419 lines
22 KiB
Python
419 lines
22 KiB
Python
import json
|
||
import openpyxl
|
||
import pandas
|
||
import re
|
||
import csv
|
||
from PySide6.QtCore import Signal, QThread
|
||
import time
|
||
from codes.common import clibs
|
||
|
||
|
||
class CurrentDataProcess(QThread):
|
||
def __init__(self, dir_path, proc, /):
|
||
super().__init__()
|
||
self.dir_path = dir_path
|
||
self.proc = proc
|
||
self.idx = 1
|
||
|
||
def initialization(self):
|
||
_, data_files = clibs.traversal_files(self.dir_path)
|
||
count, config_file = 0, None
|
||
for data_file in data_files:
|
||
filename = data_file.split("/")[-1]
|
||
if re.match(".*\\.cfg", filename):
|
||
config_file = data_file
|
||
count += 1
|
||
elif filename == "T_电机电流.xlsx":
|
||
count += 1
|
||
else:
|
||
if not re.match("^j[1-7].*\\.data$", filename):
|
||
msg = f"不合规 {data_file}<br>"
|
||
msg += "所有数据文件必须以 j[1-7]_ 开头，以 .data 结尾，比如j1_abcdef.data，请检查整改后重新运行"
|
||
clibs.logger("ERROR", "current", msg, "red")
|
||
|
||
if count != 2:
|
||
msg = "需要有一个机型配置文件\"*.cfg\"，以及一个数据处理文件\"T_电机电流.xlsx\"表格，请检查整改后重新运行"
|
||
clibs.logger("ERROR", "current", msg, "red")
|
||
|
||
return data_files, config_file
|
||
|
||
def current_max(self, data_files, rts):
|
||
clibs.logger("INFO", "current", f"正在处理最大转矩值逻辑......")
|
||
current = {1: [], 2: [], 3: [], 4: [], 5: [], 6: []}
|
||
for data_file in data_files:
|
||
if data_file.endswith(".data"):
|
||
df = pandas.read_csv(data_file, sep="\t")
|
||
else:
|
||
continue
|
||
|
||
clibs.logger("DEBUG", "current", f"正在处理 {data_file} ...")
|
||
cols = len(df.columns)
|
||
axis = int(data_file.split("/")[-1].split("_")[0].removeprefix("j"))
|
||
rt = rts[axis-1]
|
||
clibs.logger("DEBUG", "current", f"最大列数为 {cols}，{axis} 轴的额定转矩为 {rt}")
|
||
|
||
col = df.columns.values[clibs.c_servo_trq-1] # 获取 "device_servo_trq_feedback"
|
||
c_max = df[col].abs().max()
|
||
|
||
scale = 1000
|
||
_ = abs(c_max/scale*rt)
|
||
current[axis].append(_)
|
||
clibs.logger("DEBUG", "current", f"{data_file}: {_:.2f}")
|
||
clibs.logger("DEBUG", "current", f"获取到的列名为 {col}，最大转矩为 {_}")
|
||
|
||
with open(data_file, "a+") as f_data:
|
||
csv_writer = csv.writer(f_data, delimiter="\t")
|
||
csv_writer.writerow([""] * (cols-1) + [_])
|
||
|
||
for axis, cur in current.items():
|
||
if not cur:
|
||
continue
|
||
else:
|
||
_ = ""
|
||
for value in cur:
|
||
_ += f"{value:.4f} "
|
||
clibs.logger("INFO", "current", f"{axis}轴最大转矩数据：{_}")
|
||
|
||
clibs.logger("DEBUG", "current", f"获取最大转矩值结束 current_max = {current}")
|
||
clibs.logger("INFO", "current", f"最大转矩数据处理完毕......")
|
||
return current
|
||
|
||
def current_avg(self, data_files, rts):
|
||
clibs.logger("INFO", "current", f"正在处理平均转矩值逻辑......")
|
||
current = {1: [], 2: [], 3: [], 4: [], 5: [], 6: []}
|
||
for data_file in data_files:
|
||
if data_file.endswith(".data"):
|
||
df = pandas.read_csv(data_file, sep="\t")
|
||
else:
|
||
continue
|
||
|
||
clibs.logger("DEBUG", "current", f"正在处理 {data_file} ...")
|
||
cols = len(df.columns)
|
||
axis = int(data_file.split("/")[-1].split("_")[0].removeprefix("j"))
|
||
rt = rts[axis-1]
|
||
clibs.logger("DEBUG", "current", f"最大列数为 {cols}，{axis} 轴的额定转矩为 {rt}")
|
||
|
||
col = df.columns.values[clibs.c_servo_trq-1]
|
||
c_std = df[col].std()
|
||
c_avg = df[col].mean()
|
||
|
||
scale = 1000
|
||
_ = (abs(c_avg)+c_std*3)/scale*rt
|
||
current[axis].append(_)
|
||
clibs.logger("DEBUG", "current", f"{data_file}: {_:.2f}")
|
||
clibs.logger("DEBUG", "current", f"获取到的列名为 {col}，平均转矩为 {_}")
|
||
|
||
with open(data_file, "a+") as f_data:
|
||
csv_writer = csv.writer(f_data, delimiter="\t")
|
||
csv_writer.writerow([""] * (cols-1) + [_])
|
||
|
||
for axis, cur in current.items():
|
||
if not cur:
|
||
continue
|
||
else:
|
||
_ = ""
|
||
for value in cur:
|
||
_ += f"{value:.4f} "
|
||
clibs.logger("INFO", "current", f"{axis}轴平均转矩数据：{_}")
|
||
|
||
clibs.logger("DEBUG", "current", f"获取平均转矩值结束 current_avg = {current}")
|
||
clibs.logger("INFO", "current", f"平均转矩数据处理完毕......")
|
||
return current
|
||
|
||
def current_cycle(self, data_files, rrs, rts, params):
|
||
result, hold, single, scenario, dur_time = None, [], [], [], 0
|
||
for data_file in data_files:
|
||
filename = data_file.split("/")[-1]
|
||
if filename == "T_电机电流.xlsx":
|
||
result = data_file
|
||
elif re.match("j[1-7]_hold_.*\\.data", filename):
|
||
hold.append(data_file)
|
||
elif re.match("j[1-7]_s_.*\\.data", filename):
|
||
scenario.append(data_file)
|
||
dur_time = float(filename.split("_")[3])
|
||
elif re.match("j[1-7]_.*\\.data", filename):
|
||
single.append(data_file)
|
||
|
||
clibs.stop, filename = True, result.split("/")[-1]
|
||
clibs.logger("INFO", "current", f"正在打开文件 {filename}，这可能需要一些时间......", "blue")
|
||
try:
|
||
wb = openpyxl.load_workbook(result)
|
||
except Exception as err:
|
||
clibs.logger("ERROR", "current", f"{filename}文件打开失败，可能是文件已损坏，确认后重新执行！<br>{err}", "red")
|
||
|
||
ws = wb["统计"]
|
||
for idx in range(len(params)-1):
|
||
row = idx + 2
|
||
for col in range(2, 8):
|
||
ws.cell(row=row, column=col).value = params[idx][col-2]
|
||
ws.cell(row=1, column=1).value = params[-1]
|
||
|
||
if hold:
|
||
avg = self.current_avg(hold, rts)
|
||
for axis, cur_value in avg.items():
|
||
sht_name = f"J{axis}"
|
||
wb[sht_name]["P4"].value = float(cur_value[0])
|
||
|
||
if dur_time == 0:
|
||
self.p_single(wb, single, rrs)
|
||
else:
|
||
self.p_scenario(wb, scenario, rrs, dur_time)
|
||
|
||
clibs.logger("INFO", "current", f"正在保存文件 {filename}，这可能需要一些时间......", "blue")
|
||
wb.save(result)
|
||
wb.close()
|
||
|
||
def find_point(self, data_file, df, flag, row_s, row_e, threshold, step, end_point, skip_scale, axis, seq):
|
||
if flag == "lt":
|
||
while row_e > end_point:
|
||
speed_avg = df.iloc[row_s:row_e].abs().mean()
|
||
if speed_avg < threshold:
|
||
row_e -= step
|
||
row_s -= step
|
||
continue
|
||
else:
|
||
# one more time，如果连续两次 200 个点的平均值都大于 threshold，说明已经到了临界点了（其实也不一定，只不过相对遇到一次就判定临界点更安全一点点）
|
||
# 从实际数据看，这开逻辑很小概率能触发到
|
||
speed_avg = df.iloc[row_s-end_point*skip_scale:row_e-end_point*skip_scale].abs().mean()
|
||
if speed_avg < threshold:
|
||
clibs.logger("WARNING", "current", f"【lt】{axis} 轴第 {seq} 次查找数据可能有异常，row_s = {row_s}, row_e = {row_e}！", "purple")
|
||
return row_s, row_e
|
||
else:
|
||
clibs.logger("ERROR", "current", f"{data_file} 数据有误，需要检查，无法找到第 {seq} 个有效点......", "red")
|
||
elif flag == "gt":
|
||
while row_e > end_point:
|
||
speed_avg = df.iloc[row_s:row_e].abs().mean()
|
||
if speed_avg > threshold:
|
||
row_e -= step
|
||
row_s -= step
|
||
continue
|
||
else:
|
||
# one more time，如果连续两次 200 个点的平均值都小于 threshold，说明已经到了临界点了（其实也不一定，只不过相对遇到一次就判定临界点更安全一点点）
|
||
# 从实际数据看，这开逻辑很小概率能触发到
|
||
speed_avg = df.iloc[row_s-end_point*skip_scale:row_e-end_point*skip_scale].abs().mean()
|
||
if speed_avg > threshold:
|
||
clibs.logger("WARNING", "current", f"【gt】{axis} 轴第 {seq} 次查找数据可能有异常，row_s = {row_s}, row_e = {row_e}！", "purple")
|
||
return row_s, row_e
|
||
else:
|
||
clibs.logger("ERROR", "current", f"{data_file} 数据有误，需要检查，无法找到第 {seq} 个有效点......", "red")
|
||
|
||
def get_row_number(self, threshold, flag, df, row_s, row_e, axis):
|
||
count_1, count_2 = 0, 0
|
||
if flag == "start" or flag == "end":
|
||
for number in df.iloc[row_s:row_e].abs():
|
||
count_2 += 1
|
||
if number > threshold:
|
||
count_1 += 1
|
||
if count_1 == 10:
|
||
return row_s + count_2 - 10
|
||
else:
|
||
count_1 = 0
|
||
elif flag == "middle":
|
||
for number in df.iloc[row_s:row_e].abs():
|
||
count_2 += 1
|
||
if number < threshold: # 唯一的区别
|
||
count_1 += 1
|
||
if count_1 == 10:
|
||
return row_s + count_2 - 10
|
||
else:
|
||
count_1 = 0
|
||
|
||
places = {"start": "起点", "middle": "中间点", "end": "终点"} # 因为是终点数据，所以可能有异常
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴获取{places[flag]}数据 {row_e} 可能有异常，需关注！", "purple")
|
||
return row_e
|
||
|
||
def p_single(self, wb, single, rrs):
|
||
# 1. 先找到第一个速度为零的点，数据从后往前找，一开始就是零的情况不予考虑
|
||
# 2. 记录第一个点的位置，继续向前查找第二个速度为零的点，同理，一开始为零的点不予考虑
|
||
# 3. 记录第二个点的位置，并将其中的数据拷贝至对应位置
|
||
for data_file in single:
|
||
axis = int(data_file.split("/")[-1].split("_")[0].removeprefix("j"))
|
||
sht_name = f"J{axis}"
|
||
ws = wb[sht_name]
|
||
pandas.set_option("display.precision", 2)
|
||
df_origin = pandas.read_csv(data_file, sep="\t")
|
||
rr = rrs[axis-1]
|
||
addition = 180 / 3.1415926 * 60 / 360 * rr
|
||
|
||
col_names = list(df_origin.columns)
|
||
df = df_origin[col_names[clibs.c_joint_vel-1]].multiply(addition)
|
||
|
||
step = 50 # 步进值
|
||
end_point = 200 # 有效数值的数目
|
||
threshold = 5 # 200个点的平均阈值线
|
||
skip_scale = 2
|
||
row_start, row_middle, row_end = 0, 0, 0
|
||
row_e = df.index[-1]
|
||
row_s = row_e - end_point
|
||
speed_avg = df.iloc[row_s:row_e].abs().mean()
|
||
if speed_avg < threshold:
|
||
# 第一次过滤：消除速度为零的数据，找到速度即将大于零的上升临界点
|
||
row_s, row_e = self.find_point(data_file, df, "lt", row_s, row_e, threshold, step, end_point, skip_scale, axis, "pre-1")
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第二次过滤：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, "pre-2")
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第三次过滤：消除速度为零的数据，找到速度即将大于零的上升临界点
|
||
row_s, row_e = self.find_point(data_file, df, "lt", row_s, row_e, threshold, step, end_point, skip_scale, axis, "pre-3")
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 正式第一次采集：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 1)
|
||
row_end = self.get_row_number(threshold, "end", df, row_s, row_e, axis)
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 正式第二次采集：消除速度为零的数据，找到速度即将大于零的上升临界点
|
||
row_s, row_e = self.find_point(data_file, df, "lt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 2)
|
||
row_middle = self.get_row_number(threshold, "middle", df, row_s, row_e, axis)
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 正式第三次采集：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 3)
|
||
row_start = self.get_row_number(threshold, "start", df, row_s, row_e, axis)
|
||
elif speed_avg > threshold:
|
||
# 第一次过滤：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, "pre-1")
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第二次过滤：消除速度为零的数据，找到速度即将大于零的上升临界点
|
||
row_s, row_e = self.find_point(data_file, df, "lt", row_s, row_e, threshold, step, end_point, skip_scale, axis, "pre-2")
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第一次正式采集：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 1)
|
||
row_end = self.get_row_number(threshold, "end", df, row_s, row_e, axis)
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第二次正式采集：消除速度为零的数据，找到速度即将大于零的上升临界点
|
||
row_s, row_e = self.find_point(data_file, df, "lt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 2)
|
||
row_middle = self.get_row_number(threshold, "middle", df, row_s, row_e, axis)
|
||
row_e -= end_point*skip_scale
|
||
row_s -= end_point*skip_scale
|
||
# 第三次正式采集：消除速度大于零的数据，找到速度即将趋近于零的下降临界点
|
||
row_s, row_e = self.find_point(data_file, df, "gt", row_s, row_e, threshold, step, end_point, skip_scale, axis, 3)
|
||
row_start = self.get_row_number(threshold, "start", df, row_s, row_e, axis)
|
||
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴起点：{row_start}")
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴中间点：{row_middle}")
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴终点：{row_end}")
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴数据非零段点数：{row_middle-row_start+1}")
|
||
clibs.logger("DEBUG", "current", f"{axis} 轴数据为零段点数：{row_end-row_middle+1}")
|
||
if abs(row_end+row_start-2*row_middle) > 1000:
|
||
clibs.logger("WARNING", "current", f"{axis} 轴数据占空比异常，处理数据可能有误，需检查！", "purple")
|
||
|
||
data, first_c, second_c, third_c, fourth_c = [], clibs.c_joint_vel-1, clibs.c_servo_trq-1, clibs.c_sensor_trq-1, clibs.c_estimate_trans_trq-1
|
||
for row in range(row_start, row_end+1):
|
||
data.append(df_origin.iloc[row, first_c])
|
||
data.append(df_origin.iloc[row, second_c])
|
||
data.append(df_origin.iloc[row, third_c])
|
||
data.append(df_origin.iloc[row, fourth_c])
|
||
|
||
i = 0
|
||
for row in ws.iter_rows(min_row=2, min_col=2, max_row=150000, max_col=5):
|
||
for cell in row:
|
||
try:
|
||
if i % 4 == 0:
|
||
ws.cell((i//4)+2, 1).value = float(((i//4)+1)/1000)
|
||
_ = f"{data[i]:.2f}"
|
||
cell.value = float(_)
|
||
i += 1
|
||
except Exception:
|
||
if i % 4 == 0:
|
||
ws.cell((i//4)+2, 1).value = None
|
||
cell.value = None
|
||
i += 1
|
||
|
||
def p_scenario(self, wb, scenario, rrs, dur_time):
|
||
clibs.logger("INFO", "current", f"本次处理的是电机电流场景数据，场景运动周期为 {dur_time}s", "blue")
|
||
for data_file in scenario:
|
||
cycle = 0.001
|
||
axis = int(data_file.split("/")[-1].split("_")[0].removeprefix("j"))
|
||
sht_name = f"J{axis}"
|
||
ws = wb[sht_name]
|
||
pandas.set_option("display.precision", 2)
|
||
df_origin = pandas.read_csv(data_file, sep="\t")
|
||
rr = rrs[axis-1]
|
||
addition = 180 / 3.1415926 * 60 / 360 * rr
|
||
|
||
col_names = list(df_origin.columns)
|
||
df = df_origin[col_names[clibs.c_joint_vel-1]].multiply(addition)
|
||
|
||
row_start = 3000
|
||
row_end = row_start + int(dur_time/cycle)
|
||
if row_end > df.index[-1]:
|
||
clibs.logger("ERROR", "current", f"位置超限：{data_file} 共有 {df.index[-1]} 条数据，无法取到第 {row_end} 条数据，需要确认场景周期时间...", "red")
|
||
|
||
data, first_c, second_c, third_c, fourth_c = [], clibs.c_joint_vel-1, clibs.c_servo_trq-1, clibs.c_sensor_trq-1, clibs.c_estimate_trans_trq-1
|
||
for row in range(row_start, row_end+1):
|
||
data.append(df_origin.iloc[row, first_c])
|
||
data.append(df_origin.iloc[row, second_c])
|
||
data.append(df_origin.iloc[row, third_c])
|
||
data.append(df_origin.iloc[row, fourth_c])
|
||
|
||
i = 0
|
||
for row in ws.iter_rows(min_row=2, min_col=2, max_row=250000, max_col=5):
|
||
for cell in row:
|
||
try:
|
||
if i % 4 == 0:
|
||
ws.cell((i//4)+2, 1).value = float(((i//4)+1)/1000)
|
||
_ = f"{data[i]:.2f}"
|
||
cell.value = float(_)
|
||
i += 1
|
||
except Exception:
|
||
cell.value = None
|
||
if i % 4 == 0:
|
||
ws.cell((i//4)+2, 1).value = None
|
||
i += 1
|
||
|
||
def get_configs(self, config_file):
|
||
try:
|
||
if re.match("^[NXEC]B.*", config_file.split("/")[-1]):
|
||
robot_type = "工业"
|
||
else:
|
||
robot_type = "协作"
|
||
|
||
with open(config_file, mode="r", encoding="utf-8") as f_config:
|
||
configs = json.load(f_config)
|
||
|
||
version = configs["VERSION"]
|
||
sc = [0.001, 0.001, 0.001, 0.001, 0.001, 0.001] # 采样周期，sc for sample cycle
|
||
r_rrs = configs["TRANSMISSION"]["REDUCTION_RATIO_NUMERATOR"] # 减速比，rr for reduction ratio
|
||
m_avs = configs["MOTION"]["JOINT_MAX_SPEED"]
|
||
m_stall_ts = configs["MOTOR"]["STALL_TORQUE"] # 电机堵转转矩
|
||
m_rts = configs["MOTOR"]["RATED_TORQUE"] # 电机额定转矩rt for rated torque
|
||
m_max_ts = configs["MOTOR"]["PEAK_TORQUE"] # 电机峰值转矩
|
||
m_r_rpms = configs["MOTOR"]["RATED_SPEED"] # 电机额定转速
|
||
m_max_rpms = configs["MOTOR"]["MAX_SPEED"] # 电机最大转速
|
||
r_max_sst = configs["TRANSMISSION"]["MAX_TORQUE_FOR_START_AND_STOP"] # 减速器最大启停转矩，sst for start and stop torque
|
||
r_max_t = configs["TRANSMISSION"]["MAX_PEAK_TORQUE"] # 减速器瞬时最大转矩
|
||
r_avg_t = configs["TRANSMISSION"]["MAX_AVERAGE_TORQUE"] # 减速器平均负载转矩允许最大值
|
||
|
||
clibs.logger("INFO", "current", f"get_configs: 机型文件版本 {config_file}_{version}")
|
||
clibs.logger("INFO", "current", f"get_configs: 减速比 {r_rrs}")
|
||
clibs.logger("INFO", "current", f"get_configs: 额定转矩 {m_rts}")
|
||
clibs.logger("INFO", "current", f"get_configs: 最大角速度 {m_avs}")
|
||
return sc, r_rrs, m_avs, m_stall_ts, m_rts, m_max_ts, m_r_rpms, m_max_rpms, r_max_sst, r_max_t, r_avg_t, robot_type
|
||
except Exception as err:
|
||
clibs.logger("ERROR", "current", f"get_config: 无法打开 {config_file}，或获取配置文件参数错误 {err}", "red")
|
||
|
||
def processing(self):
|
||
time_start = time.time()
|
||
clibs.running[self.idx] = 1
|
||
|
||
data_files, config_file = self.initialization()
|
||
params = self.get_configs(config_file)
|
||
rts, rrs = params[4], params[1]
|
||
if self.proc == "最大值":
|
||
self.current_max(data_files, rts)
|
||
elif self.proc == "平均值":
|
||
self.current_avg(data_files, rts)
|
||
elif self.proc == "周期":
|
||
self.current_cycle(data_files, rrs, rts, params)
|
||
|
||
clibs.logger("INFO", "current", "-"*60 + "<br>全部处理完毕<br>", "purple")
|
||
time_total = time.time() - time_start
|
||
msg = f"数据处理时间：{time_total // 3600:02.0f} h {time_total % 3600 // 60:02.0f} m {time_total % 60:02.0f} s"
|
||
clibs.logger("INFO", "current", msg)
|