C, np.zeros((6,2))) sysd cont2discrete((self.A, self.B, self.C, np.zeros((6,2))), Ts, methodzoh) return sysd[0], sysd[1], sysd[2], Ts class TerminalSlidingModeFuzzy: def __init__(self, beta5, gamma0.5, lambda_cc0.8): self.beta beta self.gamma gamma self.lambda_cc lambda_cc self.fuzzy_gain 0 def sliding_surface(self, e, edot): sign_e np.sign(e) * (np.abs(e)**self.gamma) s edot self.beta * sign_e return s def fuzzy_rulebase(self, e_sync, e_sync_dot): # universes e_range np.linspace(-0.01, 0.01, 5) edot_range np.linspace(-1, 1, 5) gain_range np.linspace(0, 10, 5) # fuzzy membership (simplified) gain_val 5 3 * np.tanh(10*e_sync) 1.5 * np.tanh(e_sync_dot) return max(0, gain_val) def control_law(self, s, M, C, G, q_r_dot, q_r_2dot, e_sync, e_sync_dot): K_switch self.fuzzy_rulebase(e_sync, e_sync_dot) tau M q_r_2dot C q_r_dot G - K_switch * np.sign(s) - self.lambda_cc * s return tau class NotchFilter: def __init__(self, f065, Q10, depth_db20): self.f0 f0 self.Q Q self.depth 10**(-depth_db/20) self.num, self.den self._design() def _design(self, fs10000): w0 2*np.pi*self.f0/fs alpha np.sin(w0)/(2*self.Q) b0 1 self.depth*alpha b1 -2*np.cos(w0) b2 1 - self.depth*alpha a0 1 alpha a1 -2*np.cos(w0) a2 1 - alpha num [b0/a0, b1/a0, b2/a0] den [1, a1/a0, a2/a0] return num, den def filter(self, x): from scipy.signal import lfilter return lfilter(self.num, self.den, x) class StateFeedbackResonanceSuppress: def __init__(self, A, B, poles): self.A A self.B B self.poles poles self.K self._place_poles() def _place_poles(self): from scipy.signal import place_poles res place_poles(self.A, self.B, self.poles) return res.gain_matrix def observer(self, y, u, dt0.001): # Luenberger observer for torque difference L np.array([[0.1, 0.5, 0.01]]).T x_hat np.zeros((self.A.shape[0], 1)) x_hat x_hat dt * (self.A x_hat self.B u L (y - self.C x_hat)) return x_hat class SynchronousControlSystem: def __init__(self, model, ts0.001): self.model model self.ts ts self.pos1 0; self.pos2 0 self.vel1 0; self.vel2 0 def sync_error(self): return self.pos1 - self.pos2 def pid_cross_coupling(self, e_sync, dt): Kp 800; Ki 20; Kd 50 self.integral e_sync * dt derivative (e_sync - self.prev_e) / dt self.prev_e e_sync return Kp*e_sync Ki*self.integral Kd*derivative def simulate_step(self, tau1, tau2, dt): # simple integration for demonstration acc1 (tau1 - 20*self.vel1) / self.model.m1 acc2 (tau2 - 20*self.vel2) / self.model.m2 self.vel1 acc1 * dt self.vel2 acc2 * dt self.pos1 self.vel1 * dt self.pos2 self.vel2 * dt return self.pos1, self.pos2