深入解析C-MAPSS数据集
C-MAPSS数据集由NASA为研究航空发动机故障预测而开发的。备受发动机寿命预测科研儿的青睐，在各种顶级期刊频繁出现。因其数据比较完整，整理较为方便，且数据的拟合难度较高，是检验深度学习算法性能的试金石。本期推文采用MATLAB代码实现对该数据的深度剖析。
一、数据集概览
C-MAPSS（Commercial Modular Aero-Propulsion System Simulation）是航空发动机退化仿真数据集，官方数据获取链接：https://data.nasa.gov/dataset/cmapss-jet-engine-simulated-data
C-MAPSS数据集由四个独立子集（FD001-FD004）构成，每个子集针对不同的工况与故障场景设计，具体配置如下：
| 子集 | 训练样本数 | 测试样本数 | 运行条件 | 故障模式 |
|---|---|---|---|---|
FD001 | 100 | 100 | 单一工况 | 高压压气机（HPC）退化 |
FD002 | 260 | 259 | 六种混合工况 | HPC退化 |
FD003 | 100 | 100 | 单一工况 | HPC+风扇复合退化 |
FD004 | 248 | 249 | 六种混合工况 | HPC+风扇复合退化 |
FD001/FD003聚焦单一工况，FD002/FD004模拟多工况切换的复杂场景。FD003/FD004额外引入风扇退化问题。
以FD001为例，训练轨迹指每台发动机在其运行周期内的时间序列数据。每个训练轨迹对应于一台发动机的运行数据。100 条训练轨迹就是 100 台发动机， 每台发动机的初始磨损和制造工艺是不同的，这些差异被视为正常情况，而不是故障条件。
数据集中包含21个传感器记录的数值和不同operational setting，这些setting对发动机性能有显著影响。
C-MAPSS 数据集的物理含义
传感器分布如图：
在训练集中，模拟发动机从健康状态到故障状态的完整退化过程；而在测试集中，提供部分发动机的衰退轨迹，数据在故障发生之前随机截断（即未到真正发生故障的时间点）。测试模型根据当前的传感器值推断剩余使用寿命。
二、手撕MATLAB代码深度分析数据
读取C-MAPSS数据：
% 数据加载核心代码
clear; clc; close all;
addpath(genpath(pwd)); % 添加当前目录到路径
% 定义读取数据的函数
function data = get_data(file_name)
file_path = fullfile('C-MAPSS-Data\', file_name);
if exist(file_path, 'file') == 2
data = readtable(file_path, 'Delimiter', ' ', 'MultipleDelimsAsOne', true, ...
'ReadVariableNames', false,'HeaderLines', 0);
else
error('文件未找到: %s', file_path);
end
end
%% 加载所有数据集
% 训练数据
train_FD001 = get_data('train_FD001.txt');
train_FD002 = get_data('train_FD002.txt');
train_FD003 = get_data('train_FD003.txt');
train_FD004 = get_data('train_FD004.txt');
% RUL数据
RUL_FD001 = get_data('RUL_FD001.txt');
RUL_FD002 = get_data('RUL_FD002.txt');
RUL_FD003 = get_data('RUL_FD003.txt');
RUL_FD004 = get_data('RUL_FD004.txt');
% 测试数据
test_FD001 = get_data('test_FD001.txt');
test_FD002 = get_data('test_FD002.txt');
test_FD003 = get_data('test_FD003.txt');
test_FD004 = get_data('test_FD004.txt');
%% 设置列名
col_names = {'ID', 'cycle', 'op_set_1', 'op_set_2', 'op_set_3', ...
'sensor_1', 'sensor_2', 'sensor_3', 'sensor_4', 'sensor_5', ...
'sensor_6', 'sensor_7', 'sensor_8', 'sensor_9', 'sensor_10', ...
'sensor_11', 'sensor_12', 'sensor_13', 'sensor_14', 'sensor_15', ...
'sensor_16', 'sensor_17', 'sensor_18', 'sensor_19', 'sensor_20', ...
'sensor_21'};
col_name_RUL = {'RUL'};
% 为所有数据集分配列名
datasets = {train_FD001, train_FD002, train_FD003, train_FD004, ...
test_FD001, test_FD002, test_FD003, test_FD004, ...
RUL_FD001, RUL_FD002, RUL_FD003, RUL_FD004};
col_sets = {repmat({col_names}, 1, 8), repmat({col_name_RUL}, 1, 4)};
fori = 1:12
ifi <= 8
datasets{i}.Properties.VariableNames = col_names;
else
datasets{i}.Properties.VariableNames = col_name_RUL;
end
end
% 将更新后的数据集赋值回原始变量
train_FD001 = datasets{1};
train_FD002 = datasets{2};
train_FD003 = datasets{3};
train_FD004 = datasets{4};
test_FD001 = datasets{5};
test_FD002 = datasets{6};
test_FD003 = datasets{7};
test_FD004 = datasets{8};
RUL_FD001 = datasets{9};
RUL_FD002 = datasets{10};
RUL_FD003 = datasets{11};
RUL_FD004 = datasets{12};数据质量检查：
% 检查空值
fprintf('空值检查:\n');
fori = 1:8
null_count = sum(any(ismissing(datasets{i}), 1));
fprintf('数据集 %d: %d 个空值\n', i, null_count);
end
% 检查重复值
fprintf('\n重复值检查:\n');
fori = 1:8
[~, idx] = unique(datasets{i}, 'rows');
dup_count = size(datasets{i}, 1) - length(idx);
fprintf('数据集 %d: %d 个重复值\n', i, dup_count);
endMATLAB输出如下：
操作设置分布图：
%% 创建统一绘图样式
myColors = [0 0.4470 0.7410; 0.8500 0.3250 0.0980; 0.9290 0.6940 0.1250; ...
0.4940 0.1840 0.5560; 0.4660 0.6740 0.1880; 0.3010 0.7450 0.9330; ...
0.6350 0.0780 0.1840];
set(groot, 'defaultFigureColor', 'white', ...
'defaultAxesFontSize', 12, ...
'defaultAxesLabelFontSizeMultiplier', 1.1, ...
'defaultAxesTitleFontSizeMultiplier', 1.2, ...
'defaultLineLineWidth', 1.5);
%% 操作设置分析 (FD001)
% 提取操作设置列
op_set_cols = contains(col_names, 'op_set');
op_set_names = col_names(op_set_cols);
% 直方图分布
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
data = train_FD001.(op_set_names{i});
histogram(data, 'Normalization', 'probability', ...
'FaceColor', myColors(i, :));
title(sprintf('%s 分布', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i},'Interpreter', 'none');
ylabel('相对频率');
grid on;
box off;
if i==3
xlim([80, 120]); % 手动设置 x 轴范围
end
end
sgtitle('操作设置分布 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
操作设置随周期变化图：
% 操作设置随周期变化
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
scatter(train_FD001.(op_set_names{i}), train_FD001.cycle, 8, ...
train_FD001.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 随周期变化', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i}, 'Interpreter', 'none');
ylabel('周期');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('操作设置随周期变化 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
传感器分析 (FD001)
以FD001为例：
% 提取传感器列
sensor_cols = contains(col_names, 'sensor');
sensor_names = col_names(sensor_cols);
% 传感器直方图
figure('Position', [100, 100, 1400, 800]);
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
data = train_FD001.(sensor_names{i});
histogram(data, 'Normalization', 'probability', ...
'FaceColor', myColors(mod(i,7)+1, :));
title(sprintf('%s 分布', sensor_names{i}), 'FontSize', 10, 'Interpreter', 'none');
grid on;
box off;
end
sgtitle('传感器分布 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
传感器随周期变化
% 传感器随周期变化
figure('Position', [100, 100, 1800, 800]);
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
scatter(train_FD001.(sensor_names{i}), train_FD001.cycle, 8, ...
train_FD001.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 随周期变化', sensor_names{i}), 'FontSize', 10, 'Interpreter', 'none');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('传感器读数随周期变化 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
操作设置与RUL（剩余寿命）的关系
%% 计算RUL (FD001)
% 计算每个引擎的最大周期
[G, ID_groups] = findgroups(train_FD001.ID);
max_cycle = splitapply(@max, train_FD001.cycle, G);
max_cycle_per_engine = table(ID_groups, max_cycle, 'VariableNames', {'ID', 'max_cycle'});
% 合并数据并计算RUL
train_FD001_merged = join(train_FD001, max_cycle_per_engine);
train_FD001_merged.RUL = train_FD001_merged.max_cycle - train_FD001_merged.cycle;
%% RUL分析 (FD001)
% 操作设置与RUL的关系
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
scatter(train_FD001_merged.(op_set_names{i}), train_FD001_merged.RUL, 8, ...
train_FD001_merged.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 与RUL关系', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i}, 'Interpreter', 'none');
ylabel('RUL');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('操作设置与RUL关系 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
传感器与RUL的关系
% 传感器与RUL的关系
figure('Position', [100, 100, 1800, 800]);
colormap(jet); % 使用更清晰的配色（如 parula, jet, hsv）
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
scatter(train_FD001_merged.(sensor_names{i}), train_FD001_merged.RUL, 8, ...
train_FD001_merged.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sensor_names{i}, 'FontSize', 10, 'Interpreter', 'none');
grid on;
end
sgtitle('传感器读数与RUL关系 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
随机选择5个引擎进行趋势分析
操作设置随RUL变化趋势
engine_ids = randperm(100, 5); % 随机选择5个引擎
% 操作设置趋势分析
figure('Position', [100, 100, 1400, 600]);
tiledlayout(3, 5, 'TileSpacing', 'compact');
for i = 1:3
for j = 1:5
nexttile;
engine_data = train_FD001_merged(train_FD001_merged.ID == engine_ids(j), :);
plot(engine_data.RUL, engine_data.(op_set_names{i}), ...
'Color', myColors(i, :), 'LineWidth', 1.8);
title(sprintf('引擎 %d: %s', engine_ids(j), op_set_names{i}), 'Interpreter', 'none');
xlabel('RUL', 'Interpreter', 'none');
grid on;
end
end
sgtitle('操作设置随RUL变化趋势 (随机引擎)', 'FontSize', 16, 'FontWeight', 'bold');
传感器随RUL变化趋势
sensor_subset = sensor_names(1:6);
figure('Position', [100, 100, 1400, 800]);
tiledlayout(6, 5, 'TileSpacing', 'compact');
for i = 1:6
for j = 1:5
nexttile;
engine_data = train_FD001_merged(train_FD001_merged.ID == engine_ids(j), :);
plot(engine_data.RUL, engine_data.(sensor_subset{i}), ...
'Color', myColors(mod(i,7)+1, :), 'LineWidth', 1.8);
title(sprintf('引擎 %d: %s', engine_ids(j), sensor_subset{i}), 'Interpreter', 'none');
xlabel('RUL');
grid on;
end
end
sgtitle('传感器读数随RUL变化趋势 (随机引擎)', 'FontSize', 16, 'FontWeight', 'bold');
三、完整版代码直接复制
%% 初始化环境
clear; clc; close all;
addpath(genpath(pwd)); % 添加当前目录到路径
% 定义读取数据的函数
function data = get_data(file_name)
file_path = fullfile('C-MAPSS-Data\', file_name);
if exist(file_path, 'file') == 2
data = readtable(file_path, 'Delimiter', ' ', 'MultipleDelimsAsOne', true, ...
'ReadVariableNames', false,'HeaderLines', 0);
else
error('文件未找到: %s', file_path);
end
end
%% 加载所有数据集
% 训练数据
train_FD001 = get_data('train_FD001.txt');
train_FD002 = get_data('train_FD002.txt');
train_FD003 = get_data('train_FD003.txt');
train_FD004 = get_data('train_FD004.txt');
% RUL数据
RUL_FD001 = get_data('RUL_FD001.txt');
RUL_FD002 = get_data('RUL_FD002.txt');
RUL_FD003 = get_data('RUL_FD003.txt');
RUL_FD004 = get_data('RUL_FD004.txt');
% 测试数据
test_FD001 = get_data('test_FD001.txt');
test_FD002 = get_data('test_FD002.txt');
test_FD003 = get_data('test_FD003.txt');
test_FD004 = get_data('test_FD004.txt');
%% 设置列名
col_names = {'ID', 'cycle', 'op_set_1', 'op_set_2', 'op_set_3', ...
'sensor_1', 'sensor_2', 'sensor_3', 'sensor_4', 'sensor_5', ...
'sensor_6', 'sensor_7', 'sensor_8', 'sensor_9', 'sensor_10', ...
'sensor_11', 'sensor_12', 'sensor_13', 'sensor_14', 'sensor_15', ...
'sensor_16', 'sensor_17', 'sensor_18', 'sensor_19', 'sensor_20', ...
'sensor_21'};
col_name_RUL = {'RUL'};
% 为所有数据集分配列名
datasets = {train_FD001, train_FD002, train_FD003, train_FD004, ...
test_FD001, test_FD002, test_FD003, test_FD004, ...
RUL_FD001, RUL_FD002, RUL_FD003, RUL_FD004};
col_sets = {repmat({col_names}, 1, 8), repmat({col_name_RUL}, 1, 4)};
for i = 1:12
if i <= 8
datasets{i}.Properties.VariableNames = col_names;
else
datasets{i}.Properties.VariableNames = col_name_RUL;
end
end
% 将更新后的数据集赋值回原始变量
train_FD001 = datasets{1};
train_FD002 = datasets{2};
train_FD003 = datasets{3};
train_FD004 = datasets{4};
test_FD001 = datasets{5};
test_FD002 = datasets{6};
test_FD003 = datasets{7};
test_FD004 = datasets{8};
RUL_FD001 = datasets{9};
RUL_FD002 = datasets{10};
RUL_FD003 = datasets{11};
RUL_FD004 = datasets{12};
%% 数据质量检查
% 检查空值
fprintf('空值检查:\n');
for i = 1:8
null_count = sum(any(ismissing(datasets{i}), 1));
fprintf('数据集 %d: %d 个空值\n', i, null_count);
end
% 检查重复值
fprintf('\n重复值检查:\n');
for i = 1:8
[~, idx] = unique(datasets{i}, 'rows');
dup_count = size(datasets{i}, 1) - length(idx);
fprintf('数据集 %d: %d 个重复值\n', i, dup_count);
end
%% 创建统一绘图样式
myColors = [0 0.4470 0.7410; 0.8500 0.3250 0.0980; 0.9290 0.6940 0.1250; ...
0.4940 0.1840 0.5560; 0.4660 0.6740 0.1880; 0.3010 0.7450 0.9330; ...
0.6350 0.0780 0.1840];
set(groot, 'defaultFigureColor', 'white', ...
'defaultAxesFontSize', 12, ...
'defaultAxesLabelFontSizeMultiplier', 1.1, ...
'defaultAxesTitleFontSizeMultiplier', 1.2, ...
'defaultLineLineWidth', 1.5);
%% 操作设置分析 (FD001)
% 提取操作设置列
op_set_cols = contains(col_names, 'op_set');
op_set_names = col_names(op_set_cols);
% 直方图分布
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
data = train_FD001.(op_set_names{i});
histogram(data, 'Normalization', 'probability', ...
'FaceColor', myColors(i, :));
title(sprintf('%s 分布', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i},'Interpreter', 'none');
ylabel('相对频率');
grid on;
box off;
if i==3
xlim([80, 120]); % 手动设置 x 轴范围
end
end
sgtitle('操作设置分布 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
% 操作设置随周期变化
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
scatter(train_FD001.(op_set_names{i}), train_FD001.cycle, 8, ...
train_FD001.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 随周期变化', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i}, 'Interpreter', 'none');
ylabel('周期');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('操作设置随周期变化 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
%% 传感器分析 (FD001)
% 提取传感器列
sensor_cols = contains(col_names, 'sensor');
sensor_names = col_names(sensor_cols);
% 传感器直方图
figure('Position', [100, 100, 1400, 800]);
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
data = train_FD001.(sensor_names{i});
histogram(data, 'Normalization', 'probability', ...
'FaceColor', myColors(mod(i,7)+1, :));
title(sprintf('%s 分布', sensor_names{i}), 'FontSize', 10, 'Interpreter', 'none');
grid on;
box off;
end
sgtitle('传感器分布 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
% 传感器随周期变化
figure('Position', [100, 100, 1800, 800]);
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
scatter(train_FD001.(sensor_names{i}), train_FD001.cycle, 8, ...
train_FD001.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 随周期变化', sensor_names{i}), 'FontSize', 10, 'Interpreter', 'none');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('传感器读数随周期变化 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
%% 计算RUL (FD001)
% 计算每个引擎的最大周期
[G, ID_groups] = findgroups(train_FD001.ID);
max_cycle = splitapply(@max, train_FD001.cycle, G);
max_cycle_per_engine = table(ID_groups, max_cycle, 'VariableNames', {'ID', 'max_cycle'});
% 合并数据并计算RUL
train_FD001_merged = join(train_FD001, max_cycle_per_engine);
train_FD001_merged.RUL = train_FD001_merged.max_cycle - train_FD001_merged.cycle;
%% RUL分析 (FD001)
% 操作设置与RUL的关系
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
scatter(train_FD001_merged.(op_set_names{i}), train_FD001_merged.RUL, 8, ...
train_FD001_merged.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 与RUL关系', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i}, 'Interpreter', 'none');
ylabel('RUL');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('操作设置与RUL关系 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
% 传感器与RUL的关系
figure('Position', [100, 100, 1800, 800]);
colormap(jet); % 使用更清晰的配色（如 parula, jet, hsv）
tiledlayout(3, 7, 'TileSpacing', 'compact');
for i = 1:21
nexttile;
scatter(train_FD001_merged.(sensor_names{i}), train_FD001_merged.RUL, 8, ...
train_FD001_merged.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sensor_names{i}, 'FontSize', 10, 'Interpreter', 'none');
grid on;
end
sgtitle('传感器读数与RUL关系 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
%% 相关性分析 (FD001)
% 提取相关列
corr_cols = contains(train_FD001_merged.Properties.VariableNames, ...
{'op_set', 'sensor', 'RUL'}) | ...
strcmp(train_FD001_merged.Properties.VariableNames, 'cycle');
corr_data = train_FD001_merged(:, corr_cols);
% 计算并显示相关系数矩阵
corr_matrix = corr(table2array(corr_data), 'Rows', 'complete');
figure('Position', [100, 100, 1000, 900]);
imagesc(corr_matrix);
colorbar;
% colormap(jet);
title('特征相关性矩阵 (FD001)', 'FontSize', 16, 'FontWeight', 'bold');
axis square;
%% 随机选择5个引擎进行趋势分析
engine_ids = randperm(100, 5); % 随机选择5个引擎
% 操作设置趋势分析
figure('Position', [100, 100, 1400, 600]);
tiledlayout(3, 5, 'TileSpacing', 'compact');
for i = 1:3
for j = 1:5
nexttile;
engine_data = train_FD001_merged(train_FD001_merged.ID == engine_ids(j), :);
plot(engine_data.RUL, engine_data.(op_set_names{i}), ...
'Color', myColors(i, :), 'LineWidth', 1.8);
title(sprintf('引擎 %d: %s', engine_ids(j), op_set_names{i}), 'Interpreter', 'none');
xlabel('RUL', 'Interpreter', 'none');
grid on;
end
end
sgtitle('操作设置随RUL变化趋势 (随机引擎)', 'FontSize', 16, 'FontWeight', 'bold');
% 传感器趋势分析 (选择前6个传感器示例)
sensor_subset = sensor_names(1:6);
figure('Position', [100, 100, 1400, 800]);
tiledlayout(6, 5, 'TileSpacing', 'compact');
for i = 1:6
for j = 1:5
nexttile;
engine_data = train_FD001_merged(train_FD001_merged.ID == engine_ids(j), :);
plot(engine_data.RUL, engine_data.(sensor_subset{i}), ...
'Color', myColors(mod(i,7)+1, :), 'LineWidth', 1.8);
title(sprintf('引擎 %d: %s', engine_ids(j), sensor_subset{i}), 'Interpreter', 'none');
xlabel('RUL');
grid on;
end
end
sgtitle('传感器读数随RUL变化趋势 (随机引擎)', 'FontSize', 16, 'FontWeight', 'bold');
%% 处理FD002数据集 (示例)
[G, ID_groups] = findgroups(train_FD002.ID);
max_cycle = splitapply(@max, train_FD002.cycle, G);
max_cycle_per_engine = table(ID_groups, max_cycle, 'VariableNames', {'ID', 'max_cycle'});
train_FD002_merged = join(train_FD002, max_cycle_per_engine);
train_FD002_merged.RUL = train_FD002_merged.max_cycle - train_FD002_merged.cycle;
% 操作设置与RUL关系
figure('Position', [100, 100, 1200, 500]);
for i = 1:3
subplot(1, 3, i);
scatter(train_FD002_merged.(op_set_names{i}), train_FD002_merged.RUL, 8, ...
train_FD002_merged.ID, 'filled', 'MarkerEdgeAlpha', 0.3, 'MarkerFaceAlpha', 0.3);
title(sprintf('%s 与RUL关系 (FD002)', op_set_names{i}), 'FontWeight', 'bold', 'Interpreter', 'none');
xlabel(op_set_names{i}, 'Interpreter', 'none');
ylabel('RUL');
colormap(turbo);
colorbar('off');
grid on;
end
sgtitle('操作设置与RUL关系 (FD002)', 'FontSize', 16, 'FontWeight', 'bold');四、代码运行简单说明
从NASA官方直接下载数据：https://data.nasa.gov/dataset/cmapss-jet-engine-simulated-data
复制以上代码即可直接一键出图。
获取更多优秀代码
或者复制链接跳转：https://docs.qq.com/sheet/DU3NjYkF5TWdFUnpu端午安康
扫一扫
1992
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
