FHEAD0000000001XXXX20090901      0000009000Y1000XXX2
THEAD000000000220090901      ITM0000109393813            430143504352N22SP    000000000000RN000000010000EA  P0000000000000014390020090901      
TTAIL0000000003000000
FTAIL00000000040000000002

while IFS= read -r line
do
    case "$line" in
        FHEAD*) echo "${line:0:19}$(date +%Y%m%d)${line:27}";;
        THEAD*) echo "${line:0:15}$(date +%Y%m%d)${line:23}";;
        *) echo "$line";;
    esac
done <"file"

awk -v var="$(date +%Y%m%d)" '
/^FHEAD/{print substr($0,1,19) var substr($0,28); next}
/^THEAD/{print substr($0,1,15) var substr($0,24,106) var substr($0,138); next}
1' file > newfile

$ d="20091128"

$ echo $d
20091128

$ sed "s/\([FT]HEAD[0-9a-zA-Z]*\)[0-9]\{8\}/\1$d/g" urfile