SteadyState and verification of the Markov property for multiple Markov chains

Question

I'm building a Markov chain model to determine the probability of two states, WS (Working State) and CFS (Complete Failure State), for production machines. I applied the following code to generate a chain for each machine:

This is markov_df (the dataframe contains 3 years of history):

    markov_df <- data.frame(
  machine = c(
    "T", "T", "T", "T", "T", "T", "T",
    "Z-2H", "Z-2H", "Z-2H", "Z-2H", "Z-2H", "Z-2H", "Z-2H",
    "R-3H", "R-3H", "R-3H", "R-3H", "R-3H", "R-3H", "R-3H",
    "X-11H", "X-11H", "X-11H", "X-11H", "X-11H", "X-11H", "X-11H",
    "X-33H", "X-33H", "X-33H", "X-33H", "X-33H", "X-33H", "X-33H",
    "X-34H", "X-34H", "X-34H", "X-34H", "X-34H", "X-34H", "X-34H",
    "Y-7H", "Y-7H", "Y-7H", "Y-7H", "Y-7H", "Y-7H", "Y-7H",
    "Y-9H", "Y-9H", "Y-9H", "Y-9H", "Y-9H", "Y-9H", "Y-9H"),
  date = rep(
    c(
      "01/09/2024", "01/10/2024", "01/11/2024", "01/12/2024",
      "01/13/2024", "01/14/2024", "01/15/2024"),
    8),
  up_down = c(
    "WS", "WS", "WS", "WS", "WS", "WS", "WS",
    "WS", "WS", "WS", "WS", "WS", "WS", "WS",
    "WS", "WS", "WS", "WS", "WS", "WS", "WS",
    "WS", "WS", "WS", "WS", "WS", "CFS", "CFS",
    "WS", "WS", "WS", "WS", "WS", "CFS", "CFS",
    "WS", "WS", "WS", "WS", "WS", "CFS", "CFS",
    "WS", "WS", "WS", "WS", "WS", "WS", "WS",
    "WS", "WS", "WS", "WS", "WS", "CFS", "CFS"))


library(markovchain)

unique_machine <- unique(markov_df$machine)
data_transition <- data.frame(Machine = character(), State_origin = character(), 
                               Final_state = character(), Probability = numeric(), stringsAsFactors = FALSE)

for (machine in unique_machine) {
  df_machine <- subset(markov_df, machine == machine)
  matriz_transition_estimate <- markovchainFit(data = df_machine$up_down, method = "mle")
  matriz_transition_df <- as.data.frame(as(matriz_transition_estimate$estimate, "matrix"))
  names(matriz_transition_df) <- unique(df_machine$up_down) 
  matriz_transition_long <- reshape2::melt(matriz_transition_df, 
                                           varnames = c("State_origin", "Final_state"), 
                                           value.name = "Probability")
  matriz_transition_long$machine <- machine
  data_transition <- rbind(data_transition, matriz_transition_long)
}

data_transition <- data_transition %>% 
  group_by(machine) %>% 
  mutate(new_column = row_number())

data_transition$State <- ifelse(data_transition$new_column == 1, "CFS-CFS",
                                  ifelse(data_transition$new_column == 2, "WS-CFS",
                                         ifelse(data_transition$new_column == 3, "CFS-WS",
                                                ifelse(data_transition$new_column == 4, "WS-WS","Check"))))

Now I want to find the state probabilities (I have tried with steadyStates) and the Markov property (I have tried with verifyMarkovProperty) of each machine, and save the input in a dataframe with the machine's name and the results.

for (machine in unique_machine) {
  df_machine2 <- subset(markov_df, unique_machine == unique_machine)
  markov_chain <- markovchainFit(data = df_machine2$up_down, method = "mle")
  prob_ <- steadyStates(markov_chain)
  result[[machine]] <- prob_state
}

results_df <- do.call(rbind, lapply(seq_along(result), function(i) {
  data.frame(machine = names(result)[i], Estado = c("WS", "CFS"), Probability = result[[i]])
}))

But it's doesn´t works!