金龍哲,王夢(mèng)飛,于?露,徐明偉
?
基于PPG信號(hào)的有限空間低氧傷害評(píng)估與預(yù)警
金龍哲,王夢(mèng)飛,于?露,徐明偉
(北京科技大學(xué)土木與資源工程學(xué)院,北京 100083)
工業(yè)化迅速發(fā)展使得有限空間作業(yè)增多,有限空間作業(yè)人員在極限低氧環(huán)境(環(huán)境氧氣濃度低于15%)下易受到低氧傷害,傳統(tǒng)的環(huán)境氧濃度監(jiān)測(cè)預(yù)警未考慮作業(yè)人員個(gè)體生理狀態(tài)的差異性.本文針對(duì)作業(yè)人員生理狀態(tài)的不同提出一種快速準(zhǔn)確的特異性低氧傷害評(píng)估預(yù)警方法.通過(guò)智能穿戴設(shè)備監(jiān)測(cè)正常氧濃度(20%~21%)和人體可耐受最低氧濃度(15%~16%)情況下的人體光電容積脈搏波(photoplethysmography)信號(hào),采用振幅和基波頻率差異作為依據(jù),對(duì)作業(yè)人員進(jìn)行快速準(zhǔn)確實(shí)時(shí)的低氧傷害評(píng)估及預(yù)警.本文在信號(hào)處理過(guò)程中利用小波變換和移動(dòng)平均濾波器去除基線漂移和高頻噪聲,獲得高信噪比的信號(hào)并進(jìn)行時(shí)域分析,繼而通過(guò)快速傅里葉變換將處理后的高信噪比信號(hào)進(jìn)行轉(zhuǎn)換獲得頻譜圖進(jìn)行頻域分析.時(shí)域分析與頻域分析的結(jié)果表明:低氧環(huán)境中,人體光電容積脈搏波信號(hào)的振幅和基波頻率較正常氧濃度情況下均存在顯著性差異(<0.05),下降率的平均值分別為46%和19%;利用極大似然估計(jì)法和3原則計(jì)算下降率的臨界值,確定振幅和基波頻率下降率的臨界值分別為29%和8.6%,該臨界值的理論計(jì)算準(zhǔn)確度為84.1%;利用計(jì)算所得的臨界值建立有限空間低氧傷害評(píng)估預(yù)警表.該評(píng)估表的理論準(zhǔn)確度為97.5%,經(jīng)實(shí)驗(yàn)驗(yàn)證其預(yù)警成功率為91%且無(wú)漏報(bào)情況發(fā)生.
低氧傷害;光電容積脈搏波;有限空間;評(píng)估預(yù)警;時(shí)域分析;頻域分析
隨著工業(yè)化生產(chǎn)的高速發(fā)展,有限空間事故高發(fā)、頻發(fā),且事故后果嚴(yán)重[1],救援難度大,對(duì)工業(yè)安全生產(chǎn)、市政建設(shè)管理、建筑施工等行業(yè)安全造成了極大威脅.劉艷等[2]統(tǒng)計(jì)2013—2018年全國(guó)共發(fā)生有限空間較大事故67起,造成269人死亡.
有限空間最主要風(fēng)險(xiǎn)是空間氧氣濃度處于不安全水平.美國(guó)國(guó)立職業(yè)安全與健康研究所的研究表明由氧氣濃度過(guò)低引起的有限空間死亡事件占35%,在所有風(fēng)險(xiǎn)中占比最高[3].一般認(rèn)為氧氣濃度的安全范圍是18.5%~23.0%(體積分?jǐn)?shù)).然而最新的證據(jù)顯示,15.5%的氧氣濃度也能保證人員的基本生命安全[4].當(dāng)氧氣濃度低于15%時(shí),人體出現(xiàn)動(dòng)作不協(xié)調(diào)現(xiàn)象,逐漸失去知覺(jué).人體短暫進(jìn)入低于15%的環(huán)境即會(huì)造成身體不適或行為失控.
作業(yè)人員在進(jìn)入極限低氧環(huán)境(氧氣濃度低于15%)作業(yè)時(shí),除窒息死亡事故外[5],作業(yè)人員的健康也受到慢性環(huán)境性缺氧的損害.慢性缺氧對(duì)生理的損害主要反映在心肌代謝和腦損傷中,較長(zhǎng)時(shí)間缺氧將導(dǎo)致心肌的損傷壞死[6],感知能力及思維能力降低,部分免疫細(xì)胞的功能受到抑制[7],這些損害是導(dǎo)致職業(yè)病發(fā)生的潛在因素.
目前國(guó)內(nèi)外學(xué)者針對(duì)降低有限空間低氧傷害事故發(fā)生機(jī)率的研究思路主要集中在提高環(huán)境氧氣濃度監(jiān)測(cè)可靠性與準(zhǔn)確性上.此思路主要存在2個(gè)弊端:一是忽視了對(duì)生產(chǎn)過(guò)程中人體生理安全臨界狀態(tài)的判斷;二是未考慮作業(yè)人員的個(gè)體差異,對(duì)作業(yè)人員實(shí)時(shí)的生理狀態(tài)無(wú)法進(jìn)行差異化評(píng)估預(yù)警.
近年來(lái),可穿戴式生理電信號(hào)設(shè)備和非線性動(dòng)力學(xué)的快速發(fā)展使人體生理狀態(tài)快速無(wú)創(chuàng)測(cè)量成為可能[8].光電容積脈搏波[9]是人體最基本的生理信號(hào),蘊(yùn)含心臟搏動(dòng)、血液流量、呼吸振動(dòng)、自主神經(jīng)活動(dòng)等生理信息[10-11].汪雄等[12]通過(guò)監(jiān)測(cè)PPG信號(hào),準(zhǔn)確計(jì)算出足球運(yùn)動(dòng)員的賽時(shí)心率;吳育東等[13]提出了一種基于PPG技術(shù)的無(wú)創(chuàng)血壓測(cè)量方法,PPG技術(shù)已廣泛應(yīng)用于各領(lǐng)域[14].
本文監(jiān)測(cè)人體處于低氧環(huán)境(15%~16%)中的PPG生理信號(hào),通過(guò)分析信號(hào)的特征值,確定正常氧濃度環(huán)境與低氧環(huán)境下人體的PPG信號(hào)特征值變化率的臨界值,以該值作為評(píng)估參數(shù)提出一種新的低氧傷害預(yù)警方法,并對(duì)該方法的準(zhǔn)確度進(jìn)行檢驗(yàn)分析,為有限空間低氧環(huán)境下作業(yè)人員安全生產(chǎn)提供理論依據(jù).
實(shí)驗(yàn)場(chǎng)所為北京科技大學(xué)自主研發(fā)設(shè)計(jì)的礦用緊急避險(xiǎn)救生艙,屬于典型的有限空間.該救生艙體積為12m3,設(shè)計(jì)有1個(gè)進(jìn)入口和1個(gè)緊急逃生口,空間體積狹小,密閉性好,空間環(huán)境氣體濃度可調(diào)節(jié)控制.
本次實(shí)驗(yàn)被試通過(guò)公開(kāi)招募的方式確定,為北京科技大學(xué)的22名在校大學(xué)生,其中11名男性被試(平均年齡22歲),11名女性被試(平均年齡24歲).所有被試均健康狀況良好,無(wú)心血管或神經(jīng)方面疾病,無(wú)抽煙、長(zhǎng)期飲用咖啡或服用藥物的習(xí)慣,實(shí)驗(yàn)前10日要求無(wú)飲酒記錄.本實(shí)驗(yàn)通過(guò)了北京科技大學(xué)倫理道德委員會(huì)的批準(zhǔn).
1.2.1?救生艙密閉性檢驗(yàn)實(shí)驗(yàn)
在開(kāi)始正式實(shí)驗(yàn)之前,首先對(duì)救生艙密閉性進(jìn)行檢測(cè),以確保正式實(shí)驗(yàn)中低氧環(huán)境氛圍營(yíng)造的可操作性.采用釋放高壓普通氮?dú)庵脫Q空間氣體的方式營(yíng)造低氧環(huán)境.1h內(nèi)釋放1MPa下5cm3氮?dú)猓?h后停止釋放氮?dú)猓?0min內(nèi)空間氧濃度維持在15%~16%.認(rèn)為救生艙密閉性良好,能保證營(yíng)造低氧環(huán)境.實(shí)驗(yàn)結(jié)果如圖1所示.
1.2.2?有限空間低氧環(huán)境實(shí)驗(yàn)
共開(kāi)展22組實(shí)驗(yàn),每位被試進(jìn)行1組實(shí)驗(yàn)(1×22).實(shí)驗(yàn)時(shí)間設(shè)置在每天的14:00—16:00,該時(shí)間段內(nèi)人體氧攝取最為活躍[15].實(shí)驗(yàn)設(shè)置正常氧氣濃度水平為20%~21%,設(shè)置低氧環(huán)境氧氣濃度為15%~16%.每次開(kāi)始實(shí)驗(yàn)前,監(jiān)測(cè)被試心率、血壓與核心溫度,保證生理狀態(tài)正常后,被試人員進(jìn)入有限空間救生艙,測(cè)量正常氧氣濃度狀態(tài)下人體指端PPG數(shù)據(jù),測(cè)量結(jié)束后,封閉艙門(mén),向救生艙內(nèi)緩慢釋放氮?dú)猓?dāng)空間氧氣濃度達(dá)到低氧水平時(shí),再次測(cè)量被試PPG信號(hào)數(shù)據(jù),2組測(cè)量均要求被試手部保持靜止,以免對(duì)數(shù)據(jù)造成偽跡干擾,信號(hào)監(jiān)測(cè)頻率為500Hz,測(cè)量時(shí)長(zhǎng)為2min,保證每段PPG數(shù)據(jù)長(zhǎng)度為300±15.PPG測(cè)量方法如圖2所示.
圖1 ?氧氣濃度變化情況
圖2? PPG信號(hào)獲取
原始的PPG信號(hào)由于易受到外界因素(如運(yùn)動(dòng)、咳嗽、深呼吸、電磁等)的干擾,存在著明顯的基線漂移、高頻干擾以及運(yùn)動(dòng)偽跡等現(xiàn)象,因此需要對(duì)原始數(shù)據(jù)進(jìn)行預(yù)處理來(lái)保證特征值提取的準(zhǔn)確性.信號(hào)預(yù)處理應(yīng)盡量保留較多的原始信號(hào)信息.據(jù)此,本文利用小波變換對(duì)原始數(shù)據(jù)進(jìn)行預(yù)處理,處理步驟如下.
步驟1去除基線漂移.首先進(jìn)行的預(yù)處理為去除PPG信號(hào)的基線漂移.以db5為小波基對(duì)原始信號(hào)進(jìn)行10級(jí)分解獲取小波系數(shù),基線漂移屬于低頻干擾,因而以第8級(jí)近似分量重建信號(hào)[16],獲得信號(hào)的基線趨勢(shì),最后將原始信號(hào)減去信號(hào)的基線趨勢(shì)即可去除基線漂移.
步驟2去除噪聲信號(hào).將去除基線漂移的信號(hào)做進(jìn)一步處理,利用小波變換rigrsure閾值規(guī)則對(duì)信號(hào)進(jìn)行去噪,該閾值規(guī)則使用無(wú)偏風(fēng)險(xiǎn)估計(jì);設(shè)置閾值方法為soft閾值;閾值尺度的調(diào)整方法為sln;離散小波變換的級(jí)數(shù)為5;小波基為db5[17-18].
步驟3信號(hào)平滑處理.采用移動(dòng)平均濾波器對(duì)信號(hào)進(jìn)行平滑處理,濾波器的窗寬設(shè)置為10.
經(jīng)過(guò)信號(hào)預(yù)處理后得到適合進(jìn)行特征參數(shù)提取的高信噪比信號(hào).預(yù)處理過(guò)程及效果如圖3所示.
圖3? PPG信號(hào)預(yù)處理
人體低氧傷害預(yù)警需要對(duì)比分析PPG信號(hào)在正常氧濃度情況和低氧濃度情況的差異.本文選取PPG信號(hào)的振幅和基波頻率作為特征參數(shù).
(1) 振幅.PPG信號(hào)的振幅在一定程度上能反映心臟的泵血能力以及供輸氧情況,當(dāng)人體所處的氧環(huán)境發(fā)生變化時(shí),PPG信號(hào)的振幅也會(huì)發(fā)生相應(yīng)變化.利用MATLAB中的findpeaks函數(shù)對(duì)預(yù)處理后的PPG信號(hào)尋找全部波的峰值和谷值的索引值[19],并計(jì)算全部峰值和相鄰谷值的差值,將所有差值利用mean函數(shù)進(jìn)行平均計(jì)算,作為該段信號(hào)的振幅值,提取結(jié)果示例如圖4所示(以1次實(shí)驗(yàn)為例).
(2) 頻率.波的時(shí)域分析具有局限性,為了獲取更多的評(píng)估信息,對(duì)PPG信號(hào)進(jìn)行進(jìn)一步的特征提取.將PPG信號(hào)通過(guò)快速傅里葉變換(FFT)[20]轉(zhuǎn)換到頻域,分析其頻率的變化.利用MATLAB中的max函數(shù)尋找信號(hào)的能量最大值,以最大能量的1/2作為截取標(biāo)準(zhǔn),利用find函數(shù)獲取所有大于該能量的信號(hào),并利用mean函數(shù)計(jì)算所獲信號(hào)頻率的平均值,以此平均值作為基波頻率.頻率提取過(guò)程見(jiàn)圖5.
圖4 ?光電容積脈搏波振幅
圖5 ?光電容積脈搏波頻譜圖
隨機(jī)抽取22組實(shí)驗(yàn)數(shù)據(jù)中的11組數(shù)據(jù),分別計(jì)算11位被試正常氧濃度和低氧情況下PPG信號(hào)的平均振幅.根據(jù)PPG信號(hào)的測(cè)量原理,PPG信號(hào)振幅所代表的物理意義為血管容積內(nèi)的血液流量變化量.平均振幅的計(jì)算公式為
=∑(x-y) (1)
式中:x表示為第個(gè)光電容積脈搏波的波峰值;y表示第個(gè)光電容積脈搏波的波谷值;表示光電容積脈搏波的平均振幅.每個(gè)被試PPG信號(hào)平均振幅的計(jì)算包含300個(gè)以上完整波形,因而具有較高的可?靠性.
正常氧濃度與低氧情況下平均振幅對(duì)比見(jiàn)圖6.通過(guò)對(duì)平均振幅分析可知,人體在有限空間低氧環(huán)境中PPG信號(hào)的振幅較正常氧濃度狀態(tài)顯著下降,且大部分被試下降幅度較大,11個(gè)被試下降率的平均值為46%,中位值為47%.為確定這種變化是由氧濃度變化引起,排除抽樣誤差,本文進(jìn)一步對(duì)振幅變化的差異性進(jìn)行T檢驗(yàn)和F檢驗(yàn).PPG的振幅服從近似正態(tài)分布,由于樣本數(shù)目<30,經(jīng)計(jì)算F值為1.19,小于置信度為95%時(shí)的F值2.97,證明2組數(shù)據(jù)方差沒(méi)有顯著差異.對(duì)樣本進(jìn)行相關(guān)系數(shù)的計(jì)算,相關(guān)系數(shù)=0.6,利用相關(guān)樣本T檢驗(yàn)公式計(jì)算,=4.99>T(10)0.05,因此認(rèn)為2樣本均值在置信度為95%水平差異顯著.通過(guò)極大似然估計(jì)可知總體參數(shù),=0.46,2=0.0291.根據(jù)3原則所得分布概率情況見(jiàn)表1,為保證數(shù)值的區(qū)分度以及在工程使用中的辨識(shí)度,選擇使用29%為平均振幅變化率臨界值下限,該臨界值的理論正確率可達(dá)到84.1%.在實(shí)際生產(chǎn)中,當(dāng)人體PPG信號(hào)平均振幅變化率大于此臨界值時(shí),認(rèn)為環(huán)境氧氣濃度下降到人體可承受的最低氧氣濃度值.
圖6 ?平均振幅對(duì)比
表1 ?振幅變化數(shù)值分布概率
Tab.1 Numerical distribution probability of amplitude variation
利用第2.2節(jié)中介紹的快速傅里葉變換方法將全部11個(gè)被試的正常氧濃度與低氧情況的PPG信號(hào)由時(shí)域轉(zhuǎn)換到頻域進(jìn)行分析.統(tǒng)計(jì)結(jié)果對(duì)比如圖7所示,在圖中能夠看出,低氧情況下人體PPG信號(hào)的基波頻率低于正常氧濃度情況下的基波頻率,所有被試變化率的平均值為19%,中位值18%.PPG信號(hào)基波頻率降低代表波的周期變長(zhǎng),由于人體PPG信號(hào)的周期與心臟搏動(dòng)的周期直接相關(guān),因而PPG信號(hào)周期的延長(zhǎng)意味心臟搏動(dòng)周期變長(zhǎng),也就是說(shuō),通過(guò)本文實(shí)驗(yàn)證明,當(dāng)人體處于有限空間極端低氧環(huán)境中,人體心臟搏動(dòng)周期變長(zhǎng).對(duì)統(tǒng)計(jì)結(jié)果進(jìn)行F檢驗(yàn),檢驗(yàn)結(jié)果為6.15,在置信度為95%的水平上,數(shù)據(jù)存在顯著差異.通過(guò)極大似然估計(jì)可知,=0.185,2=0.0099.根據(jù)3原則所得概率分布情況見(jiàn)表2,使用8.6%為PPG信號(hào)基波頻率變化率臨界值下限,該臨界值的理論正確率可達(dá)到84.1%,當(dāng)人體PPG信號(hào)平均基波頻率變化率大于此臨界值時(shí),認(rèn)為環(huán)境氧氣濃度下降到人體可承受最低氧氣濃度值.
圖7? 基波頻率對(duì)比
表2 ?頻率變化數(shù)值分布概率
Tab.2 Numerical distribution probability of frequency change
上文的分析表明,在有限空間低氧環(huán)境中,人體PPG信號(hào)的振幅與基波頻率較正常氧濃度狀態(tài)都存在顯著變化,并利用極大似然估計(jì)法求得了2個(gè)特征值變化率臨界值的下限.因此,本文依據(jù)PPG信號(hào)的振幅和頻率建立了低氧傷害評(píng)估預(yù)警表,如表3所示,該評(píng)估預(yù)警表的理論正確率是97.5%.
表3 ?低氧傷害評(píng)估預(yù)警
Tab.3? Early warning form for hypoxic injury assessment
為了檢驗(yàn)該評(píng)估表在實(shí)際操作中的準(zhǔn)確性,進(jìn)一步抽取數(shù)據(jù)對(duì)表3進(jìn)行驗(yàn)證,抽取11組已知氧氣濃度情況的PPG信號(hào),利用表3進(jìn)行評(píng)估預(yù)警,驗(yàn)證結(jié)果見(jiàn)表4,表4中“是否瀕臨傷害”指人體是否進(jìn)入到人體可耐受最低氧濃度環(huán)境.在進(jìn)行的11次驗(yàn)證中,有10次評(píng)估正確,1次評(píng)估錯(cuò)誤,成功率為91%且無(wú)漏報(bào)情況,該成功率較高,因此認(rèn)為低氧傷害評(píng)估預(yù)警表為切實(shí)可行的.
表4 ?低氧傷害評(píng)估驗(yàn)證
Tab.4 ?Hypoxic injury assessment verification form
通過(guò)救生艙低氧環(huán)境實(shí)驗(yàn),監(jiān)測(cè)人體在正常氧(20%~21%)環(huán)境狀態(tài)下和低氧(15%~16%)環(huán)境狀態(tài)下的光電容積脈搏波.利用小波分析和FFT對(duì)2種狀態(tài)下PPG信號(hào)進(jìn)行處理,利用統(tǒng)計(jì)學(xué)方法進(jìn)行分析,具體得到以下結(jié)論.
(1) 傳統(tǒng)PPG信號(hào)處理中一般認(rèn)為振幅絕對(duì)值經(jīng)放大處理后失真,忽略其蘊(yùn)含的信息,本文分析發(fā)現(xiàn)振幅變化率可以有效反映人體進(jìn)入低氧環(huán)境時(shí)生理狀態(tài)變化.
(2) 人體進(jìn)入到低氧環(huán)境中時(shí)生理狀態(tài)會(huì)發(fā)生相應(yīng)的變化,人體PPG信號(hào)可以反映這種變化,在低氧環(huán)境條件下,人體的PPG信號(hào)振幅和基波頻率下降,變化率的平均值分別為46%和19%.
(3) 低氧環(huán)境導(dǎo)致的人體PPG信號(hào)振幅變化率和基波頻率變化率的臨界值分別為29%和8.6%,超過(guò)此臨界值時(shí),認(rèn)為人體所處環(huán)境的氧氣濃度已降低至人體可承受的最低氧氣濃度,即人員能保證基本生命安全的極限氧氣濃度,該臨界值利用極大似然估計(jì)法和3原則計(jì)算得出,理論準(zhǔn)確率為84.1%.
(4) 根據(jù)臨界值建立低氧傷害評(píng)估預(yù)警表,解決了作業(yè)人員實(shí)時(shí)生理狀態(tài)差異化評(píng)估預(yù)警的問(wèn)題,利用實(shí)驗(yàn)數(shù)據(jù)進(jìn)行驗(yàn)證,正確評(píng)估預(yù)警的概率為91%,為下一步智能穿戴裝備監(jiān)測(cè)預(yù)警提供理論基礎(chǔ).
[1] 臧鐵柱,張禮敬,張?麗,等. 有限空間作業(yè)意外傷之事故的成因及其對(duì)策[J]. 南京工業(yè)大學(xué)學(xué)報(bào),2005,27(3):103-106.
Zang Tiezhu,Zhang Lijing,Zhang Li,et al. Causes and countermeasures of casualty accident induced by unexpected factors in limited job space[J]. Journal of Nanjing University of Science and Technology,2005,27(3):103-106(in Chinese).
[2] 劉?艷,秦?研. 有限空間作業(yè)安全[M]. 北京:團(tuán)結(jié)出版社,2016.
Liu Yan,Qin Yan. Safety of Limited Space Operations [M]. Beijing:Unity Press,2016(in Chinese).
[3] Safety I. Worker deaths in confined spaces:A summary of NIOSH surveillance and investigative findings[J]. Industrial Safety,1994:94-103.
[4] Kamal A A R,Harness J B,Irving G,et al. Skin photoplethysmography—A review[J]. Computer Methods Programs Biomedicine,1989,28:257-269.
[5] Jason S,Jeffrey S,Janis J,et al. An investigation into the rate and mechanism of incident of work-related con?ned space fatalities[J]. Safety Science,2018,109:333-343.
[6] 陳亞妮,王延琦,萬(wàn)?紅,等. 高原高寒環(huán)境對(duì)人體功能的影響及預(yù)防[J]. 職業(yè)與健康,2014,30(10):1409-1412.
Chen Yani,Wang Yanqi,Wan Hong,et al. Effects of alpine environment on human function and prevention[J]. Occupation and Health,2014,30(10):1409-1412(in Chinese).
[7] Kokura S,Wolf R E,Yoshikawa T,et al. T-lymphocyte-derived tumor necrosis factor exacerbates anoxia-reoxygenation-induced neutrophil-endothelial cell adhesion[J]. Circulation Research,2000,86(2):205-213.
[8] Tamura T,Maeda Y,Sekine M,et al. Wearable Photoplethysmographic sensors—Past and present[J]. Electronics,2014,3(2):282-302.
[9] Giltvedt J,Sira A,Helme P. Pulsd multifrequency photoplethysmogroph[J]. Med Biol Eng Comput,1984,22(3):212-215.
[10] Avolio A P,Butlin M,Walsh A. Arterial blood pressure measurement and pulse wave analysis-their role in enhancing cardiovascular assessment[J]. Physiological Measurement,2010,31(1):1-47.
[11] Gil E,Orini M,Bailon R,et al. Photoplethysmography pulse rate variability as a surrogate measurement of heart rate variability during nonstationary conditions[J]. Physiol Meas,2010,31(9):1271-1290.
[12] 汪?雄,陳玉林,張恒亮,等. 光電容積脈搏波描記法(PPG)測(cè)試足球運(yùn)動(dòng)員賽時(shí)心率的準(zhǔn)確性研究[J]. 電子世界,2018(17):58,60.
Wang Xiong,Chen Yulin,Zhang Hengliang,et al. Accuracy of photoelectric volume pulse wave graph (PPG)in measuring the heart rate of football players in matches[J]. Electronic World,2018(17):58,60(in Chinese).
[13] 吳育東,鐘舜聰,伏喜斌. 基于光電容積脈搏波的血壓測(cè)量實(shí)驗(yàn)研究[J]. 機(jī)電工程,2017,34(8):865-869.
Wu Yudong,Zhong Shunchong,F(xiàn)u Xibin. Experimental study of blood pressure measurement based on photoelectric volume pulse wave[J]. Mechatronics Engineering,2017,34(8):865-869(in Chinese).
[14] Allen J. Photoplethysmography and its application in clinical physiological measurement[J]. Physiological Measurement,2007,28(3):1-39.
[15] Craig A,Tran Y,Wijesuriya N,et al. A controlled investigation into the psychological determinants of fatigue[J]. Biological Psychology,2006,72(1):78-87.
[16] 張文瓊,劉肖琳,吳?濤. 一種利用小波變換逼近信號(hào)濾除心電圖基線漂移的方法[J]. 計(jì)算機(jī)工程與應(yīng)用,2005(20):222-224.
Zhang Wenqiong,Liu Xiaolin,Wu Tao. A method of eliminating ECG baseline drift by using wavelet transform approximation signal[J]. Computer Engineering and Application,2005(20):222-224(in Chinese).
[17] Zhang Weipeng. Image denoising algorithm of refuge chamber by combining wavelet transform and bilateral filtering[J]. International Journal of Mining Science and Technology,2013,23(2):228-232.
[18] Zhao Yanqin. Multi-level denoising and enhancement method based on wavelet transform for mine monitoring[J]. International Journal of Mining Science and Technology,2013,23(1):163-166.
[19] 李肅義,徐?壯,熊文激,等. 一種光電容積脈搏信號(hào)的峰值點(diǎn)自動(dòng)識(shí)別方法[J]. 光譜學(xué)與光譜分析,2017,37(10):3145-3149.
Li Suyi,Xu Zhuang,Xiong Wenji,et al. An automatic peak recognition method for photoelectric volume pulse signals[J]. Spectroscopy and Spectral Analysis,2017,37(10):3145-3149(in Chinese).
[20] Lin C S,Yang S W,Lin Hung L,et al. Measurement of surface profile and surface roughness of fibre-optic interconnect by fast Fourier transform[J]. Metrology and Measurement Systems,2017,24(2):381-390.
Assessment and Early Warning Regarding Hypoxic Injury in a Confined Space Based on the Photoplethysmography
Jin Longzhe,Wang Mengfei,Yu Lu,Xu Mingwei
(School of Civil and Resource Engineering,University of Science and Technology Beijing,Beijing 100083,China)
The rapid development of industrialization has increased the number of confined space operations. Confined space workers are vulnerable to hypoxic injury under an extreme hypoxic environment,where oxygen concentration is less than 15%.The traditional monitoring and early warning system of environmental oxygen concentration does not consider the differences among the individual physiological status of the operators.In this study,a new rapid and accurate early warning method for specific hypoxic injury assessment was proposed according to the different physiological status of the workers.By monitoring the normal oxygen concentration(20%—21%)and the lowest tolerable oxygen concentration(15%—16%)via intelligent wearable equipment,the photoplethysmography signal of the human body was detected,and the differences of the amplitude and fundamental frequency were used for rapid,accurate,and real-time assessment and early warning of hypoxic injury to workers.Herein,we used the wavelet transform and a moving average filter to remove the baseline drift and high-frequency noise in signal processing,obtain a high signal-to-noise ratio(SNR)signal and analyze it in time domain,and then transform the processed high SNR signal into the frequency domain using a fast Fourier transform(FFT).The results of time-domain analysis and frequency-domain analysis show that there are significant differences in the amplitude and fundamental frequency of the human photoelectric volume pulse wave signal in a hypoxic environment compared with that in an environment with normal oxygen concentration(<0.05).The average decline ratein amplitude and fundamental frequency is 46% and 19%.The critical value of the decline rate is calculated by maximum likelihood estimation method and the principle of 3,the critical points of the decline rates of the amplitude and fundamental frequency are determined as 29% and 8.6%,respectively,and the theoretical accuracy of the critical value is 84.1%.A confined space hypoxic injury assessment and early warning table is established based on the calculated critical value.The theoretical accuracy of the evaluation table is 97.5%.The experimental results show that the accuracy rate of early warning is 91% and that there is no false negative.
hypoxic injury;photoplethysmography(PPG);confined space;evaluation and early warning;time-domain analysis;frequency-domain analysis
X912
A
0493-2137(2019)08-0822-07
10.11784/tdxbz201811069
2018-11-26;
2019-01-13.
金龍哲(1963—??),男,教授.Email:m_bigm@tju.edu.cn
金龍哲,lzjin@ustb.edu.cn.
國(guó)家“十三五”重點(diǎn)科技資助項(xiàng)目(2016YFC080176).
theNational Key Science and Technology Subsidy Project for the 13th Five-Year Plan(No.2016YFC080176).
(責(zé)任編輯:王曉燕)